CN102175202A - Ice cover thickness predicting method based on fuzzy logic - Google Patents

Abstract

Translated fromChinese

本发明公开的一种基于模糊逻辑的覆冰厚度预测方法，首先获取覆冰数据：环境温度、环境湿度、环境风速及导线温度；其次建立覆冰厚度预测模型；最后得到覆冰厚度。本发明基于模糊逻辑的覆冰厚度预测方法，适合对受多种因素影响的具有不确定性结论的事物或现象作出总的评价。此方法根据覆冰现场的环境温度、环境湿度、环境风速以及导线温度的数据，得出输电线路覆冰厚度。数据来源于覆冰在线监测系统实时监测的现场覆冰数据，从而解决了现有的覆冰预测模型现场数据来源缺乏、精确度不够的问题。

The fuzzy logic-based ice thickness prediction method disclosed by the present invention firstly acquires ice data: ambient temperature, ambient humidity, ambient wind speed and conductor temperature; secondly establishes an ice thickness prediction model; finally obtains the ice thickness. The ice thickness prediction method based on fuzzy logic of the present invention is suitable for making a general evaluation on things or phenomena with uncertain conclusions affected by various factors. This method obtains the icing thickness of the transmission line according to the ambient temperature, ambient humidity, ambient wind speed and conductor temperature data of the icing site. The data comes from the on-site icing data monitored by the icing online monitoring system in real time, thus solving the problems of lack of on-site data sources and insufficient accuracy of the existing icing prediction models.

Description

A kind of ice covering thickness Forecasting Methodology based on fuzzy logic

Technical field

The invention belongs to the powerline ice-covering monitoring technical field, be specifically related to a kind of ice covering thickness Forecasting Methodology based on fuzzy logic.

Background technology

Powerline ice-covering regular meeting causes line insulator ice sudden strain of a muscle accident, inhomogeneous icing or do not deice accident the same period, and the overload accident, ice coating wire is waved accident.Once serious wire icing of transmission line accident can cause enormous economic loss, and has a strong impact on social life.

Though existing achievement in research has related to the various relations between icing formation and the meteorological condition, but existing ice covering thickness forecast model all sums up by wind tunnel test, its valid data from the icing field monitoring are considerably less, so it is further perfect that existing ice covering thickness forecast model needs, could allow achievement in research ripe more and accurate.

Summary of the invention

The purpose of this invention is to provide a kind of ice covering thickness Forecasting Methodology, solved the problem that existing icing forecast model field data source lacks, degree of accuracy is not enough based on fuzzy logic.

The technical solution adopted in the present invention is, a kind of ice covering thickness Forecasting Methodology based on fuzzy logic is specifically implemented according to following steps:

Step 1: obtain the icing data: environment temperature, ambient humidity, ambient wind velocity and conductor temperature;

Step 2: the numerical value of environment temperature, ambient humidity, ambient wind velocity and the conductor temperature that obtains according tostep 1, set up the ice covering thickness forecast model;

Step 3: the ice covering thickness forecast model according tostep 2 obtains, calculate ice covering thickness.

Characteristics of the present invention also are,

The environment temperature in thestep 1 wherein, ambient humidity, ambient wind velocity and conductor temperature, gather by the icing on-line monitoring system, the structure of icing on-line monitoring system is: comprise MSP430F247, be connected with system power supply on the MSP430F247 respectively, liquid crystal display and clock module, icing data acquisition and processing module, data storage cell and communication module, communication module comprises Zigbee communication module and GPRS communication module, system power supply is connected with controller, controller is also respectively at sun power, accumulator is connected, icing data acquisition and processing module comprise the icing information process unit, the input end of icing information process unit respectively with Temperature Humidity Sensor, air velocity transducer and temperature sensor are connected.

Whereinstep 2 is set up the ice covering thickness forecast model, specifically implements according to following steps:

A. Fuzzy processing obtains the membership function of variable;

B. establish fuzzy rule;

C. set up the fuzzy prediction model.

Step a Fuzzy processing wherein, specifically implement: adopt four inputs, one export structure according to following steps, four input variables are consistent with output variable to be divided into five fuzzy subset: NB: very low/little, NS: lower/little, O: medium, PS: higher/big and PB: very high/big, the membership function of each variable adopts triangular function.

Wherein step b establishes fuzzy rule, specifically implement: environment temperature, ambient humidity, ambient wind velocity, conductor temperature and ice covering thickness are carried out statistical study conclude according to following steps, to every rule definition intensity G (k), be that the degree of membership u (k) that each data of composition rule belong to its fuzzy subset multiplies each other, k is the sequence number of rule, as shown in the formula

G(k)＝u(k)_ET×u(k)_EH×u(k)_EW×u(k)_CT.

Wherein, G (k) expression intensity; U (k)_ETThe size of expression environment temperature degree of membership; U (k)_EHThe size of expression ambient humidity degree of membership; U (k)_EWThe size of expression ambient wind velocity degree of membership; U (k)_CTThe size of expression conductor temperature degree of membership; The rule that falls into contradictions occurs, and then according to its intensity size, by going to stay big principle to accept or reject for a short time, finally establishes fuzzy rule.

Wherein step c sets up the fuzzy prediction model, specifically implement according to following steps: the fuzzy rule that the input variable that obtains according tostep 1, the membership function that step a obtains and step b obtain, set up the ice covering thickness forecast model by the Fuzzy logic fuzzy logic toolbox among the MATLAB.

The invention has the beneficial effects as follows that the fuzzy logic method that model adopts has the characteristics of multifactor analysis-by-synthesis, be fit to the things with uncertain conclusion or the phenomenon that are subjected to multiple factor affecting are made total evaluation.The method needs environment temperature, ambient humidity, ambient wind velocity and the conductor temperature at icing scene, the on-the-spot icing data that these are monitored in real time data from the icing on-line monitoring system, thus draw electric power line ice-covering thickness.

Description of drawings

Fig. 1 is the process flow diagram of the inventive method;

Fig. 2 is the structural representation of the icing on-line monitoring system that adopts in the inventive method;

Fig. 3 is the membership function of input variable environment temperature among the embodiment;

Fig. 4 is the membership function of input variable ambient humidity among the embodiment;

Fig. 5 is the membership function of input variable ambient wind velocity among the embodiment;

Fig. 6 is the membership function of input variable conductor temperature among the embodiment;

Fig. 7 is the membership function of output variable ice covering thickness among the embodiment.

Among the figure, 1.MSP430F247,2. system power supply, 3. liquid crystal display and clock module, 4. icing information process unit, 5. Temperature Humidity Sensor, 6. air velocity transducer, 7. humidity sensor, 8. sun power, 9. controller, 10. accumulator, 11. data are deposited defeated unit, 12.Zigbee communication module, 13.GPRS communication module.

Embodiment

The present invention is described in detail below in conjunction with the drawings and specific embodiments.

The present invention is based on the ice covering thickness Forecasting Methodology of fuzzy logic, as shown in Figure 1, specifically implement according to following steps:

Step 1: obtain the icing data, adopt the powerline ice-covering on-line monitoring system, these 4 icing influence factors of the environment temperature when monitoring icing in real time, ambient humidity, ambient wind velocity and conductor temperature are as the input variable of ice covering thickness forecast model.The structure of icing on-line monitoring system as shown in Figure 2, comprise MSP430F2471, be connected withsystem power supply 2 on the MSP430F2471 respectively, liquid crystal display and clock module 3, icing data acquisition and processing module, data storage cell 11 and communication module, communication module comprises Zigbeecommunication module 12 and GPRS communication module 13,system power supply 2 is connected with controller 9, controller 9 is also respectively atsun power 8,accumulator 10 is connected, icing data acquisition and processing module comprise icinginformation process unit 4, the input end of icinginformation process unit 4 respectively withTemperature Humidity Sensor 5,air velocity transducer 6 and temperature sensor 7 are connected.

A shaft tower monitoring unit is installed on overhead line structures, is utilizedsun power 8 andaccumulator 10 charging work, can realize round-the-clock monitoring extra high voltage line and environmental parameter.Monitoring unit real-time Monitoring Line microclimate condition and circuit icing situation, monitoring information is sent to Surveillance center by GPRS communication module 13.

Step 2: set up the ice covering thickness forecast model of fuzzy logic, specifically implement according to following steps:

A. Fuzzy processing: obfuscation is meant input is converted to fuzzy set, be about to survey the fuzzy subset that physical quantity is converted into different language value in the corresponding domain of this linguistic variable, a plurality of inputs for fuzzy logic model, the fuzzification process of each input quantity all is the same, and the prerequisite of carrying out fuzzy reasoning is to import all to pass through Fuzzy processing.Icing predictive fuzzy logical model adopts four inputs, one export structure.Based on fuzzy theory, determine the fuzzy set of each variable.In order to obtain the higher forecasting precision, four input variables and consistent five fuzzy subset: NB (very low/little), NS (lower/little), O (medium), PS (higher/big) and a PB (very high/big) of being divided into of output variable.Based on the icing database that has obtained is added up and existing experience, the membership function of each variable all adopts triangular function among the present invention.

B. establish fuzzy rule: fuzzy rule is the core of fuzzy model, and it is equivalent to the correction module or the compensating module of fuzzy model.The generation method of fuzzy rule has two kinds substantially: a kind of is the practical experience and the knowledge of the field long-term accumulation institute study or relates to according to expert or operating personnel, concludes summary and draws; Another kind is to analyze in the data to existing input-output, concludes to sum up to draw.Because of at present very few to the research of icing fuzzy analysis, expertise lacks, so take second method that a large amount of icing data (comprising environment temperature, ambient humidity, ambient wind velocity, conductor temperature) and the ice covering thickness that the icing on-line monitoring system is obtained carried out the statistical study conclusion.

In many fuzzy rules that constitute, may the fuzzy rule conflict can appear owing to reasons such as Monitoring Data errors, and promptly the fuzzy subset of some regular former piece (input variable) is the same, and the fuzzy subset of consequent (output variable) is different.For the fuzzy rule to contradiction screens choice, to every rule definition intensity G (k), promptly each data of composition rule degree of membership u (k) of belonging to its fuzzy subset multiplies each other, and k be regular sequence number, suc as formula (1):

G(k)＝u(k)_ET×u(k)_EH×u(k)_EW×u(k)_CT (1)

Wherein, G (k) expression intensity; U (k)_ETThe size of expression environment temperature degree of membership; U (k)_EHThe size of expression ambient humidity degree of membership; U (k)_EWThe size of expression ambient wind velocity degree of membership; U (k)_CTThe size of expression conductor temperature degree of membership;

Principle is accepted or rejected in screening: the rule that falls into contradictions occurs, and then according to its intensity size, decides what to use by " going to stay for a short time big " principle.Finally can establish fuzzy rule.

C. set up the fuzzy prediction model: obtain the fuzzy rule that its subordinate function and step b obtain according to the resulting input variable ofstep 1, step a, set up the ice covering thickness forecast model by the Fuzzy logic fuzzy logic toolbox among the MATLAB.

Step 3: the ice covering thickness forecast model is according to four parameters of input, obtain the output variable ice covering thickness: the ice covering thickness forecast model is set up, environment temperature, ambient humidity, ambient wind velocity and the conductor temperature of the transmission line of electricity that real-time monitors be input in the model can obtain output quantity, i.e. ice covering thickness.This draws process and comes analysis and judgement to go out according to input value through fuzzy rule, finishes by means of MATLAB software.

Fuzzy logic method is fit to the things with uncertain conclusion or the phenomenon that are subjected to multiple factor affecting are made total evaluation, and powerline ice-covering has tangible ambiguity and uncertainty, simplify the design of ice covering thickness forecast model with fuzzy logic, go to describe input, rule and output with natural language, its result more meets people's requirement, more near the form of thinking of people's intuitivism apprehension.

Embodiment

The field data of utilizing the icing on-line monitoring system to obtain, the on-the-spot icing Monitoring Data (500kV pacifies an expensive loop line, 220kV chicken sun two times, the triumphant beautiful line of 220kV, the indiscriminate two wires of 110kV, 220kV Suo Gan two loop lines, 220kV copper multitude line, 110kV soil Yang Song thatch line and 220kV and practises duck one loop line) of having collected 8 transmission lines of electricity of Guizhou electrical network.

Press the fuzzy logic model analytical procedure, at first these circuits are carried out comprehensive statistics from the icing data in year January in Dec, 2008 to 2009, the variation range that draws data such as environment temperature ET, ambient humidity EH, ambient wind velocity EW, conductor temperature CT and ice covering thickness IT is respectively :-8 ℃～17 ℃, 32%～99%, 0m/s～8.6m/s ,-12 ℃～16 ℃ and 0mm～23.39mm.The membership function of 4 input variables such as Fig. 3, Fig. 4, Fig. 5, shown in Figure 6 and output variable are as shown in Figure 7.Because the icing environment constantly changes and icing data monitoring frequency is 1 time/15min, so the data variation scope is not interior variation of scope that necessarily is confined to above statistics, in order to make data area can contain various icing situations and to be convenient to the obfuscation analysis, the variation range of above environment temperature ET, ambient humidity EH, ambient wind velocity EW, conductor temperature CT and ice covering thickness IT suitably enlarged be adjusted into-20 ℃～20 ℃, 0%～100%, 0m/s～20m/s ,-20 ℃～20 ℃ and 0mm～30mm.Carrying out statistical induction at the icing data of these 8 transmission lines of electricity sums up, draw 78 initial fuzzy rules in conjunction with expertise, accept or reject principle by screening, 25 rules have finally been obtained, fuzzy rule adopts ifET is ... andEH is ... and EW is ... and CT is ... then IT is ... this fuzzy language is described, and it is as shown in table 1 to list the part rule:

Table 1 fuzzy reasoning table

When four factors belong to different fuzzy subset, judge the fuzzy subset of ice covering thickness one by one.After fuzzy rule is established, just can be by means of setting up the ice covering thickness forecast model in the Fuzzy logic fuzzy logic toolbox among the MATLAB.

Claims (6)

Translated fromChinese

1.一种基于模糊逻辑的覆冰厚度预测方法，其特征在于，具体按照以下步骤实施：1. A fuzzy-logic-based method for predicting ice thickness, characterized in that it is specifically implemented in accordance with the following steps:步骤1：获取覆冰数据：环境温度、环境湿度、环境风速及导线温度；Step 1: Obtain icing data: ambient temperature, ambient humidity, ambient wind speed and wire temperature;步骤2：根据步骤1获取的环境温度、环境湿度、环境风速及导线温度的数值，建立覆冰厚度预测模型；Step 2: According to the values of ambient temperature, ambient humidity, ambient wind speed and wire temperature obtained in step 1, establish an ice thickness prediction model;步骤3：根据步骤2得到的覆冰厚度预测模型，计算覆冰厚度。Step 3: According to the ice thickness prediction model obtained in step 2, calculate the ice thickness.2.根据权利要求1所述的基于模糊逻辑的覆冰厚度预测方法，其特征在于，所述步骤1中的环境温度、环境湿度、环境风速及导线温度，是通过覆冰在线监测系统采集的，覆冰在线监测系统的结构为：包括MSP430F247(1)，MSP430F247(1)上分别连接有系统电源(2)、液晶显示与时钟模块(3)、覆冰数据采集与处理模块、数据存储单元(11)及通信模块，通信模块包括Zigbee通信模块(12)及GPRS通信模块(13)，所述的系统电源(2)和控制器(9)相连接，控制器(9)还分别于太阳能(8)、蓄电池(10)相连接，所述的覆冰数据采集与处理模块包括覆冰信息处理单元(4)，覆冰信息处理单元(4)的输入端分别与温湿度传感器(5)、风速传感器(6)及温度传感器(7)相连接。2. The icing thickness prediction method based on fuzzy logic according to claim 1, wherein the ambient temperature, ambient humidity, ambient wind speed and wire temperature in the step 1 are collected by an icing online monitoring system , the structure of the icing online monitoring system is: including MSP430F247 (1), the MSP430F247 (1) is respectively connected with the system power supply (2), liquid crystal display and clock module (3), icing data acquisition and processing module, data storage unit (11) and communication module, communication module comprises Zigbee communication module (12) and GPRS communication module (13), described system power supply (2) is connected with controller (9), and controller (9) is also connected with solar energy respectively (8), accumulator (10) is connected, and described icing data acquisition and processing module comprises icing information processing unit (4), and the input terminal of icing information processing unit (4) is connected with temperature and humidity sensor (5) respectively , wind speed sensor (6) and temperature sensor (7) are connected.3.根据权利要求1所述的基于模糊逻辑的覆冰厚度预测方法，其特征在于，所述步骤2建立覆冰厚度预测模型，具体按照以下步骤实施：3. The method for predicting ice thickness based on fuzzy logic according to claim 1, wherein said step 2 establishes an ice thickness prediction model, which is specifically implemented according to the following steps:a.模糊化处理，得到变量的隶属度函数；a. Fuzzy processing to obtain the membership function of the variable;b.确立模糊规则；b. Establish fuzzy rules;c.建立模糊预测模型。c. Establish fuzzy prediction model.4.根据权利要求3所述的基于模糊逻辑的覆冰厚度预测方法，其特征在于，所述步骤a模糊化处理，具体按照以下步骤实施：采用四输入一输出结构，四个输入变量和一个输出变量一致分为五个模糊子集：NB：很低/小、NS：较低/小、O：中等、PS：较高/大以及PB：很高/大，各个变量的隶属度函数采用三角形函数。4. The icing thickness prediction method based on fuzzy logic according to claim 3, characterized in that, the fuzzification process of said step a is specifically implemented according to the following steps: adopting a four-input-output structure, four input variables and one The output variables are consistently divided into five fuzzy subsets: NB: very low/small, NS: low/small, O: medium, PS: high/large and PB: very high/large, and the membership function of each variable adopts Triangle function.5.根据权利要求3所述的基于模糊逻辑的覆冰厚度预测方法，其特征在于，所述步骤b确立模糊规则，具体按照以下步骤实施：对环境温度、环境湿度、环境风速、导线温度及覆冰厚度进行统计分析归纳，对每条规则定义一个强度G(k)，即构成规则的每个数据属于其模糊子集的隶属度u(k)相乘，k为规则的序号，如下式，5. The icing thickness prediction method based on fuzzy logic according to claim 3, characterized in that, said step b establishes fuzzy rules, specifically implemented according to the following steps: ambient temperature, ambient humidity, ambient wind speed, conductor temperature and The ice thickness is statistically analyzed and summarized, and a strength G(k) is defined for each rule, that is, the membership degree u(k) of each data that constitutes the rule belongs to its fuzzy subset is multiplied, and k is the serial number of the rule, as shown in the following formula ,G(k)＝u(k)_ET×u(k)_EH×u(k)_EW×u(k)_CT.G(k)＝u(k)_ET ×u(k)_EH ×u(k)_EW ×u(k)_CT .其中，G(k)表示强度；u(k)_ET表示环境温度隶属度的大小；u(k)_EH表示环境湿度隶属度的大小；u(k)_EW表示环境风速隶属度的大小；u(k)_CT表示导线温度隶属度的大小；遇到矛盾规则出现，则根据其强度大小，按去小留大原则进行取舍，最终确立模糊规则。Among them, G(k) represents intensity; u(k)_ET represents the size of the membership degree of ambient temperature; u(k)_EH represents the size of the membership degree of environmental humidity; u(k)_EW represents the size of the membership degree of environmental wind speed; u( k)_CT represents the degree of membership of the wire temperature; when conflicting rules appear, the fuzzy rule is finally established according to the principle of eliminating the small and retaining the large according to its strength.6.根据权利要求3所述的基于模糊逻辑的覆冰厚度预测方法，其特征在于，所述步骤c建立模糊预测模型，具体按照以下步骤实施：根据步骤1得到的输入变量、步骤a得到的隶属度函数及步骤b得到的模糊规则，通过MATLAB中的Fuzzy logic模糊逻辑工具箱建立覆冰厚度预测模型。6. The icing thickness prediction method based on fuzzy logic according to claim 3, characterized in that, said step c establishes a fuzzy prediction model, which is specifically implemented according to the following steps: according to the input variable obtained in step 1, obtained in step a The membership function and the fuzzy rules obtained in step b are used to establish the ice thickness prediction model through the Fuzzy logic toolbox in MATLAB.

Images