Disclosure of Invention
In order to solve the problems, the application provides a building water supply and drainage flow speed abnormality monitoring method based on the internet of things technology, which adopts the following technical scheme:
collecting water flow velocity data in a pipeline of a building water supply and drainage system;
Obtaining monitoring anomaly of the water flow velocity values according to distribution conditions in an observation window of each water flow velocity value and clustering results of differential elements among the water flow velocity values;
Sequencing according to the monitoring anomaly degree of the water flow velocity values to obtain a characteristic value sequence of the water flow velocity values, dividing the characteristic value sequence into a plurality of initial characteristic segments, and calculating the anomaly probability of the initial characteristic segments based on element distribution characteristics and position information in the initial characteristic segments;
Obtaining a neighborhood K value of the water flow velocity value according to the monitoring anomaly degree of the water flow velocity value and the anomaly probability of the initial characteristic section to which the water flow velocity value belongs, and obtaining the abnormal water flow velocity value by utilizing an anomaly detection algorithm according to the neighborhood K value of the water flow velocity value.
Further, the monitoring anomaly degree of the water flow velocity value is obtained according to the distribution condition in the observation window of each water flow velocity value and the clustering result of the difference elements between the water flow velocity values, and the method comprises the following specific steps:
Forming flow velocity sequences of all collected water flow velocity values according to time sequence, obtaining differential sequences of the flow velocity sequences, and dividing elements in the differential sequences into a plurality of clustering clusters by using a clustering algorithm without presetting the number of the clustering clusters;
setting a sliding window with a fixed size, and taking the sliding window when each water flow velocity value is the center point of the window as an observation window of each water flow velocity value;
The monitoring anomaly of each water flow velocity value is calculated based on the distribution condition of the elements in the observation window of each water flow velocity value and the distribution characteristics of the elements in the cluster.
Further, the calculating the monitoring anomaly of each water flow velocity value based on the distribution condition of the elements in the observation window of each water flow velocity value and the distribution characteristics of the elements in the cluster comprises the following specific steps:
Determining the number of clusters where the differential elements formed by all the water flow velocity values in the observation window of each water flow velocity value are located, and calculating the product of the number of the clusters where the differential elements are located and the distribution variance of all the elements in the observation window of each water flow velocity value as a molecule;
taking the inverse of the sum of the frequencies of the differential elements formed by all the water flow velocity values in the observation window of each water flow velocity value in the differential sequence of the flow velocity sequence and the sum of the constant parameters as a first metric value;
the product of the first metric value and the molecule is taken as the monitoring anomaly of each water flow velocity value.
Further, the calculating the abnormal probability of the initial feature segment based on the element distribution feature and the position information in the initial feature segment includes the following specific steps:
For any initial characteristic section, determining coordinate information of a monitoring position of each monitoring anomaly corresponding to a water flow velocity value, and marking the coordinate information as a label value of the monitoring position of a water supply and drainage pipeline turning part in a building as 1;
calculating the product of the distribution variance of the monitored abnormal degree in each initial characteristic section and the average value of the monitored abnormal degree in each initial characteristic section, and taking the ratio of the product to the number of the monitored abnormal degrees in each initial characteristic section as the abnormal weight of each initial characteristic section;
The anomaly probability of each initial characteristic segment is determined based on the anomaly weight, the probability that the difference in the trend of the monitored anomaly degree in each initial characteristic segment and the reference characteristic segment of each initial characteristic segment is generated by the abnormal water flow velocity value.
Further, the determination manner of the anomaly probability of each initial feature segment is as follows:
Calculating Euclidean distance between monitoring position coordinate information corresponding to any two monitoring anomalies in the two initial feature segments, taking the average value of all Euclidean distances calculated between the two initial feature segments as the monitoring position distance between the two initial feature segments, and taking other initial feature segments corresponding to the minimum value of the monitoring position distance between each initial feature segment as reference feature segments of each initial feature segment;
Determining the difference value of the monitored anomaly quantity in each initial characteristic section and the reference characteristic section of each initial characteristic section;
performing straight line fitting on each initial characteristic segment and the monitored abnormal degree in the reference characteristic segment of each initial characteristic segment to obtain the sum of the absolute value of the difference value between the slopes of the straight lines and the difference value of the quantity as a first difference value;
Taking the product of the first difference value and the abnormal weight of each initial characteristic section as a molecule, and taking the ratio of the molecule to the number of monitoring positions with the label value of 1 in the monitoring positions corresponding to the monitoring abnormality degree in the reference characteristic section of each initial characteristic section as the abnormal probability of each initial characteristic section.
Further, the method for obtaining the neighborhood K value of the water flow velocity value according to the monitoring anomaly degree of the water flow velocity value and the anomaly probability of the initial characteristic section to which the water flow velocity value belongs comprises the following specific steps:
screening the monitoring anomaly degree based on the label value of the monitoring position of each water flow velocity value, and determining the distribution influence weight of the data level of each anomaly degree based on the screening result;
Determining a neighborhood distance of each water flow velocity value based on the distribution influence weight of the data level to which each monitoring anomaly belongs and the anomaly probability of the initial characteristic section to which each monitoring anomaly belongs;
if the neighborhood distance of the water flow velocity value is smaller than the preset upper limit value, taking the neighborhood distance of the water flow velocity value as a neighborhood K value of the water flow velocity value;
and if the neighborhood distance of the water flow velocity value is greater than or equal to the preset upper limit value, taking the preset upper limit value as a neighborhood K value of the water flow velocity value.
Further, the distribution of the data level of each anomaly degree affects the weight determination manner as follows:
Removing the monitoring anomaly degree with the label value of 1 at the corresponding monitoring position in all the monitoring anomaly degrees, forming a screening set by using the residual monitoring anomaly degrees, and taking the information entropy of the monitoring anomaly degree in the screening set as a first threshold value;
In the screening set, taking the monitored outliers with equal values as data levels of one outlier;
For any abnormal data level, the difference between the first threshold and the information entropy of the residual monitoring abnormal data in the screening set after deleting the abnormal data level is used as the distribution influence weight of the abnormal data level.
Further, the determining the neighborhood distance of each water flow velocity value comprises the following specific steps:
Calculating the ratio of the abnormal probability of the initial characteristic section to which the monitoring abnormality degree of each water flow velocity value belongs to the maximum value in the abnormal probability of all the initial characteristic sections;
and taking the product of the ratio and the distribution influence weight of the data level to which the monitoring anomaly of each water flow velocity value belongs as a neighborhood distance of each water flow velocity value.
Further, the dividing the characteristic value sequence into a plurality of initial characteristic segments comprises the following specific steps:
and taking a characteristic value sequence of the water flow velocity value as input, and dividing the characteristic value sequence into a plurality of subsequences by using a sequence dividing algorithm, wherein each subsequence is taken as an initial characteristic segment.
Further, the method for obtaining the abnormal water flow velocity value by using the abnormality detection algorithm according to the neighborhood K value of the water flow velocity value comprises the following specific steps:
taking all the water flow velocity values and the neighborhood K value of the water flow velocity values as input, and acquiring local abnormal factors of each water flow velocity value by using an outlier factor detection LOF algorithm;
a second threshold value is preset, and if the local abnormality factor of the water flow velocity value is greater than the second threshold value, the water flow velocity value is considered to be an abnormal water flow velocity value.
The technical scheme has the advantages that after water flow velocity data in a pipeline of a water supply and drainage system are acquired according to the Internet of things technology, the water flow velocity values change in different time periods, firstly, a clustering method which does not need to be preset for the number of clusters is utilized to divide a differential sequence formed by the water flow velocity values into a plurality of clusters, the problem that the difference between the preset number of clusters and the actually acquired data distribution influences the characteristic analysis accuracy of subsequent data is avoided, secondly, the observation window of each water flow velocity value in the preset water flow velocity data is utilized to analyze the difference of the water flow velocity values in the observation window and the clustering condition of differential elements between the water flow velocity values in the observation window, the monitoring abnormal degree of the water flow velocity values is determined, the probability that the water flow velocity values are abnormal values is estimated initially is estimated, the distribution difference of the monitoring abnormal degree is caused by the water flow velocity values of the pipeline structure in the water supply and drainage system, the abnormal degree of the water flow velocity is determined, the abnormal probability of each initial characteristic segment is eliminated, the water flow velocity in the sequence is higher due to the fact that the pipeline structure is different from the water flow velocity values, the water flow velocity values in the water flow velocity values are estimated to be detected automatically, the different from the neighborhood values in the water flow velocity values is estimated to be different from the abnormal flow velocity values, the abnormal flow velocity values in the water flow velocity values, the water flow velocity values are estimated to be detected automatically in the different neighborhood values, the water flow velocity values are estimated to be detected by the abnormal values, and the abnormal flow velocity values in the water flow velocity values, the water flow velocity values are estimated to be detected by the abnormal values in the water flow velocity values, the accuracy of detecting the abnormal water flow velocity value by using the LOF algorithm is improved.
Detailed Description
In order to further explain the technical means and effects adopted by the application to achieve the preset aim, the following detailed description is given below of the method for monitoring abnormal flow rate of water supply and drainage of a building based on the internet of things technology according to the application, which is provided by the application, with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
The specific scheme of the building water supply and drainage flow speed abnormality monitoring method based on the Internet of things technology provided by the application is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for monitoring abnormality of water supply and drainage flow rate of a building based on internet of things according to an embodiment of the present application is shown, where the method includes the following steps:
And S001, collecting flow velocity data of water flow in a pipeline in the building water supply and drainage system.
It should be noted that, the purpose of this embodiment is to perform real-time anomaly monitoring on the flow velocity data of the water in the pipeline of the water supply and drainage system of the building, and before starting the processing analysis of the data, the data needs to be collected first.
Specifically, water flow speed monitoring devices are arranged in pipelines of a plurality of building water supply and drainage systems, and one water flow speed value of each monitoring position is output every 5s by using the water flow speed monitoring devices. Wherein the water flow rate monitoring device comprises, but is not limited to, a turbine type water flow sensor, an ultrasonic water flow sensor and a water flow rate measuring instrument. The operator can select proper monitoring equipment according to the pipeline size of the building water supply and drainage system, the larger the pipeline size is, the more suitable for the monitoring equipment with large size is, the smaller the pipeline size is, the monitoring equipment with small size is selected, and the installation is convenient.
Step S002, obtaining the monitoring anomaly degree of the water flow velocity values according to the distribution condition in the observation window of each water flow velocity value and the clustering result of the difference elements between the water flow velocity values.
It should be noted that, when the LOF algorithm is used to perform anomaly analysis on the water flow rate data, the fluctuation degree of the water flow rate data in different time periods is different, which is related to the temperature in the pipeline and the supply and use conditions of the water resource, and the change of the water flow rate may cause a certain damage to the pipeline, such as "water hammer effect", i.e. a certain impact is caused to the pipeline by the abrupt change of the water flow rate in the pipeline. Therefore, the water flow velocity value with abnormal change needs to be monitored in time, and the problems of pipeline breakage and the like caused by the change of the water flow velocity are avoided.
It should be further noted that, when the abnormal monitoring is performed on the water flow velocity value, the fluctuation conditions of the water flow velocity data in different time periods need to be determined, so as to obtain the initial range of the neighborhood value K, and the magnitude of the K value has a larger influence on the local density of the water flow velocity value, so that the accuracy of the subsequent abnormal monitoring is affected. The fluctuation of the water flow velocity value in a short time has a certain similarity, so that a characteristic value sequence is obtained according to the monitoring anomaly of the water flow velocity value, then the characteristic value sequence is segmented according to the change of the characteristic value sequence, and then the segmented data are analyzed to obtain the self-adaptive K value of each water flow velocity value.
Specifically, the step of acquiring an observation window of each water flow velocity value in the water flow velocity data according to a preset numerical value is as follows:
Firstly, setting a sliding window with a size of 1*T, taking the sliding window when each water flow velocity value is a window center point as an observation window of each water flow velocity value, and for the water flow velocity value of which the sliding window exceeds the range of the collected water flow velocity data when the sliding window is taken as the window center point, interpolating filling data of the out-of-range part by using a quadratic linear interpolation method. The linear interpolation is a common technique in the field of data processing, and the specific process is not described again.
The sliding window is set to analyze the local data characteristics of the water flow velocity value by using the difference between adjacent data of the water flow velocity value with a small time interval. Thus, the magnitude of T is set to the empirical value 7 in the present embodiment, and in other embodiments, the practitioner may set to an appropriate value provided that the local data characteristics of observing each water flow velocity value can be satisfied, which the present application is not particularly limited to.
Secondly, all collected water flow velocity values form a flow velocity sequence according to a time sequence, a differential sequence of the flow velocity sequence is obtained, elements in the differential sequence are divided into a plurality of clusters by utilizing AP (Affinity Propagation) clustering algorithm, the change characteristics of the water flow velocity values are integrally reflected through intra-class differences and frequencies of the elements in the clusters, and the AP clustering algorithm is a common technology in the field of data processing and is not repeated in a specific process.
It should be noted that, the AP clustering algorithm is a clustering algorithm that does not need to set the number of clusters in advance, so that elements in the differential sequence can be divided into different clusters according to the actual difference between the collected water flow velocity values, and the situation that the number of clusters is set in advance and the actual collected data distribution is different is avoided.
It should be further noted that, in this embodiment, only one method for clustering elements in the differential sequence is provided, that is, by using an AP clustering algorithm, on the premise that element clustering in the differential sequence can be achieved and the number of clusters is not preset, in other embodiments, an implementer may use other clustering algorithms, which is not limited in particular by the present application.
Further, the monitoring anomaly of each water flow velocity value is calculated by combining the observation window of each water flow velocity value and the distribution characteristics of the elements in the cluster, so as to represent the probability that each water flow velocity value is anomaly data. The calculation formula of the monitoring anomaly degree of the ith water flow velocity value is as follows:
In the formula,Monitoring anomalies for the ith water flow rate value,Is the number of clusters in which the differential elements formed by all the water flow velocity values in the observation window of the ith water flow velocity value are located,Is the sum of the frequencies of the differential elements formed by all the water flow velocity values in the observation window of the ith water flow velocity value in the differential sequence,Is the distribution variance formed by all the water flow velocity values in the observation window of the ith water flow velocity value,Is a parameter adjusting constant for preventing the denominator from being 0,The value of (2) is any positive number not exceeding 0.01, and preferably, the empirical value is 0.001 in the embodiment of the application.
Wherein the first metric valueReflecting the possibility of abnormal data distribution in the observation window of the ith water flow rate value, the frequency of the differential elements formed by all the water flow rate values in the observation window of the ith water flow rate value in the differential sequence is smaller,The smaller the value of (c) is, the less the variation among all the water flow velocity values in the observation window of the ith water flow velocity value is in the flow velocity sequence, the greater the probability of belonging to abnormal water flow velocity fluctuation is, the molecules areReflecting the mutation degree of all water flow velocity values in the observation window of the ith water flow velocity value, the more serious the mutation of the water flow velocity value in the observation window,The larger the value of the difference element formed by all the water flow velocity values in the observation window of the ith water flow velocity value, the easier to be divided into different clusters,The greater the value of (i.e.)The greater the value of (2), the higher the probability that the i-th water flow rate value is an outlier.
Step S003, obtaining a characteristic value sequence of the water flow velocity value according to the monitoring anomaly degree of the water flow velocity value, dividing the characteristic value sequence into a plurality of initial characteristic segments, and calculating the anomaly probability of the initial characteristic segments based on element distribution characteristics and position information in the initial characteristic segments.
Specifically, according to the steps, the monitoring anomaly degree of each water flow velocity value is calculated respectively, and the characteristic value sequences of the water flow velocity values are obtained by arranging according to the time sequence, and because the position structures of different pipelines in the building water supply and drainage system and the water consumption and drainage requirements of water points in the building at different moments are different, the characteristic value sequences of the water flow velocity values are considered to be subjected to sectional processing, and the data characteristics of different sections can be used for representing the change condition of the water flow velocity values under different conditions. Firstly, dividing a characteristic value sequence of a water flow velocity value into a plurality of initial characteristic segments, wherein the characteristic value sequence is as follows:
The characteristic value sequence of the water flow velocity value is used as input, the characteristic value sequence is divided into a plurality of subsequences by utilizing BG (Bernaola Galvan) sequence dividing algorithm, each subsequence is used as an initial characteristic segment, the sequence is divided into common technologies in the field of data processing, and the specific process is not repeated.
In this embodiment, only one sequence division method is provided, that is, the characteristic value sequence is divided into a plurality of sub-sequences by using BG sequence division, and on the premise that the characteristic value sequence division can be achieved, in other embodiments, other sequence division methods may be adopted, which is not particularly limited in the present application.
Secondly, the possibility that each initial characteristic section contains abnormal water flow velocity values is estimated based on the distribution of the monitoring abnormal degrees in different initial characteristic sections and the distance between the monitoring abnormal degrees and the corresponding monitoring positions.
Further, for any one initial characteristic section, determining a monitoring position of a water flow velocity value corresponding to each monitoring anomaly, acquiring coordinate information of the monitoring position, comparing each coordinate information with a pipeline layout file of a water supply and drainage pipeline in a building, marking the coordinate information as a label value of the monitoring position of a pipeline turning position as 1, and otherwise marking the label value as 0. In each initial characteristic section, monitoring anomaly degree with stable distribution exists, if the number of monitoring anomaly degree in one initial characteristic section is small and the interval between monitoring positions is not stable, the water flow velocity value in the initial characteristic section can be caused by the change of the flow velocity at the turning position in the water supply and drainage pipeline or the abnormal condition at different monitoring positions, and the comparison analysis is needed to be carried out with the monitoring anomaly degree in the other initial characteristic sections. The pipeline layout file of the water supply and drainage pipeline in the building refers to a file for recording information about the size, the position, the radius and the like of the pipeline when the water supply and drainage pipeline is placed in the building, and the file usually exists in the form of a building information model BIM and the like at the present stage.
Specifically, the Euclidean distance between the coordinate information of the monitoring positions corresponding to any two monitoring anomalies in the two initial characteristic sections is calculated, and the average value of all Euclidean distances calculated between the two initial characteristic sections is used as the monitoring position distance between the two initial characteristic sections. And taking other initial characteristic sections corresponding to the minimum value of the monitoring position distance between each initial characteristic section as reference characteristic sections of each initial characteristic section.
Here, the probability of abnormality inclusion of each initial characteristic segment is calculated to characterize the magnitude of the probability of monitoring abnormality caused by the abnormal water supply and drainage flow rate phenomenon existing in each initial characteristic segment. The calculation formula of the abnormal probability of the a-th initial feature segment is as follows:
In the formula,Is the outlier weight of the a-th initial feature segment,Is the number of monitored anomalies in the a-th initial feature segment,Is the distribution variance of the monitored outliers in the a-th initial feature segment,Is the average value of the monitoring anomaly degree in the a-th initial characteristic section;
is the probability of inclusion anomaly for the a-th initial feature segment,、The method is characterized in that the a initial characteristic section and the reference characteristic section of the a initial characteristic section monitor abnormal degrees to perform straight line fitting to obtain the slope of the straight line,Is the difference in the number of monitored anomalies in the a-th initial feature segment and the reference feature segment of the a-th initial feature segment,Is the number of monitoring positions with the tag value of 1 in the monitoring positions corresponding to the monitoring abnormality degree in the a-th initial characteristic section and the reference characteristic section of the a-th initial characteristic section.
Wherein,The probability that the monitored outliers in the a-th initial feature segment are outliers compared to the outliers in the remaining initial feature segments is characterized,The larger the value of (c) is, the larger the probability of larger monitoring abnormality calculated from the water flow velocity values caused by the abnormal phenomenon at different monitoring positions is,The smaller the value of (a) is, the more prominent the distribution of the monitored anomalies in the a-th initial feature segment is compared with the overall distribution of all the monitored anomalies;
first difference valueThe difference characteristic of the monitoring abnormal degree in the a-th initial characteristic section and the reference characteristic section is represented on the premise that the monitoring positions are close, the larger the difference of the slopes of the two fitting straight lines is, the larger the difference of the variation trend of the monitoring abnormal degree in the a-th initial characteristic section and the reference characteristic section is, the more likely to be caused by abnormal water flow velocity values, the larger the number of the monitoring abnormal degrees in the monitoring positions in the a-th initial characteristic section and the reference characteristic section is, the larger the probability of the difference characteristic caused by the pipeline structure is, otherwise,The smaller the value of (2), the higher the likelihood that the differential feature is caused by abnormal water flow monitoring values, i.eThe greater the value of a, the greater the likelihood of monitoring anomalies calculated from the values of the flow rate of the abnormal water contained in the a-th initial characteristic segment.
Step S004, obtaining a neighborhood K value of the water flow velocity value according to the monitoring anomaly degree of the water flow velocity value and the anomaly probability of the initial characteristic section to which the water flow velocity value belongs, and obtaining the anomaly water flow velocity value by utilizing an anomaly detection algorithm according to the neighborhood K value of the water flow velocity value.
Firstly, according to the steps, the monitoring anomaly degree of each water flow velocity value and the anomaly probability of each initial characteristic section can be obtained respectively.
And secondly, further analyzing the influence degree of each monitoring anomaly degree in each initial characteristic section on the whole initial characteristic section, evaluating the possibility that each monitoring anomaly degree belongs to an abnormal water flow velocity value, combining the significance degree of the monitoring anomaly degree of each water flow velocity value compared with the monitoring anomaly degree of the water flow velocity value acquired at the monitoring position of all non-pipeline curves, and comprehensively determining the neighborhood K value of each water flow velocity value.
Specifically, from all monitoring anomalies, the monitoring anomalies of the water flow velocity value acquired by the monitoring position with the label value of 1 are removed, the rest monitoring anomalies are utilized to form a screening set, and the information entropy of the monitoring anomalies in the screening set is used as a first threshold value. And in the screening set, counting the data level of the monitoring anomaly degree, and taking the monitoring anomaly degree with equal values as the data level of one anomaly degree.
Further, for any one abnormal data level, the difference between the first threshold and the information entropy of the residual monitoring abnormal degree in the screening set after deleting the abnormal data level is used as the distribution influence weight of the abnormal data level.
Here, a neighborhood K value for each water flow velocity value is calculated based on the distribution impact weight of the data level for each anomaly and the inclusion anomaly probability for each initial feature segment. The calculation formula of the neighborhood K value of the jth water flow velocity value is as follows:
In the formula,Is the neighborhood distance of the jth water flow velocity value,Is a function of the rounding-off,Is the distribution influence weight of the data level to which the monitoring anomaly of the jth water flow velocity value belongs,Is the abnormal probability of the initial characteristic section to which the monitoring abnormality degree of the jth water flow velocity value belongs,Is the maximum value of the inclusion anomaly probabilities for all initial feature segments;
Is the neighborhood K value of the jth water flow rate value,Is the upper limit value of the neighborhood K value when detecting abnormal data of the water flow velocity,The magnitude takes the empirical value 50, which is set to avoid that too large a value of K may result in a reduced sensitivity of the algorithm to outliers. This is because the local anomaly factor is based on a comparison of local densities, and if the K value is too large, the estimate of the local density may be affected by more distant points, thereby reducing the ability to detect true anomalies.
Further, according to the steps, respectively obtaining a neighborhood K value of each water flow velocity value in the water flow velocity data, and obtaining an abnormal water flow velocity value through an LOF algorithm.
Specifically, all the water flow velocity values and the neighborhood K values of the water flow velocity values are used as inputs, an outlier factor detection LOF algorithm is used to obtain local anomaly factors of each water flow velocity value, a second threshold is preset, the embodiment describes with the second threshold being 0.85, and if the local anomaly factors of the water flow velocity values are greater than the second threshold, the water flow velocity values are abnormal water flow velocity values. The LOF algorithm is a common technology for detecting the neighborhood of the abnormal data, and the specific process is not repeated, and when abnormal water flow velocity values occur, the condition that the water flow velocity in a water supply and drainage pipeline in a building is abnormal is indicated.
It should be noted that the foregoing embodiments are merely illustrative of the technical solutions of the present application, and not restrictive, and although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions described in the foregoing embodiments may be modified or some of the technical features may be replaced equivalently, and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application and are included in the protection scope of the present application.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.