US20020011570A1 - Fire detector and housing - Google Patents

Abstract

A process and system for flame detection includes a microprocessor-controlled detector with at least three sensors. A wide band infrared sensor is used as the primary detector, with near band and visible band sensors serving to detect false-alarm energy from nonfire sources. Digital signal processing is used to analyze sensed data and discriminate against false alarms. A multistage alarm system can be provided, which is selectively triggered by the microprocessor. Spectral recording and analysis of prefire data is provided for. The detector can be housed in an enclosed, sealed, removable, plastic housing that may include an integral plastic window lens.

Description

CROSS REFERENCE TO RELATED APPLICATIONS
• This application is a continuation of U.S. application Ser. No. 09/384,808, filed on Aug. 27, 1999, which is a continuation application of U.S. application Ser. No. 08/866,029, filed on May 30, 1997, now U.S. Pat. No. 6,064,064, which is a continuation-in-part of U.S. application Ser. No. 08/690,067, filed on Jul. 31, 1996, now U.S. Pat. No. 6,046,452, which is a continuation-in-part of U.S. application Ser. No. 08/609,740, filed on Mar. 1, 1996, now U.S. Pat. No. 5,773,826, and also claims priority to PCT International application Ser. No. PCT/U.S. 97/03327, filed on Feb. 28, 1997. Each of the foregoing applications is hereby incorporated by reference as if set forth fully herein.[0001]
1. FIELD OF THE INVENTION
• The field of the present invention pertains to apparatus and methods for detecting sparks, flames, or fire. More particularly, the invention relates to a process and system for detecting a spark, flame, or fire with increased sensitivity, faster processing and response times, intelligence for discriminating against false alarms, and selective actuation of multi-stage alarm relays.[0002]
2. BACKGROUND
• To prevent fires, and the resulting loss of life and property, the use of flame detectors or flame detection systems is not only voluntarily adopted in many situations, but is also required by the appropriate authority for implementing the National Fire Protection Association's (NFPA) codes, standards, and regulations. Facilities faced with a constant threat of fire, such as petrochemical facilities and refineries, semiconductor fabrication plants, paint facilities, co-generation plants, aircraft hangers, silane gas storage facilities, gas turbines and power plants, gas compressor stations, munitions plants, airbag manufacturing plants, and so on are examples of environments that typically require constant monitoring and response to fires and potential fire hazard situations.[0003]
• To convey the significance of the fire detection system and process proposed by this patent application, an exemplary environment, in which electrostatic coating or spraying operations are performed, is explained in some detail. However, it should be understood that the present invention may be practiced in any environment faced with a threat of fire.[0004]
• Electrostatic coating or spraying is a popular technique for large scale application of paint, as for example, in a production painting line for automobiles and large appliances. Electrostatic coating or spraying involves the movement of very small droplets of electrically charged “liquid” paint or particles of electrically charged “Powder” paint from an electrically charged (40 to 120,000 volts) nozzle to the surface of a part to be coated.[0005]
• While facilitating efficiency, environmental benefits, and many production advantages, electrostatic coating of parts in a production paint line, presents an environment fraught with fire hazards and safety concerns. For example, sparks are common from improperly grounded workpieces or faulty spray guns. In instances where the coating material is a paint having a volatile solvent, the danger of a fire from sparking, or arcing, is, in fact, quite serious. Fires are also a possibility if electrical arcs occur between charged objects and a grounded conductor in the vicinity of flammable vapors.[0006]
• Flame detectors have routinely been located at strategic positions in spray booths, to monitor any fires that may occur and to shut down the electrostatics, paint flow to the gun, and conveyors in order to cut off the contributing factors leading to the fire.[0007]
• Three primary contributing factors to a fire are: (1) fuel, such as atomized paint spray, solvents, and paint residues; (2) heat such as derived from electrostatic corona discharges, sparking, and arcing from ungrounded workpieces, and so on; and (3) oxygen. If the fuel is heated above its ignition temperature (or “flash point”) in the presence of oxygen, then a fire will occur.[0008]
• An electrical spark can cause the temperature of a fuel to exceed its ignition temperature. For example, in a matter of seconds, a liquid spray gun fire can result from an ungrounded workpiece producing sparks, as the spray gun normally operates at very high voltages (in the 40,000 to 120,000 volt range). An electrical spark can cause the paint (fuel) to exceed its ignition temperature. The resulting spray gun fire can quickly produce radiant thermal energy sufficient to raise the temperature of the nearby paint residue on the booth walls or floor, causing the fire to quickly spread throughout the paint booth.[0009]
• A fire may self-extinguish if one of the three above mentioned factors is eliminated. Thus, if the fuel supply of the fire is cut off, the fire typically stops. If a fire fails to self-extinguish, flame detectors are expected to activate suppression agents to extinguish the fire and thereby prevent major damage.[0010]
• Flame detectors, which are an integral part of industrial operations such as the one described above, must meet standards set by the NFPA, which standards are becoming increasingly stringent. Thus, increased sensitivity, faster reaction times, and fewer false alarms are not only desirable, but are now a requirement.[0011]
• Previous flame detectors have had many drawbacks. The drawbacks of these previous devices have led to false alarms which unnecessarily stop production or activate fire suppression systems when no fire is present. These prior flame detectors have also failed to detect fires upon occasion, resulting in damage to the facilities in which they have been deployed and/or financial repercussions due to work stoppage or damaged inventory and equipment caused by improper release of the fire suppressant.[0012]
• One drawback of the most common types of flame detectors is that they can only sense radiant energy in one or more of either the ultraviolet, visible, near band infrared (IR), or carbon dioxide (CO[0013]₂) 4.3 micron band spectra. Such flame detectors tend to be unreliable and can fail to distinguish false alarms, including those caused by non-fire radiant energy sources (such as industrial ovens), or controlled fire sources that are not dangerous (such as a lighter). Disrupting an automated process in response to a false alarm can, as noted, have tremendous financial setbacks.
• Another drawback of previous fire detectors is their lack of reliability, which can be viewed as largely stemming from their approach to fire detection. The most advanced fire detectors available tend to involve simple microprocessor controls and processing software of roughly the same complexity as those used for controlling microwave ovens. The sensitivity levels of these previous devices are usually calibrated only once, during manufacture. However, the sensitivity levels often change as time passes, causing such conventional flame detectors to fail to detect real fires or to false alarm.[0014]
• Many of the conventional flame detectors also are limited by their utilization of pyroelectric sensors, which detect only the change in radiant heat emitted from a fire. Such pyroelectric sensors depend upon temperature changes caused by radiant energy fluctuations, and are susceptible to premature aging and degraded sensitivity and stability with the passage of time. In addition, such pyroelectric sensors do not take into account natural temperature variations resulting from environmental temperature changes that occur, typically during the day, as a result of seasonal changes or prevailing climatic conditions.[0015]
• Other types of conventional flame detectors identify fires by relying primarily on the ability to detect a unique narrow band spectral emissions radiated from hot CO[0016]₂(carbon dioxide) fumes produced by the fire. Hot CO₂gas from a fire emits a narrow band of radiant energy at a wavelength of approximately 4.3 microns. However, cold CO₂(a common fire suppression agent) absorbs energy at 4.3 microns, and can therefore absorb a hot CO₂spike emission generated by a fire. In such situations, conventional CO₂-based flame detectors can miss detecting a fire.
• Another type of conventional IR flame detector monitors radiant energy in two infrared frequency bands, typically the 4.3 micron frequency band and the 3.8 micron frequency band, while others use as many as three infrared frequency bands. The dual IR frequency band flame detector commonly utilizes an analog signal subtraction technique for subtracting a reference sensor reading at approximately 3.8 microns from the sensed reading of CO[0017]₂at approximately 4.3 microns. The triple IR frequency band flame detector uses an analogous technique, with an additional reference band at approximately 5 microns. These types of multiband flame detectors can false alarm when cold CO₂obscures the fire source from the flame detector, thereby misleading the detector into believing that a strong CO₂emission spike from a fire is detected, when, in fact, a negative absorption spike (caused by e.g., a CO₂suppression agent discharge or leak) has been detected.
• Conventional flame detectors using ultraviolet (“UV”) sensors also exist, but these too have drawbacks. Flame detectors with UV sensors may be sensitive to electrostatic spray gun flashes and corona discharges from waterborne coatings, which can cause false alarms and needlessly shut down production in paint spray booths. Also, because arc welding produces copious amounts of intense ultraviolet energy which can be reflected or transmitted over long distances, UV flame detectors can generate false alarms from such UV energy sources, even when the non-fire UV energy is located at a far distance from the spray booth. Moreover, after deployment, conventional UV detectors eventually can become highly de-sensitized as a result of absorbing smoke from a fire and/or solvent mist, causing the UV detector to become blinded. As a result, UV detectors can provide a false sense of security that they are operating at their optimum performance levels, when, in fact, the facility may be vulnerable to a costly fire.[0018]
• As an additional disadvantage, UV flame detectors generally require a relatively clean viewing window lens for the UV sensor, and can therefore become blinded or degraded by the presence of paint or oil contaminants on the viewing window lens. Moreover, the sensing techniques utilized with conventional UV detectors usually do not take into account the effects of such types of degradation.[0019]
• Besides problems with flame detection, many or all conventional flame detectors also have limitations or drawbacks relating to their housing and/or mounting that can affect their performance or longevity, in addition to being relatively expensive to manufacture. For example, most optical flame detectors have been built with metal housing made from costly aluminum, stainless steel, or similar materials. Such housings can be heavy, difficult to mount and may not be suitable for certain corrosive environments such as “wet-benches” used in semiconductor fabrication facilities for manufacturing silicon chips and the like.[0020]
• Further, most or all optical flame detector housings require a window lens (necessary for high optical transmission in the spectral bands used, and typically made of glass, quartz, sapphire, etc.), but it is usually quite difficult to obtain a tight seal of the window lens to metal housings, particularly in chemical manufacturing, or integrated circuit manufacturing or other applications having extremely rigorous environmental requirements. If the flame detector is not tightly sealed, then corrosive chemicals can leak into the electronic circuitry and degrade or destroy the unit.[0021]
• In flame detectors that detect UV energy, the protective window lens must be constructed from highly expensive quartz, sapphire, or other similar material that does not block UV energy. Moreover, the quartz or sapphire window lenses are typically placed in a metal detector housing, and are collectors of dust and contaminants due to the electrostatic effect of the high voltage field (around 300 to 400 volts) used in the UV detectors. To ensure that the UV detector's sensor(s) can “see through” the window lens, complex and costly “through the lens” tests are necessary. To conduct built-in “through the lens” window lens tests, a UV source tube is generally required to generate a UV test signal. Such UV source tubes require a high voltage for gas discharge sources and/or a large current for incandescent sources. Also, UV source tubes are subject to high failure rates. In sum, these self tests are expensive, require extra power and space, and are prone to breakdowns.[0022]
• There is a need for a sensitive, reliable, fully enclosed, inexpensive, light-weight, intelligent, and effective method and system for detecting sparks, flames, or fire with little or no interruptions caused by false alarms.[0023]
SUMMARY OF THE INVENTION
• The present invention is directed in various aspects to a sensitive, reliable, fully enclosed, inexpensive, light-weight, intelligent, and effective method and system for detecting sparks, flames, or fire with little or no interruptions caused by false alarms. According to one embodiment of the present invention, a process and system for flame detection includes a sensor array for providing sensor signals, a temperature sensor for providing signals indicative of ambient temperature conditions, and either internal or external signal processing electronics for processing the sensor signals and generating a response, if necessary.[0024]
• In a first aspect of the invention, a microprocessor-controlled detector advantageously includes at least three sensors. Preferably, a wide band infrared sensor is used as the primary detector, with near band and visible band sensors serving to detect false-alarm energy from nonfire sources. The system may comprise a single or a series of detector units with wide spectrum sensing capabilities (quantum sensors) located within a desired facility such as, e.g., inside a paint spray booth.[0025]
• In a second, separate aspect of the invention, a multistage alarm system is provided. In a preferred embodiment, the multistage alarm system is selectively triggered by a microprocessor. Sensor data captured at detector units can be interfaced to either internal or external signal processing electronics which process and analyze the sensor data, and selectively trigger multistage (e.g., two- or three-stage) alarm relays. The signal processing electronics and the relays may advantageously be located within the detector unit, or can be remotely located in a controller unit.[0026]
• In a third, separate aspect of the invention, digital signal processing is used to analyze sensed data and discriminate against false alarms. False alarms also are avoided through periodic conduction of comprehensive diagnostic evaluations of the system components. The system programs parameters for its system components and varies the parameters depending upon ambient conditions. Algorithms and techniques for eliminating false alarms are employed, thereby providing effective detection of any sign of a spark, flame, or fire.[0027]
• In a fourth, separate aspect of the invention, prefire spectral data is recorded both before and after a fire situation. The recorded spectral data may be later analyzed to identify the cause of the fire and thereby help eliminate the occurrence of future fires. Sensor data is captured, processed, and analyzed at the detection location. Preferably, the spectra of the radiated detected energy from a fire or potential fire is stored in memory, providing a comprehensive record of sensor array spectral data (processed or unprocessed). This information may be retrieved after a fire occurs for analysis.[0028]
• In a fifth, separate aspect of the invention, a fire detector is preferably housed in a sealed, self-contained housing. The housing may include a window region to protect the sensors wherein the window region is constructed from a different material than the housing material. The housing may be placed within a wall, workbench, wet-bench, or other suitable mounting structure, and sealed therewith, welded or similarly attached thereto. In another embodiment, the housing comprises a base portion and a removable upper lid portion. Attached to the upper lid portion is a module containing the sensors and sensitive electronic circuitry used in the flame detector. The removable upper lid portion of the housing allows relatively quick and easy replacement of the primary detection components of the fire detector.[0029]
• Further embodiments and variations of the invention are also disclosed herein.[0030]
BRIEF DESCRIPTION OF THE DRAWINGS
• FIG. 1 is a diagrammatic illustration of an electrostatic coating booth, in which a fire detector according to the present invention may be employed.[0031]
• FIG. 2 is a graphical representation of the wide spectrum sensitivity afforded by the process and system of FIG. 1, in which the protective cover has the same transmittance characteristics as the detector.[0032]
• FIG. 3 is a graphical representation of the sensitivity of a flame detector having wide band IR, near band IR and visible band sensors.[0033]
• FIG. 4 is a perspective view of one housing embodiment in accordance with certain aspects of the present invention.[0034]
• FIG. 4[0035]ais a perspective view of a protective cover with wide spectrum transmittance characteristics.
• FIG. 4[0036]bis a diagrammatic illustration of a fire/flame detector having a fiber optic cable assembly with a protective cover to facilitate use in confined or unaccessible areas.
• FIG. 4[0037]cis a cross-sectional view taken along line4d-4dthrough FIGS. 4 and 4a.
• FIG. 5 is a graph illustrating the regions of sensitivity of a particular wideband sensor array in accordance with various aspects of the present invention.[0038]
• FIG. 6 is a diagrammatic illustration of an enclosed, removable, self-contained module for optical fire/flame detectors.[0039]
• FIG. 7[0040]ais a diagrammatic side view illustration of a plastic, sealed housing with an integral window lens.
• FIG. 7[0041]bis a top view of the housing of FIG. 7a.
• FIG. 8 is a diagrammatic illustration of a side view of a plastic, sealed housing with a thin window-lens area for high transmittance.[0042]
• FIG. 9 is a diagrammatic illustration of a side view of a plastic, sealed housing with an embedded window lens made of such materials as quartz or sapphire.[0043]
• FIG. 10 is a diagrammatic illustration of a side view of housing that is heat welded to a plastic cable.[0044]
• FIG. 11 is a block diagram representation of one embodiment of a flame/fire detection system in accordance with various aspects of the present invention, wherein a single or a series of flame detector components are located inside a desired facility, such as a paint booth, and a controller component of the system is located outside the facility for processing data captured by the sensors.[0045]
• FIG. 12 is a block diagram representation of an alternative embodiment of a flame/fire detection system, wherein a single or a series of detectors incorporate a microprocessor and process data captured by the system in the detector component.[0046]
• FIGS. 13[0047]aand13bdepict a table comparing fire/flame temperature and radiant energy calculations in various spectral regions.
• FIG. 14 is a graph of radiant energy as a function of fire temperature.[0048]
• FIGS. 15 and 16 are graphs comparing a detected radiant energy for a wide band spectral detector versus a narrow band 4.3 micron infrared detector as a function of fire temperature.[0049]
• FIG. 17 is a graph illustrating relative radiant emittance at various wavelengths for a 2500 K degree fire.[0050]
• FIG. 18 is an illustration of a record and fields of data that may be stored upon occurrence of a fire.[0051]
• FIG. 19 is a diagram of an event log generated by the system upon detection of a fire signature warranting an “alert” condition.[0052]
• FIG. 19[0053]ais an exemplary fire signature which upon observation would result in an “alert” condition being declared.
• FIG. 20 is a diagram of an event log generated by the system upon detection of a fire signature warranting a “fire early warning” condition.[0054]
• FIG. 20[0055]ais an exemplary fire signature which upon observation would cause a “fire early warning” condition to be declared.
• FIG. 21 is a diagram of an event log generated by the system upon detection of a fire signature warranting an “alarm” condition.[0056]
• FIG. 21[0057]ais an exemplary fire signature which upon observation would cause an “alarm” condition to be declared.
• FIG. 22 is a logic flow diagram of processing as may be embodied in the present system, illustrating diagnostic evaluations or tests performed by the system.[0058]
• FIG. 23 is a logic flow diagram of processing as may be embodied in the present system, illustrating a lens test performed by the system.[0059]
• FIG. 24 is a portion of a logic flow diagram of processing as may be embodied in the present system, illustrating a preferred logic flow and sequence of steps performed during overall operation of the system.[0060]
• FIG. 25 is a portion of a logic flow diagram of processing as may be embodied in the present system, illustrating a logic flow and continued sequence of steps for detecting an “alert” condition.[0061]
• FIG. 26 is a portion of the logic flow diagram of processing as may be embodied in the present system, illustrating a logic flow and continued sequence of steps for detecting a “fire early warning” condition and an “alarm” condition.[0062]
• FIG. 27 is a logic flow diagram illustrating operation of a system with a two-stage alarm relay.[0063]
• FIG. 28 is a functional diagram illustrating a preferred fire discrimination algorithm.[0064]
• FIG. 29 is a graph showing filtered sensor outputs after processing through two filters with different time response characteristics.[0065]
• FIG. 30 is a graph showing a compensation made to a slow filter output of the wide band IR sensor.[0066]
• FIG. 31 is a diagram illustrating the effect of asymmetrical digital filtering on the visible band sensor output for the purpose of rejecting certain false alarm sources such as a flashlight.[0067]
• FIG. 32 is a diagram illustrating operation of a flicker detection algorithm for rejection of other false alarm sources.[0068]
• FIG. 33 is a diagram illustrating an algorithm for rejecting false alarm sources such as industrial ovens.[0069]
• FIG. 34 is a diagram of a circuit for processing a sensor input signal.[0070]
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
• A process and system for detecting sparks, flames, or fire in accordance with a preferred embodiment of the present invention is described herein. It should be noted that the terms “fire detector,” “flame detector” and “fire/flame detector” are used interchangeably in the present text and refer generally to any process and/or system for detecting sparks, flames, or fires, including explosive type fires or fireballs and other dangerous heat-energy phenomena.[0071]
• A particular embodiment of a process and system for fire detection is described in conjunction with an exemplary situation of an electrostatic coating operation. However, it should be understood that the process and system may be effectively utilized in any environment facing a threat from sparks, flames, or fire. For example, the process and system may be used in such applications as petrochemical facilities and refineries, semiconductor fabrication plants, co-generation plants, aircraft hangars, gas storage facilities, gas turbines and power plants, gas compressor stations, munitions plants, airbag manufacturing plants, and so on.[0072]
• FIG. 1 illustrates an[0073]exemplary environment10, as for example, a coating zone, such as a spray or paint booth or enclosure, in which electrostatic coating operations are routinely performed. As illustrated in FIG. 1,parts12 are transported through thespray booth14 by aconveyor16 connected to a reference potential orground18. The direction in which the conveyor moves is indicated by anarrow20. Theparts12 are typically supported from the conveyor by a conductive hook-like support orhanger22. Theparts12 are passed proximate ahigh voltage source13 with ahigh voltage antenna15. Thehigh voltage source13 may be one available from Nordson as Model number EPU-9. Electrical charge is transferred from the high voltage source, which may operate between 60,000-120,000 volts, to theparts12 to be coated.
• The electrostatic coating system illustrated in FIG. 1 represents an air electrostatic spray system of a type used in many industrial operations. A typical[0074]industrial spray system24 includes aspray gun26 coupled to apower supply27, a paint supply container28 (for example, a pressure tank), and some form ofspray control mechanism30. Thespray control mechanism30 may include an air compressor and an air regulator (not separately shown).
• A single[0075]flame detector component32 is located at a strategic position within thespray booth14. Thedetector component32 can be advantageously manufactured from a substantially explosion-proof material, as discussed in greater detail below. Depending upon the size of thespray booth14 or other facility, a plurality of suchflame detector components32 may be strategically located throughout thespray booth14 or other facility.
• Referring also to FIG. 4, the[0076]flame detector32 embodying features in accordance with a preferred embodiment of the present invention is sensitive to radiant energy in the visible (VIS) band, near band infrared (NIR), and wide band infrared (including middle band infrared (MIR)) spectra. Theflame detector32 preferably has a spectrum sensitivity for infrared energy, within a range from roughly 700 to 5000 nanometers (0.7 to 5 microns), and for visible energy, within a range from approximately 400 to 700 nanometers. Theflame detector32 is preferably enclosed within aprotective housing132. The type of housing132 (i.e., shape, material and/or configuration) may vary depending upon application and such things as environmental factors. Various housings are discussed in more detail hereinafter.
• The[0077]housing132 of theflame detector32 is, in a particular embodiment, constructed with aviewing window132bdisposed over the sensors of the flame detector. Theviewing window132bis advantageously provided where one or more IR sensors (by itself or in conjunction with other sensors) are used so that the flame detector may detect IR frequency bands. In such embodiments, theviewing window132bwould be comprised of an IR-transparent material. In such embodiments, theviewing window132bmay be quartz, sapphire, glass or a plastic material such as hydrocarbon or fluorocarbon polymer, for example. Other embodiments of housings are also disclosed herein that do not utilize a viewing window that is constructed from a material that is different from the housing material.
• In one embodiment, the[0078]flame detector32 has aprotective cover132adisposed over it that is conformed to fit snugly over theprotective housing132. Theprotective cover132ais preferably constructed from a relatively inexpensive material such as a plastic, such as polypropylene or polyvinyl chloride, that is transmissive with respect to IR and visible wavelengths. Thus, theprotective cover132amay be easily disposed, recycled, or reused, as desired. To avoid accumulation of paint and grime on aviewing window132b, theprotective cover132acan be configured to slip easily over thehousing132 of theflame detector32.
• So as not to obstruct the wide spectrum sensitivity of the[0079]flame detector32, theprotective cover132a(see FIG. 4a) preferably has wide spectrum transmittance characteristics that enable optimum sensing of any flame, spark or ignition that may occur. The transmittance characteristics of theprotective cover132aare also illustrated in FIG. 2.
• Referring to FIGS. 4, 4[0080]a, and4c, theprotective cover132a, which appears somewhat like a top hat, is preferably configured to conform around a cylindrical protrudingportion272 of thehousing132 of theflame detector32. In order to prevent accumulation of spray paint, grime, oil contaminants, or the like on aviewing window132bof theflame detector housing132, theprotective cover132ashould completely cover theviewing window132b. Preferably, theprotective cover132ahas aplanar face270 and acylindrical body271. Thecylindrical body271 extends sufficiently along the protrudingportion272 and is sufficiently detached from the protrudingportion272 to prevent any movement of airborne paint particles toward theviewing window132b.
• As specifically illustrated in FIGS. 4 and 4[0081]a, thecylindrical body271, at itsbase273, terminates in aperpendicularly projecting flange274. Theflange274 also serves to prevent airborne paint particles from moving toward theviewing window132b. A centrally locatedgroove275 runs along its circumference, almost contacting the protrudingportion272 of thehousing132 of theflame detector32, which serves to further prevent airborne paint particles from reaching theviewing window132b. Slight pressure applied on theprotective cover132a, to ease theprotective cover132aover thehousing132 of theflame detector32, causes thegroove275 to slide over alocking mechanism278 of thehousing132 of the flame detector32 (best illustrated in FIG. 4c). Thegroove275 serves to hold theprotective cover132a, albeit flexibly, in place.
• The[0082]flange274 has a plurality of reinforcingmembers276 projecting outwardly toward its outer periphery. The reinforcingmembers276 preferably lend theflange274 enough rigidity to allow a person to easily pull it off theflame detector housing32 when replacing theflame detector32.
• The[0083]protective cover132amay be constructed from any suitable material having the required transmittance characteristics. The material used in the illustrated embodiment is relatively inexpensive, has some rigidity, yet is also resilient. In the illustrated embodiment of theprotective cover132a, a clear polyvinyl chloride (PVC), with an “ORVIS®-K” coating to serve as an anti-static agent, is used. Theprotective cover132ais preferably fabricated from clear PVC with a starting gauge of 20 mil, which is vacuum drawn over a machined, metal mold to yield thin, flexible protective covers. Theprotective cover132amay alternatively be fabricated from materials such as LEXAN®, which may be injection molded. Other plastics with similar transmittance characteristics may alternatively be used. The illustratedprotective cover132amay be easily disposed, recycled, or reused after cleaning, as desired.
• Alternatively, the[0084]protective cover132amay be configured as a bag or a planar surface in any shape or form necessary to cover theviewing window132b, with a string or wire to fasten theprotective cover132ato the protrudingportion272 of thehousing132 of theflame detector32. Additionally, although the protective cover is preferably constructed from a light weight, inexpensive and disposable material, any material that transmits radiant energy having wavelengths between 700-5000 nanometers is appropriate.
• Referring now to an embodiment shown in FIG. 4[0085]b, theprotective cover132amay vary in dimensions to suit various applications. Flame detectors are routinely used in confined areas, such as cabinets, processing equipment (including mixers of explosive materials), extruders, and the like. For example, a small, almost miniature, version of theprotective cover132a, as illustrated in FIG. 4b, may be used at aviewing end277 of afiber optic cable279, which is attached at asecond end280 to aflame detector320. Use of thefiber optic cable279 can facilitate remote location of theflame detector320 and enable transmission of the radiant energy patterns detected to theflame detector320.
• Sensor data captured by the[0086]flame detector32 can be relayed to a central control system34 (see FIG. 1), which, in paint spray booth applications, may be located outside thespray booth14. Thecentral control system34 may take the form of a computer with a central microprocessing unit, a display monitor, a suitable memory, and printing capabilities. Thecentral control system34 may coordinate functioning of theflame detectors32 with other detection systems, as for ungrounded parts or the like.
• The[0087]housing132 of theflame detector32 also can be constructed from polypropylene, which is inert to harsh chemical environments and beneficial for transmittance of infrared spectra. In addition to these advantages, use of polypropylene may permit the housing to be heat-sealed so as to create a sealed, watertight environment. The housing may incorporate structural elements or be otherwise reinforced to enable it to withstand explosions and to give it an explosion proof rating. Potting resins can be added to reduce cavities that typically trap gases or fumes.
• The[0088]flame detector32 preferably operates by searching for radiant energy characteristics or patterns of a flame or fire. A continuous stream of spectral data from a sensor array38 (as illustrated in FIG. 11 or12 and described hereinafter) may be analyzed by a controller (microprocessor, or microcomputer)unit39 or the controller (or microprocessor, or microcomputer)36. In a preferred embodiment, an Intel 8051 microprocessor or microcomputer is utilized.
• In another embodiment, a removable, enclosed, self-contained electro-[0089]optical module220, illustrated in FIG. 6, is used for optical fire/flame detection. The electro-optics and electronics of the optical fire/flame detector220 (whether analog, digital, or a combination thereof) preferably comprise one or more sensors operating in one or more of the ultraviolet band, visible band, near band infrared, wide band infrared, or narrow band infrared (such as a 4.3 micron narrow band IR).
• FIG. 6 shows a surface-[0090]mountable flame detector220 constructed according to the present invention, having a structure particularly well suited to ease of replacement of the electronics and electro-optical components of the flame detector without having to re-mount an entirely new flame detection unit. Theflame detector220 comprises adetector housing base222 having a relatively flat back portion orplate214 which may be placed flush against a surface to which theflame detector220 is mounted (e.g., a wall). Theback plate214 of thedetector housing base222 has mountingholes215 placed at appropriate locations to secure thedetector housing base222 to the mounting surface, and may be mounted to a swivel assembly or hard-mounted to a surface.
• The[0091]detector housing base222 is connected to one ormore conduits210 which contain signal andpower wires226, as illustrated in FIG. 6 through the cut-away portion of thedetector housing base222. The signal andpower wires226 are connected to a removable plug-inconnector213. Thedetector housing base222 is preferably cylindrical in shape, and has circular threadings216 around its outer, upper periphery, as illustrated in FIG. 6. A threadeddetector housing lid227, having a cylindrical shape, is adapted to fit over thedetector housing base222 and has threadings on its interior portion that allow it to be placed snugly over thedetector housing base222 in a manner similar to a threaded nut and screw. Thedetector housing lid227 has a solid, relatively flat upper surface orplate217 on one side so as to enclose an inner detector module224 when thedetector housing lid227 is placed securely over thedetector housing base222.
• The inner detector module[0092]224 comprises the electronics and electro-optics, including thesensors225 of theflame detector220. The inner detector module224 from a physical standpoint comprises an enclosed chamber in which the sensitive electronics reside, and can be self-contained in the sense that the flame detection circuitry and other components reside within the enclosed chamber. In the alternative, the sensors are enclosed within the enclosed chamber and the processing circuitry can be external to the enclosed chamber. The inner detector module224 is physically attached directly to thedetector housing lid227, and more specifically to theupper plate217 of thedetector housing lid227, such that when thedetector housing lid227 is removed the inner detector module224 removes along with it. Thesensors225 are preferably located adjacent to theupper plate217, as illustrated in FIG. 6.
• The[0093]detector housing lid227 may optionally be provided with aviewing window212. Alternatively, thedetector housing lid227 need not have aviewing window212. In such a case, theupper plate217 of the detector housing lid227 (and possibly the entiredetector housing lid227 as well as the detector housing base222) is preferably comprised of a material that has low bulk adsorption characteristics for radiant energy having frequency components between approximately 400-5000 nanometers, such as glass, polypropylene or other suitable plastic material.
• If it is desired to replace the[0094]flame detection unit220 to perform maintenance or tests on the unit, or because of a functional problem in the sensors or circuitry, or for any other reason, the electronics and electro-optical components may be relatively easily replaced by unscrewing thedetector housing lid227 and replacing it with anotherdetector housing lid227. Because the inner detector module224 is connected to thedetector housing lid227, the key components of the flame detector can be relatively quickly and easily replaced without having to re-mount the flame detector to the surface. The samedetector housing base222 that was used for the olddetector housing lid227 continues to remain operable for the new one. The removable plug-inconnector213 also facilitates rapid and relatively easy substitution of the new unit for the old one.
• There are a number of advantages associated with the housing structure of the flame detector embodiment shown in FIG. 6. For example, sensitive detector electronics and electro-optics which form a part of the inner detector module[0095]224 are better protected from handling-induced electrostatic discharge and physical handling damage. Also, as noted above, the critical components of theflame detector220 can be replaced easily and quickly in the field without removing or dismantling thedetector housing base222, which would otherwise necessitate detaching theback plate214 from the surface to which it is mounted as well as detaching the attached conduit(s)210. Because the sensitive electronics are largely protected within the inner detector module224, the installation process can in many instances be quickly and easily performed, once thedetector housing base222 is secured and connected to the conduit piping. Benefits may also be achieved in testing, storing and shipping theflame detector220 with less concern for electrostatic discharge damage at either the factory or the distributor/integrator facility.
• The[0096]detector housing lid227 anddetector housing base222 can be of any shape or size, so long as thedetector housing lid227 fits securely within thedetector housing base222. Further, thedetector housing lid227 anddetector housing base222 need not necessarily be threaded so as to attach by screwing together, but may also snap together or be secured by other suitable or conventional means. Also, thedetector housing lid227 anddetector housing base222 can be made from plastic, metal, or any other suitable material, or any combination thereof.
• A second embodiment of a flame detector with a housing is illustrated in various angles, cross-sectionals and details in FIGS. 7[0097]a,7b,8,9 and10. Referring first to FIG. 7a, a self-contained, sealedhousing230 encloses the electronics and electro-optical components of a flame detector. At least one printed circuit board (PCB)233 is mounted to one or more walls of the sealedhousing230. On the PCB233 integrated chips could be mounted which comprise the electronic circuitry and/or electro-optical circuitry of the flame detector. Also on the PCB233 asensor236 or array of sensors for detecting radiant light can be mounted. The sealedhousing232 can be made from a variety of materials and may comprise awindow region232 that is formed of thinner housing material or material different from the housing. Advantageously, thewindow region232 is integral with thehousing230, thereby allowing thehousing230 to be constructed of lighter and less costly material than conventional metal housings. Depending upon the environment in which the detector will be used, different materials may be used to construct the housing. For example, in environments that may subject the housing to corrosive chemicals, such as acids, polypropylene may be utilized. In environments in which greasy hydrocarbons are present such as an oil rig, Teflon may be appropriate. In other environments, ceramic or metal housings may be desirable. In addition, the housing may be constructed of a first material and then coated with a second material that provides the best protection for the device in the intended environment.
• As a result of the fact that no UV sensors are necessary to practice the present invention, the[0098]window region232 of thehousing230 need not be a quartz or sapphire lens as is required in flame detectors that rely upon UV sensors. Historically, quartz or sapphire windows have been necessary with UV sensors because these two materials have low bulk absorption characteristics for UV wavelengths. However, since the present invention utilizes infrared and visible wavelengths, any material that does not significantly absorb the desired IR and visible wavelengths can be used as a window for the IR sensors. In addition some materials such as Teflon or polypropylene are resistant to the accumulation of contaminants. Accordingly, thehousing230 is preferably constructed of a corrosion-resistant and contamination-resistant material such as Teflon or polypropylene, or some similar material or hybrid. Use of such a contamination-resistant material largely eliminates the need for conventional “through the lens” testing as is commonly required for prior flame detectors (particularly those using UV sensors) having a glass, quartz or sapphire window lens that is not contamination-resistant.
• Further, use of a wideband infrared sensor or sensor array (spanning a light frequency band from about 0.7 microns to 5 microns) as a primary sensor in the flame detector, and elimination of UV sensors, results in a construction wherein the wideband infrared sensors are able to effectively “see through” the[0099]housing230 due to the longer wavelengths detected by such sensors. The wide band infrared sensor or sensors should be able to “see through” most contaminants because most contaminants do not absorb significant amounts of infrared energy.
• Therefore, the[0100]flame detector housing230 can be manufactured from a material such as polypropylene, polyvinyl chloride, ABS plastic, Teflon, glass, fiberglass, spun glass, or a combination of several materials and/or additives. Theintegral window region232 can be made from the same material as thehousing230, SO that sealing and mounting problems associated with prior art flame detectors that required quartz, or sapphire window lenses can be provided. Thehousing230 may be heat sealed or sealed with an “O”ring239, as illustrated in FIG. 7a. A shieldedcable238 may be welded or otherwise connected to thehousing230, as further described hereinafter in more detail with respect to FIG. 10.
• In an alternate embodiment, shown in FIG. 8, the[0101]detector housing230 can be constructed from a combination of polypropylene, polyvinyl chloride, ABS plastic, Teflon, fiberglass, spun glass, quartz, glass, sapphire, etc., or a combination of several materials and/or additives. The outerintegral window region232 of thehousing230 can be made from the same material as thehousing body230 and reinforced under thewindow region232 with a suitable material such as polypropylene or glass, so that sealing and mounting problems associated with conventional glass, quartz, or sapphire window lenses are eliminated.
• A[0102]Teflon housing230 with a thin area immediately adjacent to thewindow lens area232 is used for the embodiment shown in FIG. 8. Because the bulk absorption of Teflon is relatively high for longer (infrared) wavelengths, the area of thewindow region232 is preferably thin so that the absorption of the desired wavelengths is minimized. A reinforcingplug234 of low absorption material can be screwed in, snapped in, molded in (e.g., in the injection molding process), heat welded, or otherwise attached to thehousing230 as shown in FIG. 8 in the hollow space caused by thethin window region232, such that thethin window region232 is more resistant to puncture or physical damage. The reinforcing plug can be constructed from any material having low bulk absorption characteristics for the desired wavelengths, including the materials mentioned above.
• In another alternate embodiment, illustrated in FIG. 9, the[0103]detector housing230 is made from a combination of polypropylene, polyvinyl chloride, ABS plastic, Teflon, fiberglass, spun glass, quartz, glass, sapphire, etc., or a combination of several materials and/or additives. Thewindow region232 can be made from any different material, which material is secured or embedded (e.g., by heat sealing) into theplastic housing230 during the injection molding process, or is heat sealed into the material after the housing fabrication. Thus, problems associated with sealing and mounting problems glass, quartz, or sapphire window lenses are eliminated.
• As discussed above, few contaminants will adhere to the[0104]housing230 when constructed of Teflon, polypropylene or another of the preferred corrosion-resistant and contamination-resistant materials. The use of such contaminant-resistant materials, without the need for a window lens of sapphire, quartz or other UV-transparent material, is made possible by the present flame detectors which do not require UV detectors. The housing for the present flame detectors are easier to maintain than conventional types of flame detectors, particularly those detectors that use UV sensors, which require regular and frequent cleaning of the sensor window lens to remove contaminants. In addition, because plastics are generally lighter than metal, quartz and sapphire, mounting and installation costs can be reduced. The preferred, light-weight, low-cost, sealed,plastic detector housing230 is constructed so as to fall withinNEMA 4, 12, etc. outdoor ratings and hazardous ratings such as Class I and II andDivisions 1 and 2. These ratings are necessary for operating the optical detector in a hazardous, corrosive, and/or outdoor environment.
• In another embodiment, the self-contained[0105]housing230 is sealed using one or more “O” rings and screws, rivets, etc. The preferred method of sealing is heat sealing, which melts the housing materials (e.g., plastic or glass) into essentially one solid piece. The optical flame/fire detectors can usevarious sensors236, including wide band IR, nearband IR, visible band, or narrow band IR (such as 4.3 micron infrared, for example), with one or more IR rejection bands for multiple frequency IR detection. The optical transmittance of thewindow region232 should be high enough for the selectedsensors236 to sense radiant energy in their respective light frequency detection bands.
• The[0106]window region232 for an all-plastic or allfiberglass detector housing230 for example, can be thinner than the rest of thehousing230 so as to decrease the bulk absorption losses of the window region. In a an embodiment using a combination of wide band IR, near band IR and visible light sensors, preferably thewindow region232 has a low bulk absorption for wavelengths in the range of approximately 0.4 to 5 microns, or about 400 to 5000 nanometers. Preferably, various reinforcement techniques may be used such as ribbing or braces, or the high-optical-transmittance plug described previously with respect to FIG. 8.
• As an additional advantage, manufacture of the[0107]housing230 from a material such as polypropylene, Teflon, fiberglass and the like (with or without additives, if desired) can render thepreferred detector housing230 resistant to corrosion and/or acid. Thedetector housing230 can be machined out of the bulk material, injection molded or blow molded, for example, and is preferably heat sealed to ensure complete isolation of the internal electro-optics and electronic from any corrosive elements that could leak through conventional sealing techniques such as “O” rings and screws.
• A power and[0108]signal cable238 connects to thehousing230 and can be made from the same material as thehousing230. The power/signal cable238 can be sealed to thehousing230 with clamps or “O” rings, for example, or can be heat sealed to thehousing230 itself, as illustrated in FIG. 10. For example, in an embodiment in which thehousing230 and power/signal cable238 are made from polypropylene, the housing/cable interface can be heat sealed for optimum sealing because thecable238 andhousing230 can be melted into an integral piece. In addition, in an acidic environment such as a semiconductor wet-process bench, one or more protective sleeves are advantageously placed around the power/signal cable238 to add further protection to the electronic wires enclosed within the power/signal cable238.
• There is currently a need to deploy fire/flame detectors in hostile, caustic, acidic environments, both indoors and outdoors. In a preferred embodiment suitable for many such environments, a flame detector is physically installed within a wall, workbench, or similar structure. Preferably, the flame detector housing is such as described for any of the embodiments shown in FIGS. 7[0109]a,7b,8,9 and10. For example, a self-contained, sealed flame detector housing made of polypropylene can be installed inside the corrosive compartments of a semiconductor clean-room chemical wet-process bench. Installing a sealed, corrosion-proof, plastic-housed optical fire/flame detector inside the wet bench compartments has the advantage of reducing exposure to contaminants and placing the detector closer to the potential fire source for improved fire response. Also, the flame detector housing itself is less likely to become a repository of miscellaneous contaminants present in the worksite.
• A sealed, corrosion-proof, plastic-housed optical fire/flame detector can similarly be used on an offshore oil platform where maintenance costs are high and low weight is important. Preferably, the housing is constructed from or coated with Teflon, which is highly resistant to the accumulation of oil deposits. In addition, because the flame detector does not require frequent maintenance, i.e., window-lens cleaning, as do conventional fire detectors, substantial cost savings from reduced maintenance can be realized.[0110]
• FIGS. 11 and 12 are block diagrams depicting embodiments of a flame detector utilizing wide band IR detection. In accordance with one embodiment of the present system, a[0111]single flame detector32 located at a particular location, indicated by reference letters FD1, or a plurality of flame detectors, located at a plurality of different locations, indicated by reference letter FDN, may be located, for example, inside the spray booth14 (see FIG. 1). Apower supply46, typically operating at 24 volts, supplies power to theflame detector32.
• In addition to the[0112]sensor array38, theflame detector32 may include an analog to digital (A/D)converter50, which receives a continuous stream of analog sensor signals from each of thesensors40,42,44 of thesensor array38, and converts the analog signals into digital signals for storage and selective processing by amicroprocessor36 or acontroller39, or both. Atemperature sensor52 located within theflame detector32 serves to indicate ambient temperature values for calibration purposes. Amemory component54 within theflame detector32 comprises ROM (Read Only Memory) and RAM (Random Access Memory) for temporary and permanent storage of data, as for storing instructions for themicroprocessor36, for performing intermediate calculations, or the like. In a preferred embodiment, thesensor array38 preferably has asensor40 for sensing radiant energy within the visible band spectrum, asensor42 for sensing radiant energy within the near band infrared spectrum, and asensor44 for sensing radiant energy within a wide band infrared (WBIR, or MIR) spectrum.
• Referring now to FIG. 2, the[0113]first sensor40 searches for and detects radiant energy within the visible band range extending from about 400 nanometers to approximately 700 nanometers, indicated in FIG. 2 by the frequency band designated as “VIS.” Thesecond sensor42 searches for and detects radiant energy within the near band infrared range extending from roughly 700 nanometers to about 1100 nanometers, indicated in FIG. 2 by the frequency band designated as “NEAR BAND IR.” Thethird sensor44 searches for and detects radiant energy within a wide band infrared range extending from about 700 nanometers to about 5000 nanometers, indicated in FIG. 2 by the frequency band designated as “WIDE IR SPECTRUM.”
• In a preferred embodiment, wide band IR (WBIR) is used as the primary sensor in the optical fire/flame detectors. The WBIR sensor preferably detects radiant energy over a spectral band from about the end of the visible spectrum (about 0.7 microns) to the band comprising the longer IR wavelengths (up to about 3.5 microns). The[0114]WBIR sensor44 can, however, be susceptible to various false-alarm sources, including sunlight, bright lights, ovens, and other sources of wideband IR radiation. In order to successfully use WBIR as the primary sensor without false alarms from broadband energy sources, information obtained from sensing energy in the visible band (VB) (from about 0.4 to 0.7 microns) and/or the near band IR (NBIR) (from about 0.7 to about 1.5 microns) is used in conjunction with signal processing algorithms to prevent false triggering. A preferred process and system therefore comprises an “intelligent” multi-spectral approach to optical fire/flame detection.
• In a preferred embodiment, compensating digital signal processing algorithms are performed in the optical fire/flame detectors by the[0115]microprocessor36 orcontroller39 to distinguish between actual fires and non-fire energy sources. Such algorithms preferably include time correlation of the sensor signals along with VB and NBIR energy bands and a comparison of relative energy levels in the different energy bands.
• In order to rapidly detect all types of fires, whether hydrocarbon and nonhydrocarbon in nature, a preferred optical fire/flame detector senses energy over a wide, continuous spectral band of infrared radiant energy. Preferably, the energy band observed by the fire/flame detector covers the range from about 0.4 to about 5 microns (i.e., spectral ranges in the VB, NBIR, and WBIR) to ensure that virtually all types of fires are detected. This spectral range constitutes the bulk of the radiant heat energy generated by an unwanted fire, including, for example, burning polypropylene or PVC plastic.[0116]
• FIG. 17 is a graph illustrating the energy emitted at various wavelengths by an exemplary fire source. As shown in FIG. 17, a large portion of the energy emitted by a typical fire occurs at wavelengths other than the 4.3 micron range. Accordingly, fire detections that rely solely on observing a CO[0117]₂spike in the narrow 4.3 micron band are in effect observing only a small fraction of a fire's total energy radiation. In contrast, the preferred fire/flame detector observes a much wider portion of the energy emitted by a fire. However, because nonfire sources such as sunlight and artificial light may also be observed by the fire/flame detector, a mechanism for discrimination between fire and non-fire energy sources is desireable.
• The discriminator in the fire/flame detector can be programmed or otherwise configured to make advantageous use of known or observed characteristics of different types of fires, in order to more readily distinguish fire and non-fire energy sources. By way of general background, all materials that burn in the condition known as an unwanted fire, which can be described as uncontrolled rapid oxidation, emit wideband blackbody radiant energy and molecular narrow band line emissions, such as the 4.3 micron CO[0118]₂spike. (The term blackbody refers to a material's emissivity, and not its color.) The blackbody radiant emissions of a fire are always present and predictable because they are a function of the temperature of the materials being consumed by the fire, the temperature of the fire's gaseous flames and solid particulates, and the average emissivity of the flames, particulates, and burning material. Radiant emissions are the transfer of heat from one body to another without a temperature change in the medium; they are electromagnetic in nature and travel at the speed of light. They are, for example, the physical mechanism that transfers energy (heat) from the sun to the earth through airless outer space.
• Blackbody radiant emissions are the primary reason that a fire feels hot at a distance. Because Kirchoff's Law states that a good emitter is also a good absorber for each wavelength, a blackbody may be defined as an ideal body that completely absorbs all radiant energy striking it, and therefore, appears perfectly black at all wavelengths. Emissivity may be defined as the ratio of an object's radiance to that emitted by a blackbody radiator at the same temperature and at the same wavelength. A perfect blackbody has an emissivity of one. Highly reflective surfaces have a low emissivity, but most materials that burn easily have emissivities of 0.5 or greater. The radiation emitted by a blackbody is referred to as Planck's Blackbody Radiation Law.[0119]
• FIGS. 13[0120]a,13b,14,15 and16 can be used to compare the amount of energy emitted at different energy bands by fires of different temperatures, and therefore the amount of energy that can be detected by sensors operating at different energy bands. FIGS. 13aand13bare tables containing data comparing the heat of a fire (in degrees Kelvin), its total radiant energy (in watts/cm²) over the visible, nearband IR and wideband IR ranges, its radiant energy in the narrowband IR range, and the relative percentages of narrowband versus the composite of nearband IR, visible and wideband IR. FIGS. 14, 15 and16 are graphs illustrating the information in the table of FIGS. 13aand13b. The table and graphs of FIGS. 13a,13b,14,15 and16 indicate that a fire emits far more energy in the wideband energy spectra than in the narrowband IR band.
• FIG. 14 is a graph showing total radiant energy as a function of a fire's temperature. As shown in FIG. 14, the total radiant energy generally increases as a function of the fire's temperature. FIGS. 15 and 16 compare in different aspects the amount of the fire's energy observable by a wideband detector versus a narrowband detector. FIG. 15 compares the percentage of radiant energy detected by a wideband detector and a narrowband detector as a function of the fire's temperature, while FIG. 16 is a graph showing a plot of the relative increase in energy detected by a wideband detector over a narrowband detector, as a function of the fire's temperature.[0121]
• The information appearing in FIGS. 13[0122]a,13b,14,15 and16 has been derived as follows. First, the formula for calculating the total heat radiation at all wavelengths from a perfect blackbody is known as the Stefan-Boltzmann Law:
• W=σT⁴  (1)
• where W is the total radiation emitted in watts/m[0123]², T is the absolute temperature in °K (degrees Kelvin), and σ is the Stefan-Boltzmann constant, 5.67×10⁻⁸watt/m²K⁴. The Stefan-Boltzmann Law indicates that the total radiant emitted energy from a surface is proportional to the fourth power of its absolute temperature; consequently, the hotter the body is, the greater the wide-band infrared radiation that is emitted. To obtain a more precise value of W, the total radiant blackbody energy emitted using equation (1), W can be multiplied by the average emissivity of the burning materials, which can be approximated by 0.5.
• Planck's Radiation Law may be used to calculate the continuous radiant energy distribution among the various wavelengths. For all the wavelengths from 0 to 100 microns, the radiated energy should be equal to the total radiated energy calculated by the Stefan-Boltzmann Law. The detected percentage of radiated energy is found by calculating the energy in the wavelength span covered by an optical fire detector with the total energy radiated by the fire using the Stefan-Boltzmann Law. Planck's formula for calculating the total radiated energy between first and second wavelengths λ[0124]1 and λ2 is as follows:
• W_λ1-λ2=∫((2πhc²)dλ)/λ⁵(e^hc/λ(kT)−1))  (2)
• where[0125]
• h=Planck's constant, 6.63×10[0126]⁻³⁴joule-sec.,
• c=speed of light, 3.00×10[0127]¹⁰cm/sec.,
• λ=wavelength in cm (10−2 meters),[0128]
• T=absolute temperature in degrees Kelvin, and[0129]
• k=Boltzmann constant, 1.38×10[0130]⁻²³joules/°K.
• Using equation (2) and integrating over the wavelength range from 0.4 to 3.5 microns, it can be determined that an “ideal” optical fire detector with a wide band spectral range of 0.4 to 3.5 microns is theoretically capable of sensing, for example, about 88.23% of the total radiated energy at a fire/flame temperature of 2500 degrees Kelvin (K) (2226.85 degrees Celsius), as appears in the information contained in the tables of FIGS. 13[0131]aand13b. For reference, the temperature of a typical clean burning flame generally varies from between 1400 and 3500 degrees K.
• There is another kind of infrared (IR) radiation that is discontinuous and made up of individual, very narrow emission lines. An example is the line spectrum produced by a heated gas, such as carbon dioxide, which is a by-product of hydrocarbon fires. This phenomenon is related to the excitation or heating of certain types of gases. The atoms or molecules of a gas have certain natural frequencies of vibration and rotation depending upon their structure, bonding forces, and masses. If certain gases are suitably excited (i.e., heated), they will emit a line emission spectrum characteristic of the particular gas. Such gases can also absorb radiation in the same line spectrum (i.e., according to application of Kirchoff's Law). For CO[0132]₂gas, these narrow band line emissions and absorption regions include the 4.3 micron band. Nonhydrocarbon fires, however, such as silane and hydrogen fires, do not produce CO₂gas as a by-product because no carbon atoms are involved in the combustion process.
• Thus, fires that are not oxidizing carbon based materials may not emit CO[0133]₂gas or have a narrow band line emission at 4.3 microns. The CO₂gas emission line from an uncontrolled fire is therefore unpredictable. Moreover, the total radiant energy of line emissions represents a very small fraction of the total blackbody radiant energy and does not measurably affect the total radiant energy output. Using Planck's Radiation Law, Equation (2), with a wavelength range from 4.2 to 4.4 microns to calculate the energy in the 4.3 micron band, the percentage of Planckian blackbody energy is about 0.82% at a fire/flame temperature of 2500 degrees K (see FIGS. 13aand13b). Thus, optical fire detectors that use the 4.3 micron narrow band to sense fires can fail to detect nonhydrocarbon fires, and do not present a proportional measure of energy consumed during a fire, as they often can see less than one percent of the fire's total radiated energy.
• Besides observing radiant energy over a given spectral band, a fire/flame detector can also observe the changes in radiant energy over time to make a better determination of whether a fire is occurring, as opposed to a non-fire event that may otherwise result in a false alarm trigger. Further characteristics of a fire can therefore be used to improve the detection ability of the fire/flame detector.[0134]
• For example, it has been observed by the inventors that the wide, continuous band of blackbody radiant energy pulsates as the fire's rising thermal energy causes the burning material(s) to further outgas, consuming more oxygen in rapid, irregular, exothermic chemical reactions. As the temperature of the fire further rises, the radiant blackbody energy correspondingly increases and the carbon particulates, if the fire is a hydrocarbon type (i.e., a fire involving hydrogen and carbon), remain after the other outgassing components are consumed. These hot carbon particles also radiate blackbody emissions and their emissivity is high.[0135]
• A calm, controlled fire, such as a candle burning in still air, radiates a constant blackbody radiant heat of the gaseous flame and particulates that can be felt within about one foot of the flame. In contrast, for an uncontrolled, unwanted fire, especially a growing fire, the radiant heat that is felt by the hand at a distance is pulsating and irregular. It has been observed by the inventors that most threatening fires tend to pulsate at a rate of approximately 2 to 100 Hertz. This flickering or pulsating causes ripples to occur at a similar rate in the detected blackbody energy.[0136]
• The differences between uncontrolled, unwanted fires and calm, controlled fires or non-fire energy sources can be advantageously used by the fire/flame detector to discriminate between fire situations which call for a response and situations which call for no response or merely continued monitoring. For example, the discriminator in the fire/flame detector may observe the frequency and regularity at which the detected radiant energy is pulsating, and thereby weed out potential false alarm situations. For example, if the fire/flame detector observes that the detected energy has a time-varying component between 2 and 100 Hertz, the discriminator may conclude that a potential unwanted fire situation exists. If, on the other hand, the fire/flame detector observes that the detected energy has no time-varying component, or has one or more time-varying components outside of the 2 to 10 Hertz frequency band, then the discriminator may conclude that it is unlikely that a fire situation exists. The discriminator may combine this information with other information to arrive at a final conclusion of whether a potential unwanted fire situation exists.[0137]
• The ultimate criteria used to determine whether a fire situation exists can largely depend upon the application in which the fire/flame detector is employed. The fire/flame detector is therefore advantageously configured with a programmable microprocessor so that the discrimination mechanism can be tailored to each particular environment.[0138]
• The discriminator can also be programmed with “false alarm profiles” to further assist in distinguishing between fire and non-fire situations. To accomplish this type of programming, the output of the fire/flame detector sensors is measured in response to various non-fire or non-dangerous flame sources anticipated to occur in the area where the fire/flame detector will be deployed. For example, the response of the fire/flame detector sensors to an oven, flashlight or a lit match may be measured and recorded. The energy profile resulting from the oven, flashlight, lit match or other non-dangerous fire source can be stored in a memory within the fire/flame detector. After the detector is deployed, and when a potential fire situation later occurs, the fire/flame detector can compare the current energy profile with the stored “false alarm” profiles and can prevent itself from declaring a fire situation if a close match is found.[0139]
• Referring now to FIGS. 3 and 5, sensor sensitivities and sensor types that are used in a fire/[0140]flame detector32 are illustrated. It should be understood that a variety of different sensors may be used in different configurations to accomplish the same or equivalent purpose. In accordance with one illustrated embodiment (FIG. 3), suitable silicon (Si) photodiode sensors are used for detecting radiant energy within the visible band and near band infrared spectrums. The wavelength (in nanometers) of the radiant energy is indicated along the x-axis and the sensor sensitivity in relative percentage is indicated on the y-axis. For a wide infrared spectrum, a suitable lead sulfide (PbS) sensor can be used. With reference specifically to FIG. 5, in accordance with an alternative embodiment, a Germanium photodiode sensor may be sandwiched on top of the lead sulfide (PbS) sensor.
• In accordance with the embodiment illustrated in FIG. 11, sensor digital data (once converted by the A/D converter) is continuously transmitted to the[0141]controller39. Thecontroller39 analyses the sensor digital data and determines if there is any sign of sparks, flames, or fire (whether or not visible to the human eye). Thecontroller39 therefore acts in this embodiment as a discriminator so as to discriminate between an unwanted and/or uncontrolled fire, and a non-fire or controlled fire source.
• To this end, the[0142]controller39 compares the radiant energy sensed by the visible band and nearband IR sensors against the radiant energy sensed by the wideband IR sensor. The radiant energy sensed by the visible band and nearband IR sensors can be from nonfire sources such as electrical lights, reflected and/or direct sunlight modulated by such things as tree branches or leaves in a breeze, reflected sunlight off water, steady-state IR sources such as infrared curing ovens, and the like. The radiant energy sensed by the wideband IR sensor is indicative of blackbody radiation. The relative levels of visible/nearband IR energy versus wideband IR energy, and the time relationship of those energy levels, can be analyzed by thecontroller39 in order to determine the presence or lack of presence of a fire.
• In a preferred embodiment, the radiant energy sensed by the visible band and nearband IR sensors (i.e., the A/D sampled outputs of the visible band and nearband IR sensors) is digitally subtracted from the radiant energy sensed by a wide band IR sensor (i.e., the A/D sampled output of the wideband IR sensor), resulting in a “compensated” measured energy level. The compensated measured energy level is compared against a predetermined threshold level. If the predetermined threshold level is exceeded, a possible fire situation is declared. The visible, nearband IR and wideband IR sensor outputs are then compared against false alarm profiles to verify that known false alarm sources have not caused the measured energy level to exceed the threshold.[0143]
• As described further hereinafter, multiple threshold levels for comparison may be established, with each threshold level resulting in a different response or action by the fire/flame detector.[0144]
• A preferred mechanism for discriminating between fires potentially requiring a responsive action and other radiant energy sources (including non-fires or certain types of controlled fire or energy sources) may be described with reference to FIGS. 28 through 34. FIG. 28 is a functional diagram illustrating the basic steps of the discrimination technique. While this technique is explained with reference to an embodiment which takes advantage of microprocessor processing speed and power, it will be understood that some or all of the functions described can be implemented, if desired, using analog circuitry or a combination of digital and analog circuitry.[0145]
• FIG. 34 shows a circuit for processing a sensor input signal and, more particularly, a circuit for processing an output from a wide band IR sensor[0146]502 (such as a lead sulfide (PbS) sensor) and generating afirst signal521 indicative of a DC level of the sensor output and asecond signal522 indicative of a transient level of the sensor output. The wideband IR sensor502 acts similar to a variable resistor, having a resistance that depends on the amount of radiant energy detected in the wide band IR range. The output of the wideband IR sensor502 is provided to anamplifier505 which is biased using a 2.5volt reference signal512 and a 5volt reference signal511 with suitable resistance values as shown in FIG. 3r. Theamplifier505 produces a first amplifiedsignal508 that is low pass filtered by the collective action of resistor R32 and capacitor C16, and then integrated and scaled usingamplifier507 to arrive at a wide band IRDC output signal521, designated MIR_DCherein.
• The first amplified[0147]signal508 is also high pass filtered using capacitor C14 and then amplified byamplifier506, which essentially acts as a buffer, to arrive at a wide band IRtransient output signal522, designated MIR, herein. Thecircuit501 of FIG. 34 thereby outputs both a wide band IRDC output signal521 and a wide band IRtransient output signal522. It will be understood that the wide band sensor is sometimes referred to in the text and the drawings as “WBIR” and at other times as “MIR”, and likewise with the output signal(s) from the wide band IR sensor; however, the designations WEIR and MIR are used interchangeably herein and are not intended to refer to different aspects of the described embodiments.
• A circuit similar to the circuit shown in FIG. 34 is provided for processing the output of the near band IR detector and producing two signals indicative of a DC level and transient level, respectively, of the near band IR detector output, except the component values of the elements of the circuit would be altered to match the characteristics of the near band IR sensor as may be readily accomplished by one skilled in the art. Likewise, a circuit similar to the circuit shown in FIG. 34 is provided for processing the output of the visible band detector and producing two signals indicative of a DC level and transient level, respectively, of the visible band detector output, except the component values of the elements of the circuit would be altered to match the characteristics of the visible band sensor.[0148]
• FIG. 28 shows a wide[0149]band IR sensor302 outputting a wide band IR DC signal and a wide band IR transient signal, a nearband IR sensor303 outputting an near band IR DC signal and a near band IR transient signal, and avisible band sensor304 outputting a visible band DC signal and a visible band IR signal. Each of the DC and transient signals for the wideband IR sensor302, nearband IR sensor303, andvisible band sensor304 is sampled and converted into the digital domain by A/D converters306,307 and308, respectively. (In a preferred embodiment, a single A/D converter is shared among all threesensors302,303 and304.)
• The A/[0150]D converters306,307 and308 output digitally sampled sensor signals312,313 and314, which are designated in FIG. 28 as MIR_DC, NIR_DC, and VIS_DC, respectively, each representing the “raw” DC component of the DC signal from the correspondingsensor302,303 or304. The A/D converters306,307 and308 also output digitally sampledsensor signals317 and318, which are designated in FIG. 28 as MIR_Tand VIS_T, respectively, each representing the “transient” component of the transient signal from correspondingsensor302 or304.
• The MIR[0151]_DCsignal312 is provided to a “fast”digital filter321 and to a “slow”digital filter322. Likewise, the NIR_DCsignal313 is provided to a “fast”digital filter323 and to a “slow”digital filter324, and the VIS_DCsignal314 is provided to a “fast”digital filter325 and to a “slow”digital filter326. In each case, the “fast” digital filter has a relatively fast time constant, and the “slow” digital filter has a relatively slow time constant. In each case, the slow digital filter outputs a slow filtered signal that (with adjustment or correction, in some cases) is used as a “baseline” for comparison against a fast filtered signal that is output from the fast digital filter. The “fast” digital filter and “slow” digital filter may each comprise a low pass filter, with the low pass filter of the “fast” digital filter having a higher cutoff frequency than that of the “slow” digital filter.
• [0152]Subtractors332,333 and334 operate to generate comparison signals between the fast filtered signal and the slow filtered signal for each of the digitized sensor signals, with the adjustments as noted below. FIG. 29 is a graph illustrating the output of any one of thesubtractors332,333 or334. At predetermined intervals of time (although it may be done continuously as well), the slow filtered signal is subtracted from the fast filtered signal to arrive at a difference signal designated as ΔMIR, ΔNIR and ΔVIS forsubtractors332,333 and334, respectively. In essence, the slow filtered signal represents a steady state or DC baseline for comparison, and the fast filtered signal represents a more rapidly occurring change in the radiant energy detected by the particular sensor. The difference signals ΔMIR, ΔNIR and ΔVIS therefore represent the change in radiant energy for each of the particular energy bands with respect to a measured baseline that varies gradually over time.
• In two situations the output of a slow or fast digital filter is adjusted prior to being applied to a substractor. First, a compensation is made in[0153]compensation function block328 to the output of the slowdigital filter322 of the wide band IR sensor input. Incompensation function block328, the output of the slowdigital filter322 is forced to the same value as the output of the fastdigital filter321 if the output of the fastdigital filter321 is less than the output of the slowdigital filter322. This effect is illustrated in FIG. 30, which is a graph showing a plot of a slowfilter output signal402 and a fastfilter output signal403. The dottedline portion404 indicates what the output of the slow filter would have been without the compensation provided by the compensation function block428, and is present where the output of the fastdigital filter321 is less than the output of the slowdigital filter322. The output of compensation function block428 is sent to thesubtractor332.
• One purpose of compensation function block[0154]428 is to ensure the output ofsubtractor332 will always be positive, which will ensure proper monitoring where a fire suddenly loses energy (such as where it is doused with fire suppressant) or else a source of wide band IR radiant energy is removed from the view of the WBIR (or MIR) sensor. Thus, the fire detector will be able to more easily determine if a temporarily suppressed fire is going to self-extinguish or is starting to regrow, in which case further treatment may be required.
• A second compensation is made in[0155]compensation function block329. This secondcompensation function block329 is connected to the output of the fastdigital filter325 associated with the visible band sensor input and can be used to reject false alarm sources such as flashlights.Compensation function block329 provides the equivalent of a peak detection function which holds the output of the fastdigital filter325 at its peak and thereafter allows it to steadily decline. Thecompensation function block329 may be realized as an asymmetrical digital filter with a fast rise time and slow decay time.
• FIG. 31 is a graph illustrating the effect of asymmetrical digital filtering on the fast digital filter output of the visible band sensor. In FIG. 31, a plot of the fast digital[0156]filter output signal411 is shown along with a plot of the slow digitalfilter output signal412. The dottedline portion413 indicates the output of the fast digital filter after asymmetrical filtering carried out by thecompensation function block329. As can be seen in FIG. 31, the signal after asymmetrical filtering decays more slowly than the actual output of the fast digital filter. By such asymmetrical filtering, the fire detector in essence creates an “artificial” visible light that lingers when an actual visible light is shut off. In this manner, false alarms caused by flashlights or other man-made electrical or battery light sources can in many instances be avoided. The output of the compensation function block329 (referred to as the “flashlight rejection” algorithm) is provided tosubtractor334.
• The output of subtractor[0157]332 (associated with the wide band IR sensor input signal) is connected to atemperature coefficient block341 which adjusts the difference signal ΔMIR by a coefficient to compensate for ambient temperature fluctuations. This compensation is made where, for example, a lead-sulfide (PbS) based wide band IR sensor is used, because such sensors are sensitive to temperature variations. The near band IR sensor and visible band sensor may be constructed of silicon, and therefore, on the other hand, would be largely temperature independent.
• The temperature compensated difference signal MΔ[0158]_Tis compared against the difference signal relating to the visible band sensor. However, because the visible band sensor input signal and wide band IR input signal are in terms of different units, a unit conversion adjustment is made to the visible band difference signal ΔVIS in unitconversion function block342. The temperature compensated difference signal MΔ_Tis thus compared by use of asubtractor345 against the unit converted visible band difference signal VΔ_C, to arrive at a compensated energy value MΔ_T′ which generally provides an indication of wide band radiant energy less visible radiant energy, with the adjustments for certain false alarm sources as noted above.
• The near band IR difference signal ΔNIR output from[0159]subtractor333 is used to set a threshold for comparison against the compensated energy value MΔ_T′. The output ofsubtractor333 is applied to a threshold value lookup table347 (which may be stored, for example, in non-volatile RAM) . The threshold value output from the threshold lookup table347 preferably changes as a step function of ΔNIR, with a change in ΔNIR leading to a proportional change (usually a fractional change) in the selected threshold value output from threshold value lookup table347.
• The compensated energy value MΔ[0160]_T′ is compared against the selected threshold value M_THby a subtractor (or comparator)350. Aresponse function block360 monitors the output of thesubtractor350 and declares a suitable early warning, alert or alarm condition in response to the compensated energy value MΔ_T′ exceeding the selected threshold value M_TH, indicating that a certain “dangerous” or potentially dangerous amount of radiant energy has been detected.
• The[0161]response function block360 preferably makes a response decision based not only on the output ofsubtractor350, but also on certain false alarm rejectors that are built in to the system. Two particular such false alarm rejectors are shown in FIG. 28. First, a flickerpulse count detector352 is utilized to provide an indication that the radiant energy source is flickering or pulsating in a “chaotic” manner typical of uncontrolled or growing fire sources. FIG. 32 is a diagram illustrating operation of a flicker detection algorithm in accordance with various aspects of the present invention. According to a preferred flicker detection algorithm, a “flicker” is detected by comparing positive transitions in the digitized transient wide band IR signal (MIRT)317 against positive transitions in the digitized transient visible band signal (VIS_T)318. Each positive transition of the digitized transient wideband IR signal317 is counted as a “flicker”, unless the digitized transientvisible band signal318 also has a positive transition at about the same time.
• If both of the digitized transient wide[0162]band IR signal317 and the digitized transientvisible band signal318 have a positive transition at about the same time (within a certain programmable tolerance), then the event is not deemed a “flicker,” regardless of how wide the respective pulses turn out to be. Thus, in FIG. 32, a first pulse (positive transition)432 in the digitized transient wideband IR signal317 is not counted as a valid flicker event, but a second pulse (positive transition)433 and a third pulse (positive transition)434 are counted as valid flicker events.
• According to the false alarm rejection technique embodied in the flicker[0163]pulse count detector352, a certain number (e.g., two) of valid flicker pulses must be counted in each consecutive time frame (e.g., two seconds) (which may, if desired, be a sliding time frame) or else the flickerpulse count detector352 outputs a value of “FALSE” indicating that the detected energy pattern does not appear to conform to an uncontrolled or growing fire situation. If, on the other hand, the flickerpulse count detector352 detects at least two valid flicker pulses in each consecutive time frame, and if it does so for a predetermined number of consecutive time frames (spanning, e.g., ten seconds), then the flicker detection algorithm is satisfied and the flickerpulse count detector352 outputs a value of “TRUE,” indicating that the detected energy pattern could be caused by an uncontrolled or growing fire.
• The above process performed by the flicker[0164]pulse count detector352 is known as flicker pulse count aging because it monitors the flickering nature of the detected radiant energy over time. If the flickering is not maintained for a sufficient amount of time (e.g., ten seconds), then the flickerpulse count detector352 provides an indication that the energy source may not be an actual fire for which a responsive action is advisable. In one aspect, the flickerpulse count detector352 may be viewed as applying a digital bandpass filter that outputs a “TRUE” value when valid flickers occur within a certain rate range.
• In addition to the flicker[0165]pulse count detector352, another false alarm rejector is a wide band IR peakenergy pulse detector353 that serves to weed out false alarm sources producing wide band IR energy of a steadily growing nature such as industrial ovens. As shown in FIG. 28, the digitized transient wide band IR signal (MIR,) is provided to a “fast”digital filter319 and a “slow”digital filter320, in a manner analogous to the digitized DC wide band IR signal (MIR_DC). The output of the “fast”digital filter319 and the “slow”digital filter320 are sent to the wide band IR peakenergy pulse detector353, which qualifies a fire by looking for a narrow peak in the transient wide band IR signal.
• FIG. 33 is a diagram illustrating a wideband peak energy pulse detection algorithm, and shows a wide band IR fast filter output signal[0166]461 (such as may be output from the “fast” digital filter319) plotted over time on a graph along with a wide band IR slow filter output signal462 (such as may be output from the “slow” digital filter320). FIG. 33 also shows a peak energy pulsedetection threshold level463 that varies over time such that it remains a preset amount above the slowfilter output signal462.
• Each time the wide band IR fast[0167]filter output signal461 exceeds the peak energy pulsedetection threshold level463, the amount of time it spends above the peak energy pulsedetection threshold level463 is measured to arrive at a peak pulse width Δt. If the peak pulse width Δt is “narrow”—i.e., less than a predefined narrowness criteria, then the peak is deemed a “valid” narrow peak. If a valid narrow peak has occurred within a predetermined amount of time (e.g., two seconds), then the wide band IR peak energy pulse detection algorithm has been satisfied, and the wide band IR peakenergy pulse detector353 outputs a “TRUE” value indicating that the detected wide band IR energy may conform to an uncontrolled or growing fire. Otherwise, the wide band IR peakenergy pulse detector353 outputs a “FALSE” value.
• The wide band IR peak[0168]energy pulse detector353 operates under the assumption that uncontrolled or growing fires tend to have rapid or narrow spikes of broadband radiant energy, and that such spikes of broadband radiant energy can be observed using the wide band IR sensor. Thus, wide band IR energy profiles that fail to display such characteristics are assumed to be related to non-fire or controlled fire sources. The wide band IR peakenergy pulse detector353 need not be used in all fire detection applications and may, for example, be employed only when the fire detector is placed in any environment having sources of wide band IR radiant energy such as industrial ovens.
• Referring again to FIG. 28, the[0169]response function block360 makes a response decision based on the output ofsubtractor350, as well as the flickerpulse count detector352 and the wide band IR peakenergy pulse detector353 if those false alarm rejectors are employed. If thesubtractor350 indicates that the compensated energy value MΔ_T′ exceeds the selected threshold value M_TH, then a certain amount of dangerous or potentially dangerous radiant energy has been detected. Theresponse function block360 then, if desired, uses the outputs of the flickerpulse count detector352 and the wide band IR peakenergy pulse detector353 to weed out false alarms. For example, theresponse function block360 may require that both the outputs of the flickerpulse count detector352 and the wide band IR peakenergy pulse detector353 be “TRUE” in order to declare a certain status level, such as an “ALARM” situation calling for an appropriate response.
• In addition, the fire detector may be provided with a second threshold level (M[0170]_TH2) against which the compensated energy value MΔ_T′ is compared, in order to support various types of multi-stage responses. In this manner, the fire detector can have a threshold value corresponding to different energy levels, such as 3 kW and 13 kW, for example. The second threshold level may be implemented by using a second threshold lookup table (which can be integrated with the first threshold lookup table) and a second subtractor connected to the second threshold level M_TH2and to the compensated energy value MΔ_T′.
• The[0171]response function block360 may or may not use the false alarm rejectors for arriving at a fire detection decision. For example, it may be desired to program the fire detector to respond rapidly to a 3 kW fire. In such a case, theresponse function block360 could issue an appropriate response by observing when the compensated energy value MΔ_T′ exceeds the threshold level of 3 kW, while disregarding the outputs provided by the flickerpulse count detector352 and the wide band IR peakenergy pulse detector353. Theresponse function block360 may be programmed to provide a different type of response to a 13 kW fire. In such a case, theresponse function block360 may be programmed to observe not only when the compensated energy value MΔ_T′ exceeds the second threshold level of 13 kW, but also to issue a response only when the outputs provided by the flickerpulse count detector352 and the wide band IR peakenergy pulse detector353 are also both TRUE.
• In addition, the fire detector could be provided to circuitry analogous to that shown in FIG. 28 for processsing the digitized transient wide[0172]band IR signal317 and comparing the detected transient wide band IR energy to a threshold level (such as 3 kW), after compensation for the transient visible band energy in a manner similar to that shown for the DC wide band IR and visible band signals in FIG. 28.
• In an alternative embodiment, mathematical techniques such as Fast Fourier Transforms (FFT's) are used to separate the temporal radiant energy spectral response of the WBIR, NBIR, and VB spectra into the individual Fourier components, thereby transforming the spectral radiant energy received as a function of time into a representation of radiant energy received as a function of frequency. By subtracting the individual frequency components of the nonfire sources from the individual frequency component of a real fire, a compensated energy level can be obtained, which is then used to eliminate potential false alarm sources as described above.[0173]
• In more detail, FFT's can be used to obtain individual Fourier components for each of the WBIR, NBIR and VB sensors at each a plurality of predetermined frequencies (such as 2, 5, 7 and 10 Hertz, for example). The magnitudes of WBIR frequency components are compared for each of the predetermined frequencies against the magnitudes of the VB frequency components, to arrive at a first set of energy level comparison values. Similarly, the magnitudes of WBIR frequency components are compared for each of the predetermined frequencies against the magnitudes of the NIR frequency components, to arrive at a second set of energy level comparison values. The first set of energy level comparison values and second set of energy comparison values may be applied to a lookup table to determine whether the profile matches that of a fire or potential fire situation.[0174]
• Use of a wideband IR sensor as the primary sensor realizes several benefits, particularly where use of a UV sensor is eliminated. First, the fire/flame detector may be housed in a self-contained, low-cost plastic, polypropylene, Teflon or fiberglass housing, whereas UV-based sensors require relatively expensive sapphire or quartz window lenses. Being able to use such a housing also eliminates problems found in UV-based sensors relating to sealing the UV sensor window. A similar advantage may be experienced over narrow band IR sensors. Because impurities in the housing material can have a significant impact on the quality of detection of a narrow band IR sensor, materials such as polypropylene, Teflon or plastic may be unacceptable for use as a housing and/or window with such narrow band IR sensors, but will, in contrast, have far less impact on a wide band IR sensor which does not rely on one narrow IR band or a few narrow bands for fire detection.[0175]
• As another significant benefit, a WBIR-based fire/flame detector according to the present invention can detect most any type of hydrocarbon (propane, butane, gasoline, etc.) and nonhydrocarbon-based fires (such as silane, hydrogen, sodium azide, etc.). In contrast, narrow band IR detectors typically only look for the CO[0176]₂spike emission line output centered at approximately 4.3 microns, which is best generated by a well-oxygenated, clean-burning, hydrocarbon fire (such as a propane flame). UV-based sensors are subject to “blindness” caused by cold CO₂suppressant gas and/or black smoke generated by a polypropylene fire. Such detectors can also be blinded by chemicals, acetones, vapors and gases in the atmospheres, or by contaminants that foul the window lens such as oil, paint residue (e.g., liquid and powder), dirt, etc. With WBIR as the primary sensor, the fire/flame detector can “see through” most contaminants because the longer infrared wavelengths of WBIR can penetrate them, with the exception of certain contaminants such as thick black paint, or thick layers of contaminants.
• In addition, a wide band IR sensor can have a significantly improved field of view over a narrow band IR sensor. Narrow band IR sensors require relatively precise filtering (such as may be carried out by a Fabry-Perot filter) tuned by the thickness of a material used in the sensor. When an energy source (such as a fire) is located at an angle to the narrow band IR sensor, the radiant energy emitted by the source strikes the sensor material at an angle and travels through a greater amount of the sensor material. This results in a distortion that makes the radiant energy appear to be located at a different wavelength than it actually is, and can cause a narrow band IR sensor to fail to detect a fire. Conventional narrow band IR sensors avoid this problem by maintaining a relatively tight field of view, such as 90°, for the sensor. A wide band IR sensor, on the other hand, can detect fires while having a 120° or even greater field of view.[0177]
• Using wide band IR can reduce maintenance problems because frequent, costly, manual window-lens cleaning of the UV lens can be eliminated. The flame/fire detector can also be made more rugged, reliable, and trouble-free without a UV sensor, which is typically made of UV transmitting glass or quartz.[0178]
• As an additional benefit to relying on wideband IR as the primary sensory input, the need for UV “through the lens” testing is eliminated. Each UV test source generally requires extra circuitry, which is usually high-voltage circuitry that can be susceptible to reliability problems. Such high-voltage circuitry is unnecessary where no UV sensor is used.[0179]
• It will be understood that while a preferred embodiment is described with respect to use of three sensors (i.e., a visible band sensor, a nearband IR sensor, and a wideband IR sensor), other sensor arrangements can be used to obtain the same or equivalent results. For example, a fire/flame detector may use a number of sensors each operating over a distinct narrow energy band range, and sum up the sensor outputs so as to obtain an indicia of total blackbody energy. While such a design would be more complicated due to the greater number of sensors, the same principles of fire/flame detection as previously described would apply to such a configuration.[0180]
• In accordance with a particular embodiment as illustrated in FIG. 11, sensor data is A/D converted and the resulting digital data is transmitted to the[0181]controller39. Thecontroller39 analyses the sensor digital data and determines if there is any sign of sparks, flames, or fire. Upon detecting an “alert,” a “fire early warning,” or an “alarm” condition, thecontroller39 selectively triggers one or more of three individual relays within an alarm unit56 (one-, two-, or three-stage). In accordance with one embodiment, a three-stage version of themulti-stage alarm unit56 comprises an “alert”relay58, a “fire early warning”relay60, and an “alarm”relay62. Alternatively, in accordance with another embodiment, a two-stage version of themulti-stage alarm unit56 comprises only the “alert”relay58 and the “alarm”relay62. Each of the relays may be coupled to distinctive LED indicators, audible alarms, or the like.
• In accordance with one approach, the[0182]controller39 compares the sensor digital data against programmed threshold values (of characteristics of fire signatures or false alarm models), to determine if the observed data indicates a cause for concern. Thecontroller39, upon detecting characteristics that warrant an “alert” condition, triggers the “alert” relay. Likewise, thecontroller39, upon detecting characteristics that warrant a “fire early warning” condition (in the three-stage embodiment) or an “alarm” condition, triggers either the “fire early warning” (in the three-stage embodiment)relay60 or the “alarm”relay62. The appropriate relay may in turn trigger an associated LED indicator or audible alarm. Atimer64 is set in every instance to either reject false alarm situations or allow the flame or fire sufficient time to self-extinguish. Only upon detecting an “alarm” condition, and that also after a predetermined time limit, are the suppression agents activated.
• In accordance with the general operation, the present system typically observes a fire in as little as 16 milliseconds (but can be less than one millisecond), then verifies the fire condition multiple times to ensure its existence. Following this exercise, the system (in the three-stage alarm embodiment) declares a “fire early warning” condition. For example, if the fire is a spray gun fire, the present system declares an “alert” condition to cause shutdown of the spray gun paint flow, electrostatics, and[0183]conveyor16. The present system continues to monitor the fire condition during a predetermined limit of time to allow it to self-extinguish. In the event the fire persists, the system declares an “alarm” condition and activates release of suppression agents to quell the fire.
• Alternatively, the system (in the two-stage alarm embodiment) looks for any sign of fire (small) and reports it so that personnel on the monitored facility can immediately respond to it. If the fire continues to grow, the system activates the “alarm” condition to activate release of the suppression agents to quell the fire.[0184]
• The following discussion relates to a preferred embodiment, in which multiple levels of responses are provided for different responsive actions in the optical fire/flame detector based upon both the type of fire and the radiant wide band continuous spectral output of the fire. There are different types of fires, including explosive fires; fast-burning, “fireball” fires; slow-burning, flickering fires; large, growing fires; etc. Different kinds of responsive actions may need to be taken depending upon the type of fire.[0185]
• In a particular embodiment, WBIR is used as the primary sensor in a multispectral sensor array of an optical fire/flame detector with digital signal processing in the electrostatic finishing industry. In this industry, spray guns are used to apply paint coatings in an electrostatic paint line or booth. Such spray guns apply approximately 100,000 volts to the atomized paint (e.g., liquid or powder). Because of the possibility of a malfunctioning spray gun or an improperly grounded part, arcing can occur, which can quickly ignite the paint mist, resulting in a “fireball”-type fire. The majority of the time, if the fireball is detected within one-half second and the paint flow and electrostatics are immediately shut down, the fire will self-extinguish. In a paint line or booth, paint mist accumulates on the detector's window lens, even with air shields. While this might blind a UV sensor, the preferred WBIR-based fire/flame detector is far less affected by accumulated paint.[0186]
• In another preferred embodiment, WBIR is used as the primary sensor in a multispectral array including NBIR and VB sensors and digital signal processing in semiconductor clean-room applications such as, e.g., chemical wet benches. An acid-proof, plastic housing can be used because WBIR (and NBIR and VB) can see through the plastic integral window, allowing all types and classes of hydrocarbon and nonhydrocarbon fires to be detected.[0187]
• By way of example, the following are different responsive actions for different types of fires: For an explosive type fire, the action taken could be to signal for the release of a high-speed water jet in several milliseconds in order to suppress the explosive fire. In a “fireball” fire, such as occurs, e.g., in a paint-booth spray-gun fire, the action taken is usually to shut off the paint flow and the electrostatic high-voltage supply to the spray gun, which will usually self-extinguish the fireball. For a slow-burning, flickering fire, such as a solvent rag burning in a paint spray booth or a solvent fire in a polypropylene chemical wet bench, the action may be to warn the operator with lights and/or sound annunciators only, as a suppression release would be premature and could damage costly product. For a large, growing fire, the action may be to warn the operator with annunciators and signal for a release of suppression agent. Also, different types of suppression agent may be released depending upon the type of fire.[0188]
• Another type of multilevel response to a fire is illustrated by the following example. Suppose a small, solvent rag ignites spontaneously in an automatic paint spray booth. The preferred, multilevel-response optical fire/flame detector detects the slow-burning, small flickering fire and signals a first-stage-level response (such as “Fire Early Warning” stage). The first-level response signals strobe lights and audio annunciators to warn the operator. No suppression agent would be released at the first level, as the fire might self-extinguish or be extinguished manually with a portable extinguisher, thereby averting a costly cleanup and process shutdown. Also, as discussed below, the preferred detector can record the spectral history of the first-level fire to aid in diagnosing the cause of the fire, especially if it should self-extinguish without being discovered.[0189]
• If the fire were to ignite the paint overspray and generate a fireball fire, the second-stage level would be declared (“Alert” stage), which could be to shut down the paint flow and electrostatics to the spray guns. If the fireball did not self-extinguish, and instead ignited the paint residue on the booth floor and began growing into a larger, dangerous fire, the preferred, multilevel optical fire detector would sense the radiant output (using the WBIR sensors) of the fire, and when the fire exceeded a certain energy criteria, would signal for release of suppression agents (“Alarm” stage).[0190]
• The following is another example of a multilevel response to a fire that is possible in a preferred embodiment. Suppose a solvent chemical leaks inside a semiconductor process wet bench and is ignited by an electrical spark. The preferred, multilevel optical fire detector detects the small, flickering fire, and when it grows to certain criteria (such as, e.g., 3 kW) energy level, a first-stage response level is advantageously declared. The first-level response signals strobe lights and audio annunciators to warn the operator that a hidden fire is occurring inside the wet bench. Also, the detector preferably digitally records the spectral history of the first-level fire, as discussed in greater detail below, to aid in diagnosing the cause of the fire, especially if it should self-extinguish without being discovered. This can assist the operator in future fire prevention.[0191]
• When a first-level response is declared, the operator has the option of manually extinguishing the fire and/or completing the process, or waiting to see if the fire will self-extinguish because of the limited supply of solvent and/or the high air flow “blowing out” the fire without a costly suppression release. In this regard, it has been estimated that about 80% of wet-bench fires self-extinguish. Allowing the fire a chance to self-extinguish can therefore save many thousands of dollars by avoiding the release of fire suppressants and the consequent negative effects on the materials at the manufacturing site (such as computer chips fabricated on large wet-benches).[0192]
• If the fire continues to burn, consuming more solvent fuel, and rises in temperature to ignite the polypropylene bench material, the preferred multilevel detector will signal a second-stage response when the fire's energy level reaches a preset level (such as, e.g., a 13 kW energy level threshold), which will cause activation of lights and/or sound to alert the operator, and the second-stage response will signal for a suppression release. The preferred WBIR detector is capable of “seeing through” the suppression agent and continuing to signal for suppression until the fire is extinguished. Should the fire “reflash” later, the WEIR detector can again respond.[0193]
• The fire detector may make response level declarations based on radiant energy output and fire signal characteristics, in addition to using temporal information such as the relative rate of growth or periodic fluctuations of the energy level in each of the observed spectral bands. Once the fire has been detected, spectral data from the fire (including the time period just before the fire detection) can be digitally recorded and analyzed at a later time.[0194]
• In accordance with another feature of the present invention, the microcomputer (otherwise referred to as controller or microprocessor)[0195]36 and the controller verify proper operation of each other, and upon detecting any sign of failure, trigger thefault relay66.
• In a preferred embodiment, a real-time graphical display of the digital sensor data detected by the[0196]flame detector32 is generated and viewed at a “SnapShot™”display68. The digital sensor data is represented in the form of relative spectral intensities versus present time. The “Snapshot” display is preferably viewed with an IBM compatible personal computer (with an RS-232 interface port). An associated memory (RAM)68amay store a particular display.
• A “FirePic™”[0197]generator70 facilitates retrieval of sensor spectral data stored prior to an occurrence of fire. A graphical display of relative spectral intensities versus time preceding the fire provides evidence to enable analysis and determine the true cause of the fire. The “FirePic” data may be stored, for example, in anon-volatile RAM72. As indicated in FIG. 18, the “FirePic” data may indicate a “FirePic” number and data such as the date, the time, the temperature, the MIR_{(DC and T)}, NIR_{(DC and T)}, and VIS_{(DC and T)}readings of sensor signal data, the input voltage, and the control switch settings. FIGS. 19, 20, and21 indicate “alert,” “fire early warning,” and “alarm” events relating to exemplary fire signatures. FIG. 19arepresents an exemplary fire signature that would trigger the “alert”relay58. FIG. 20arepresents an exemplary fire signature that would trigger the “fire early warning”relay60. FIG. 21arepresents an exemplary fire signature that would trigger an “alarm”relay62. A printout of a graphical display (from the “FirePic” generator, or the “SnapShot” display, or of fire signatures) may be obtained with aprinter76.
• In more detail, the spectral data of the optical fire/flame detector—i.e., the digitally converted output values from the visible band (VIS), near band IR (NIR) and wideband IR (WBIR or MIR) sensors—can be digitally recorded (with or without other digital processing information such as parameter settings, etc.) immediately prior to a fire and/or during a fire. The digitized sensor data is maintained in a circular buffer, so that the most recent sensor data over a predefined time period (e.g., eight seconds) is held at any given time. When a detection event occurs, the[0198]controller39 determines that a recording will be made, at which time the data stored in the circular buffer is transferred to a location in a nonvolatile memory.
• The type of nonvolatile memory may comprise, for example, a CMOS RAM backed with a lithium battery and shutdown logic, or an SRAM combined with a “shadow” EEPROM (electrically erasable programmable read only memory). Such nonvolatile RAM memory can offer years of storage life in the absence of external power. In addition to storing sensor data in the nonvolatile memory, detection events and parameters may also be recorded, including warning status level. The stored data may then be used for post-fire event analysis. Such data can be used to help determine the cause of the fire and measures can be taken to ensure that the fire does not recur in the future.[0199]
• While an EEPROM-based system may be used, use of other types of nonvolatile memory provides several advantages over an EEPROM based system. While it is possible to continuously write to volatile RAM during a fire event and, after the fire event is over, store the data in the EEPROM, this is a relatively time consuming process (taking several seconds). Should the fire event cause the electrical power to be interrupted or reset, the stored sensor data is likely to be lost before transfer to the EEPROM is complete. Also, if after a fire event, an operator uses a match, butane lighter, or handheld tester to make sure the detector is working properly, the real fire data in volatile RAM can be overwritten with useless test data.[0200]
• Thus, in a preferred embodiment, spectral data of the fire/flame detector is stored in nonvolatile, battery-backed RAM, which records substantially faster (i.e., in a matter of milliseconds) than an EEPROM. A large-capacity nonvolatile RAM is preferably used, so that multiple fire events can be stored, each with a predefined amount of data (e.g., eight seconds of data). Moreover, the time and date of each event can be stored, thereby enabling discrimination between real fire data and test data after a fire event. Preferably, added to each data package is more signal processing data including parameters such as, for example, and indication of which warning, alert or alarm stages were declared in the multilevel-response optical fire detector.[0201]
• The[0202]controller39 initially and routinely after preselected periods of time, such as every ten minutes, performs diagnostic evaluation or tests on select system components, such as checking for continuity through the relay coils, checking to ensure that the control settings are as desired, and so on. Upon detecting some cause for concern, the diagnostic test relating to the area of concern may be performed every thirty seconds or any such preselected period of time. It should be understood that any or all the parameters including reaction times, etc., may be programmed to address particular requirements. A digital serial communication circuit69 (see FIGS. 11 and 12) controls serial connections of one or more of a plurality offlame detectors32 to thecontroller39 to ensure clear communication through the otherwise noisy environment.
• Referring now to FIG. 12, in accordance with an alternative embodiment of the present system, the[0203]microprocessor36 located within theflame detector32 itself processes all the sensor digital data to determine the nature of the prevailing condition and triggers an appropriate one of the multistage (e.g., two- or three-stage)alarm unit56. In this embodiment only the “SnapShot”display68 and its associatedmemory68ais located external to thedetector component32. The digital communicationserial circuit68 controls serial connections of one or more of a plurality offlame detectors32 to any peripheral devices such as theprinter76, “SnapShot”display60, etc.
• The system performs extensive diagnostic evaluations, the logic of which will now be considered with reference to FIG.[0204]22. A start of the diagnostic evaluation operations is indicated byreference numeral80. To ensure that the “alarm” relays are functioning properly, current is passed through each of the relay coils, as illustrated by ablock82. Continuity of current through the relay coils is determined as illustrated by ablock84.
• The diagnostic evaluation proceeds to check the control settings for the various system components. The step is illustrated by a[0205]block86. The control settings are compared against stored data on control settings desired by a user, as indicated by adecision block88. If the control settings are as desired, the diagnostic evaluation operation proceeds to the next step. If the control settings are not as desired, they are initialized in accordance with the stored data, as indicated by ablock90. The next step in the diagnostic evaluation is a test to determine communication between the detector unit and the controller unit. This step is illustrated by aquery block92. In the event the communications are satisfactory, operation proceeds to a step illustrated by ablock94 that indicates that the system is ready to commence its detection operations. In the event the communications are not satisfactory, steps to correct any existing problems may be taken, as indicated by ablock96. After the communication problems are corrected or solved, operation returns to thequery block92 until communications between the detector and the controller are found to be satisfactory.
• With reference to FIG. 23, a lens test may if desired be performed by the system to ensure its optimum performance, in those embodiments where the housing includes a lens (such as[0206]lens100 shown in FIG. 4) as part of theviewing window132bon the face of thedetector32. Such a lens test, as noted previously, is not ordinarily necessary in embodiments of the invention where no UV sensor is utilized. However, should a lens test be desired, the start of the lens test is indicated at104 in FIG. 23. Light from an infrared LED or any other infrared source (not shown) is transmitted from within thedetector32 through thelens100 of thedetector32. This step is illustrated by ablock106. The intensity of light reflected back by thedetector grill102 is measured to determine the transmittance level of thedetector lens100. This step is illustrated by ablock108. Preset threshold intensity values (of transmittance) provided by the lens manufacturer are stored as indicated by ablock110. As illustrated by adecision block112, the measured intensity values are compared against the stored values to determine if there is any degradation in transmittance characteristics or levels. In the event the measured intensity values are greater than the stored values, the system proceeds to the next step indicated by ablock114. At that point, the overall system operation for detection can commence. In the event the measured intensity values are less than the stored values, indicating degraded transmittance characteristics, operation proceeds to the next step indicated by ablock116.
• The measured intensity values are registered in memory and a number is assigned to each registered value. A[0207]decision block118 determines if the number of intensity values registered exceeds the number ten. If the answer is affirmative, the test proceeds to one of two options. Under option one, in the event there are multiple detectors, a fault condition is relayed to an external computer, as illustrated by ablock120. Under option two a fault condition is declared and an alarm is sounded, as illustrated by a block122. If the answer to decision block118 is negative, as illustrated by ablock124, the lens test is repeated every thirty seconds.
• In addition to the above tests, an additional diagnostic test could also consist of moisture detection to verify the seal integrity of a housing. If a moisture detection test is desired, a moisture detector is preferably located within the housing and provides an output for use by the[0208]controller39 in determining whether the housing seal has been damaged or is leaking.
• With reference to FIGS. 24, 25, and[0209]26, the logic for the overall system operation for detection is described. Once the system is installed at a desired facility, prior to operation of the system the control settings for the various system components are programmed. Referring now to FIG. 24, a start is indicated at ablock150. The system of the present invention runs diagnostic evaluations, such as those described above, at the very outset and repeatedly during system operation to ensure proper functioning of all its system components. This step is illustrated by ablock152. Following the diagnostic evaluations, the parameters for the system components are adjusted in accordance with ambient conditions (e.g., ambient temperature adjustments, ambient light adjustments). Ablock154 illustrates this step. Sensor signals from each of the sensors MIR_{(DC and T)}, NIR_{(DC and T)}, and VIS_{(DC and T)}are received from thesensor array38, as illustrated by ablock156. At this point operations split into two paths, indicated as a path1 (for detecting an “alert” condition) and a path2 (for detecting “fire early warning” and “alarm” conditions).
• To determine an “alert” condition, transient sensor signals MIRT, NIRT, and VIST are monitored, as illustrated by a[0210]block158. The MIRT signal value is compensated by subtracting from it the VIST signal value, as illustrated by ablock159.
• Referring now to FIG. 25, the compensated MIRT signal value is compared against predetermined threshold values (one or more as desired), as indicated by a[0211]decision block160. If the compensated MIRT signal value exceeds the predetermined threshold value, an “alert” timer is set as indicated by ablock162. In the next step, illustrated by adecision block163, the system determines if a predetermined time limit has passed. Once the predetermined time limit is passed, an “alert” condition is declared, as illustrated by ablock164. Following that step, another predetermined period of time is enforced or allowed to pass, during which no action is taken, in order to allow the fire to self-extinguish. This step is illustrated by ablock165. After that point, operation loops back to point A, whereby the system again receives sensor signals. Of course, until the predetermined time limit actually expires, operation loops back to the point beforedecision block163.
• Referring again to FIG. 24, to determine a “fire early warning” condition or an “alarm” condition, sensor signals MIR[0212]_DC, NIR_DC, and VIS_DCare passed through long-term and short-term averaging filters as illustrated by ablock166. These signals are monitored to obtain values as illustrated by ablock168. To eliminate false alarm rejection, in the event the short-term filter output values are less than the long-term filter output values, as illustrated by adecision block170, the long-term filter output values are jam set (forced) to adopt the short-term filter output values, as illustrated by ablock172.
• Referring now to FIG. 26, the sensor signal MIR[0213]_DCreading is compensated by the sensor signals NIR_DCand visible_DCreadings, as illustrated by ablock174. This step is taken to distinguish a real fire from other sources more likely to emit substantial visible light. Once the MIR_DCsignal is compensated to eliminate declaring a false alarm, the MIR_DCsignal value is compared against programmed parameters, as indicated by adecision block176. In the event the MIR_DCsignal value is determined to be less than the programmed parameters, operation loops back to point A, beginning the cycle of receiving the sensor signals from thesensor array38, and so on.
• In the event the MIR[0214]_DCsignal value is greater than the programmed parameters, adecision block178 determines if the variations in the MIR_DCsignal values are significant. If it is determined that the variations in the MIR_DCsignal values are not significant, as illustrated by ablock180, the system ensures that the “fire early warning” and “alarm” timers are set to zero. Following that, operation once again loops back to point A.
• If it is determined that the variations in MIR[0215]_DCsignal values are significant, the timers for the “fire early warning” and the “alarm” are set to begin counting. This step is illustrated by ablock182. If the “fire early warning” timer indicates that a predetermined time limit has passed, as indicated by adecision block184, a “fire early warning” condition is declared, as illustrated by ablock186. Once the “fire early warning” condition is declared, the appropriate relay is activated as illustrated by ablock188. At that point, operation may ultimately loop back to point A. Of course, until the predetermined time limit has expired, operation loops back to the point beforedecision block184.
• Once a “fire early warning” is declared, as illustrated by a[0216]decision block190, the system determines if the “alarm” timer indicates that a predetermined time limit has passed. If not, operation loops back to the point beforedecision block190, to ensure that the appropriate time limit has passed. If the “alarm” timer indicates that a predetermined time limit has passed, as indicated by adecision block190, a “alarm” condition is declared, as illustrated by ablock192. Once the “alarm” condition is declared, the appropriate relay is activated, as illustrated by ablock194. At that point, operation ultimately may loop back to point A.
• Referring now to FIG. 27, starting at point C, operation of the system in accordance with its two-stage alarm embodiment compares the MIR[0217]_DCsignal value against a first predetermined threshold value, as indicated bydecision block198. The first predetermined threshold value corresponds to a “small” fire of a size considered to be a hazard. If the MIR_DCsignal value is less than the first predetermined threshold value, operation loops back to point A, where the system continues to read signal values. If the MIR_DCsignal value exceeds the first predetermined threshold value (stored in memory), the system declares a “pre-alarm” (or “alert”) condition as indicated by ablock200. Subsequently or immediately, the system activates the appropriate “alert” relay, as indicated by ablock201, enabling personnel at the monitored facility to investigate the fire, and ceases all ambient considerations. The system continues to monitor for a rise in the fire, as indicated by ablock202. This may be done by reading optical radiation amounts emitted by the fire. It should be recognized that other ways of monitoring a rise in fire known to those skilled in the art may alternatively be used. As indicated bydecision block204, the system compares the optical radiation amounts emitted by the fire against a second predetermined threshold amount (stored in memory). If the optical radiation amount emitted by the fire exceeds the second threshold amount, the system declares an “alarm” condition as indicated by ablock206, and activates the appropriate relay and suppression agents, as indicated by ablock208.
• It should be recognized that the system could compare readings against more than two energy thresholds. The energy thresholds can be empirically determined by performing fire tests of the sizes and at distance desired by the monitored facility. Also, the specific thresholds used may vary depending on the choice of sensor, amplification of signals, etc. For example, in a clean room environment with chemicals, an alcohol fire having a four-inch diameter viewed at a distance of eight feet has an energy output of 3 kilowatts (kW) and may be predetermined as the first threshold. Similarly, a fire having an eight-inch diameter viewed from a distance of eight feet has an energy output of 13 kW and may be predetermined as the second threshold.[0218]
• While some embodiments (such as the fire detection system of FIG. 11) have been described with a controller external to the flame detector, it will be appreciated that the functionality of the controller can be located within the flame detector. Having an external controller can be efficient where multiple detectors are deployed in the same locality, so that the multiple detectors can share the same controller and therefore be implemented with reduced cost. In some situations it may be preferable for each fire detector to have its own controller, and to have all of the controller electronics along with the sensor electronics enclosed within a self-contained unit.[0219]
• Certain alternative embodiments involve use of different types of sensors for the primary wide band IR (WBIR) sensor. In a preferred embodiment, the WBIR sensor is manufactured from lead sulfide (PbS), which has a high sensitivity to wide band IR over the necessary operational temperature ranges, in addition to having relatively low cost, proven reliability, and ready availability. Other sensors can also be used to sense wide band IR in certain fire detection applications.[0220]
• There are two main classes of practical sensors that can be used for wide band IR fire detection. The first class includes photon detectors, which have time constants typically in the microsecond range, and the second class includes thermal detectors, which have time constants typically in the millisecond range.[0221]
• Photon detectors may include any of a number of quantum photodetectors such as photoconductive (or photoresistive) detectors or photovoltaic detectors. A photoconductive (or photoresistive) sensor is one in which a change in the number of incident photons causes a fluctuation in the number of free charge carriers in the semiconductor material. The electrical conductivity of the responsive element is inversely proportional the number of photons. This change is conductivity is monitored and amplified electrically. A photovoltaic sensor is one in which a change in the number of photons incident on a p-n junction causes fluctuations in the voltage generated by the junction. This change in voltage is monitored and amplified electrically.[0222]
• Photon detectors include sensors made from material(s) in the lead salt family such as lead sulfide (PbS), lead telluride (PbTe), lead selenide (PbSe), lead tin telluride (PbSnTe), the doped germanium family (Ge:AuSb, Ge:Cu, Ge:Hg, Ge:Cd, Ge:Zn, Ge-Si:Zn, Ge-Si:Au, etc.), indium antimonide (InSb), indium arsenide (InAs), telluride (Te), mercury cadium telluride (HgCdTe), and other such materials.[0223]
• Thermal detectors and sensors that can be used for wide band IR sensing include both pyroelectric sensor types and thermopile sensor types. Pyroelectric sensor types include such sensors as deuterated triglycine sulfate (DTGS or TGS), lithium tantalate (LiTAO[0224]₃), barium titanate (BaTiO₃), and the like. Pyroelectric sensors are thermal sensors and use a crystal which develops a charge on opposite crystal faces (similar to a capacitor) when incident radiation causes the crystal temperature to change. Thermopile sensor types are those in which thermovoltaic and generate a voltage when thermal radiation strikes their surface. Usually thermopile sensors are manufactured as a small matrix array. The way they operate is by sensing an output voltage across a junction of dissimilar metals. When the temperature of the junction fluctuates because of changes in the level of incident radiation, the output voltage generated by the junction will fluctuate. This voltage is monitored and amplified electrically.
• For any of the above wide band IR sensors, to set the wavelength cutoffs for the desired wide band IR spectral range, appropriate interference type or absorption type filters, or a combination thereof, may be used.[0225]
• Further information of interest may be found in U.S. Pat. application Ser. No. 08/866,024 (attorney docket no. 226/036) entitled “Fire Detector With Multi-Level Response,” U.S. Pat. application Ser. No. 08/865,695 (attorney docket no. 226/037) entitled “Fire Detector With Event Recordation,” U.S. Pat. application Ser. No. 08/866,023 (attorney docket no. 226/038) entitled “Fire Detector and Housing,” and U.S. Pat. application Ser. No. 08/866,028 (attorney docket no. 226/039) entitled “Fire Detector With Replaceable Module,” each of which applications is filed concurrently herewith, and each of which is incorporated by reference herein as if set forth fully herein.[0226]
• While the present invention has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations will be apparent to those skilled in the art in view of the foregoing description. Accordingly, the invention is intended to embrace all such alternatives, modifications, and variations that fall within the spirit and scope of any appended claims.[0227]

Claims (24)

What is claimed is:

1. A fire detector, comprising:

a wide band infrared sensor;

means for detecting when a level of wide band infrared radiant energy detected by the wide band infrared sensor exceeds an energy level threshold; and

means responsive to said detecting means for initiating a fire response protocol.

2. The fire detector ofclaim 1, wherein said means for detecting when a level of wide band infrared radiant energy detected by the wide band infrared sensor exceeds an energy level threshold comprises a filter connected to the wide band infrared sensor for measuring the level of wide band infrared radiant energy, and a comparator connected to an output of said filter and to said energy level threshold.

3. The fire detector ofclaim 2, wherein said means for detecting when a level of wide band infrared radiant energy detected by the wide band infrared sensor exceeds an energy level threshold further comprises a visible band sensor connected to a second filter for measuring a level of visible band radiant energy, and a subtractor whereby said level of visible band radiant energy is subtracted from the output of the filter connected to the wide band infrared sensor prior to said comparator.

4. The fire detector ofclaim 1, further comprising a near band infrared sensor, wherein said energy level threshold is set as a function of a level of energy detected by the near band infrared sensor.

5. A fire detection system, comprising:

a sensor array for providing sensor signals, said sensor array including a false-alarm sensor for detecting visible band energy, a false-alarm sensor for detecting near band infrared energy, and a primary sensor for detecting wide band infrared energy;

a temperature sensor for providing signals indicative of ambient temperature conditions; and

a controller for processing said sensor signals, said controller coupled to said sensor array and said temperature sensor for receiving said sensor signals from said sensor array and said signals from said temperature sensor, said controller calibrating operation parameters for said fire detection system in accordance with said signals from said temperature sensor and said sensor signals.

6. The fire detection system ofclaim 5, further comprising a multistage alarm system coupled to said controller, said controller selectively activating a select one of a plurality of relays of said multistage alarm system depending upon one of a plurality of conditions sensed by said controller.

7. The fire detection system ofclaim 6, wherein said plurality of relays comprises two relays and said plurality of conditions comprises two conditions.

8. The fire detection system ofclaim 6, wherein said plurality of relays comprises three relays and said plurality of conditions comprises three conditions.

9. The fire detection system ofclaim 5, further comprising means for storing spectral information associated with a detected fire event.

10. The fire detection system ofclaim 9, wherein said means for storing comprises nonvolatile RAM.

11. The fire detection system ofclaim 5, further comprising means for storing spectral information preceding a detected fire event.

12. The fire detection system ofclaim 11, wherein said means for storing comprises nonvolatile RAM.

13. The fire detection system ofclaim 5, further comprising an enclosed, sealed, removable housing containing said sensor array, said temperature sensor, and said controller, said housing including a window lens.

14. The fire detection system ofclaim 13, wherein said window lens is integral with said housing.

15. The fire detection system ofclaim 14, wherein said window lens is thinner than the remaining part of said housing.

16. The fire detection system ofclaim 13, wherein said window lens is embedded in said housing.

17. The fire detection system ofclaim 13, wherein said housing is heat sealed.

18. The fire detection system ofclaim 17, further comprising a plastic-coated, electrical power cable, said housing being heat sealed to said cable.

19. A fire detection system, comprising:

a sensor array including a primary fire sensor for detecting wide band infrared energy, a first false-alarm sensor for detecting near band infrared energy, and a second false-alarm sensor for detecting visible band energy;

a temperature sensor;

a microprocessor configured to receive signals from said sensor array and said temperature sensor, said microprocessor calibrating operation parameters for said fire detection system in accordance with said signals received from said sensor array and said temperature sensor;

a multistage alarm system coupled to said microprocessor, said microprocessor selectively activating a select one of a plurality of relays of said multistage alarm system depending upon one of a plurality of conditions sensed by said microprocessor;

an enclosed, sealed, removable housing containing said sensor array, said temperature sensor, and said microprocessor, said housing including a window lens integral thereto; and

a nonvolatile storage medium for storing spectral data associated with a detected fire event, said nonvolatile storage medium coupled to said microprocessor.

20. A method of detecting a spark, flame, or fire, comprising the steps of:

sensing wide band infrared energy, near band infrared energy, and visible band energy;

eliminating false-alarm, nonfires based upon the sensed near band infrared energy and visible band energy; and

detecting a spark, flame, or fire based upon the sensed wide band infrared energy.

21. The method ofclaim 20, further comprising the step of storing the detected energy.

22. The method ofclaim 20, further comprising the step of varying the response to the results of said detecting step based upon the results of said detecting step.

23. The method ofclaim 22, wherein said varying step comprises first and second sequentially performed steps, said first step comprising initiating an alarm signal when the results of said detecting step exceed a first predetermined energy threshold, said second step comprising releasing a suppressing agent when the results of said detecting step exceed a second predetermined energy threshold, said second threshold greater in magnitude than said first threshold.

24. The method ofclaim 22, wherein said varying step comprises first, second, and third sequentially performed steps, said first step comprising initiating an alarm signal when the results of said detecting step exceed a first predetermined energy threshold, said second step comprising shutting off electric power when the results of said detecting step indicate that a fire remains burning for a predetermined time after performance of said initiating step, said third step comprising releasing a suppressing agent when the results of said detecting step exceed a second predetermined energy threshold, said second threshold greater in magnitude than said first threshold.

Images