A method and device for abnormal diagnosis of power system

Abstract

The present invention provides an abnormal diagnosis method and device for an electric power system, which relates to the technical field of fault detection, wherein the method comprises: determining the importance matrix of all structures that affect the operating state of the electric power system according to a hierarchical structure tree of the electric power system; wherein the hierarchical structure tree comprises a device layer, an indicator layer and a parameter layer that are layered from top to bottom; calculating the final health assessment value of each device according to the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device; determining the health level to which each final health assessment value belongs according to a preset health score table, and inputting the real-time parameter data of the device whose health level is lower than the good level into a preset abnormal diagnosis model after preprocessing, and outputting the abnormal diagnosis result of the device. This scheme can realize real-time abnormal detection of the operating state of the electric power system.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to an abnormality diagnosis method and device for an electric power system. Background Art

[0002] With economic construction and development, in order to meet the needs of people's daily living standards, the scale and complexity of the power system are also constantly increasing. In order to ensure the safe operation of various equipment in the power system, it is usually necessary to conduct regular inspections and maintenance of the power system. The large increase in various equipment also makes the detection and diagnosis of faults and abnormal conditions of the power system face huge challenges.

[0003] In the related technology, when the power system is maintained and repaired, the real-time performance of detection and diagnosis is often poor due to the large volume of system data, and the safety hazards in the system cannot be discovered in time, which leads to emergency repairs only starting after the fault occurs.

[0004] Based on this, there is an urgent need for an abnormal diagnosis method and device for an electric power system to solve the above technical problems. Summary of the invention

[0005] The embodiments of the present invention provide a method, a device, an electronic device and a storage medium for abnormal diagnosis of an electric power system, which can realize real-time detection and diagnosis of the operating status of the electric power system.

[0006] In a first aspect, an embodiment of the present invention provides a method for diagnosing abnormality in a power system, comprising:

[0007] According to the hierarchical structure tree of the power system, the importance matrix of all structures that affect the operation status of the power system is determined; wherein the hierarchical structure tree includes the equipment layer, the indicator layer and the parameter layer which are layered from top to bottom, and the importance matrix is ​​used to characterize the influence of different structures of the power system on the normal operation of the system;

[0008] The final health assessment value of each device is calculated based on the final weight value and initial health assessment value of each structure and the temperature coefficient of the device; wherein the final weight value is obtained by iteratively calculating the importance matrix, and the temperature coefficient is determined by the heat absorption power and heat dissipation power of the system power transmission line;

[0009] The health level to which each final health assessment value belongs is determined according to a preset health score table, and the real-time parameter data of the equipment whose health level is lower than the good level is input into a preset abnormal diagnosis model after preprocessing, and the abnormal diagnosis result of the equipment is output; wherein the abnormal diagnosis model is obtained by training a preset neural network with abnormal data samples of the equipment as input and abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation processing and normalization processing on historical abnormal data sets.

[0010] In a second aspect, an embodiment of the present invention further provides an abnormality diagnosis device for a power system, comprising:

[0011] A determination module is used to determine the importance matrix of all structures that affect the operation status of the power system according to the hierarchical structure tree of the power system; wherein the hierarchical structure tree includes a device layer, an indicator layer and a parameter layer layered from top to bottom, and the importance matrix is ​​used to characterize the influence of different structures of the power system on the normal operation of the system;

[0012] A calculation module, used to calculate the final health assessment value of each device according to the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device; wherein the final weight value is obtained by iteratively calculating the importance matrix, and the temperature coefficient is determined by the heat absorption power and heat dissipation power of the system power transmission line;

[0013] A diagnosis unit is used to determine the health level to which each final health assessment value belongs according to a preset health score table, and input the real-time parameter data of the equipment whose health level is lower than the good level into a preset abnormal diagnosis model after preprocessing, and output the abnormal diagnosis result of the equipment; wherein the abnormal diagnosis model is obtained by training a preset neural network with abnormal data samples of the equipment as input and abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation processing and normalization processing on historical abnormal data sets.

[0014] In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in any embodiment of this specification is implemented.

[0015] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, enables the computer to execute the method described in any embodiment of this specification.

[0016] The embodiment of the present invention provides an abnormal diagnosis method and device for an electric power system. First, the structure of each device in the electric power system is determined, and then the degree of influence of each structure on the normal operation of the device is determined. An importance matrix is ​​established based on this degree. Then, the weight value of each structure is calculated by the importance matrix, and the health status of the device is evaluated in combination with the temperature coefficient of the device. Finally, the device that does not meet the good health level is determined based on the evaluation value, and the device is diagnosed abnormally to obtain the corresponding diagnosis result. This method can not only avoid directly detecting all data of the electric power system, greatly improving the detection speed of the detection and diagnosis system, but also monitor the electric power system in real time, complete the fault diagnosis of the electric power system when the abnormality first appears, and avoid greater losses caused by the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 is a flow chart of a method for diagnosing abnormalities in a power system provided by an embodiment of the present invention;

[0019] Figure 2 is a schematic diagram of a hierarchical structure tree of a wind power generation device provided by an embodiment of the present invention;

[0020] Figure 3 is a hardware architecture diagram of an electronic device provided by an embodiment of the present invention;

[0021] Figure 4 It is a structural diagram of an abnormality diagnosis device for a power system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] As mentioned above, when conducting operation and maintenance inspections of power systems, relevant technologies often perform maintenance and repairs after a fault occurs, which greatly affects the normal operation of the power system.

[0024] Based on this, the idea of ​​the present invention is to score the real-time operating data of the power system by establishing a health evaluation system, so as to find out the hidden dangers that do not meet the standards and detect and diagnose them, so as to ensure that the power system is always in a healthy operating state.

[0025] The specific implementation of the above concept is described below.

[0026] Please refer to Figure 1 , an embodiment of the present invention provides a method for diagnosing abnormality in a power system, the method comprising:

[0027] Step 100, determining the importance matrix of all structures that affect the operating state of the power system according to the hierarchical structure tree of the power system;

[0028] Step 102, calculating a final health assessment value of each device according to the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device;

[0029] Step 104, determine the health level to which each final health assessment value belongs according to a preset health score table, and input the real-time parameter data of the equipment with a health level lower than the good level into the preset abnormal diagnosis model after preprocessing, and output the abnormal diagnosis result of the equipment.

[0030] In the embodiment of the present invention, the structure of each device in the power system is first determined, and then the degree of influence of each structure on the normal operation of the device is determined. An importance matrix is ​​established based on this degree, and then the weight value of each structure is calculated by the importance matrix, and the health status of the device is evaluated in combination with the temperature coefficient of the device. Finally, the device that does not meet the good health level is determined based on the evaluation value, and the device is diagnosed abnormally to obtain the corresponding diagnosis result. This method can not only avoid directly detecting all data of the power system, greatly improving the detection speed of the detection and diagnosis system, but also monitor the power system in real time, complete the fault diagnosis of the power system when the abnormality first appears, and avoid greater losses caused by the fault.

[0031] Described below Figure 1 How the various steps are performed.

[0032] First, for step 100, the importance matrix of all structures that affect the operating state of the power system is determined according to the hierarchical structure tree of the power system.

[0033] A hierarchical structure tree is a data structure widely used in computer science. It represents hierarchical relationships in the form of a tree. For a power system, a hierarchical structure tree includes a device layer, an indicator layer, and a parameter layer from top to bottom.

[0034] For example, taking the wind power generation equipment included in the power system as an example, for the wind power generation equipment, the equipment layer is the wind power generation equipment. According to the historical information such as the price standard of the wind power generation equipment, the indicator layer includes environmental indicators, operating status indicators and economic indicators, among which the environmental indicators include the data of the parameter layer such as the ambient wind speed, air density, and operating years; the operating status indicators include the data of the parameter layer such as the power generation, power conversion rate, and working voltage; the economic indicators include the data of the parameter layer such as the operation and maintenance cost, the life cycle cost, etc., thereby obtaining the following: Figure 2 The hierarchical structure of a wind turbine is shown.

[0035] It is worth noting that, in addition to the above parameters, the equipment in the power system also includes other parameters and indicators that affect its normal operation. These can be adjusted by technical personnel in this field according to the needs of actual operation and maintenance inspection, and will not be elaborated here.

[0036] Furthermore, after obtaining the hierarchical structure tree of the power system, it is necessary to determine the importance matrix of each structure at the same level according to a preset importance scale. The importance scale is used to indicate: compare every two structures at the same level with each other to determine the importance of the two relative to the parent node of the level. Specifically, 0.1 means that when two structures are compared, the latter structure is particularly important than the former structure; 0.2 means that when two structures are compared, the latter structure is strongly important than the former structure; 0.3 means that when two structures are compared, the latter structure is obviously more important than the former structure; 0.4 means that when two structures are compared, the latter structure is slightly more important than the former structure; 0.5 means that when two structures are compared, the latter structure is equally important as the former structure; 0.6 means that when two structures are compared, the former structure is slightly more important than the latter structure; 0.7 means that when two structures are compared, the former structure is obviously more important than the latter structure; 0.8 means that when two structures are compared, the former structure is strongly important than the latter structure; 0.9 means that when two structures are compared, the former structure is particularly important than the latter structure.

[0037] Furthermore, by consulting the data and consulting the relevant staff, it is known that among the three indicators that affect the working level of wind power generation equipment, the influence of the operating status indicator is the most important, and the importance of the environmental indicator is greater than that of the economic indicator. Combining the above importance scale values, we can get the importance matrix M with the economic indicators, operating status indicators and environmental indicators in order:

[0038]

[0039] Then, for step 102, the final health assessment value of each device is calculated based on the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device.

[0040] In an embodiment of the present invention, the final weight value is obtained by iteratively calculating the importance matrix, and the calculation includes the following steps: calculating the row data sum of each row of data and the column data sum of each column of data in the importance matrix; fuzzy processing is performed on all row data sums and all column data sums to obtain a fuzzy matrix that meets the consistency requirements; hierarchical weight sorting is performed on the fuzzy matrix to obtain the initial weight value of each parameter; and the initial weight value is iteratively updated using the fuzzy matrix after reciprocal transformation to obtain the final weight value that meets the iteration conditions.

[0041] Specifically, suppose that a structure set C containing n structures at a certain level is:

[0042] C={C1 C2...C n}

[0043] Then the importance matrix D, which represents the importance between each structure in the set, is:

[0044]

[0045] Then calculate the row data sum of each row of data and the column data sum of each column of data in the importance matrix:

[0046]

[0047] In the formula, s i For row data and; s j is the column data and; d ij Represents the importance relationship between the i-th structure and the j-th structure; i = 1, 2, ..., n; j = 1, 2, ..., n.

[0048] Then, all row data and all column data are fuzzy transformed by the following formula to obtain each element s in the fuzzy matrix R that meets the consistency requirements: ij :

[0049]

[0050] Then use the following formula to calculate the hierarchical weight sorting of the fuzzy matrix:

[0051]

[0052] In the formula, is the element of the normalized fuzzy matrix; m i is the process variable; w i is the initial weight value.

[0053] Furthermore, the fuzzy matrix R is reciprocally transformed by the following formula to obtain the reciprocal matrix E = (eij ) n×n :

[0054]

[0055] In the formula, e ij are the elements of the reciprocal matrix E.

[0056] Finally, the initial weight value is iteratively calculated using the following formula until the infinite norm of the iterative weight value reaches the maximum eigenvalue, and the weight value is output as the final weight value:

[0057] V k+1 =E·V k

[0058] W=V k+1

[0059] Wherein, k = 1, 2, ..., n; V0 is the initial weight value of each element calculated in the above process; W is the final weight value output when the infinite norm of the weight value reaches the maximum eigenvalue.

[0060] In an embodiment of the present invention, the final health assessment value of the equipment is calculated, including the following steps: normalizing the real-time parameter data of the power system and the preset parameter threshold to obtain the initial health assessment value of each parameter; performing weighted average calculation on the final weight values ​​and the initial health assessment values ​​of all parameters under the same indicator layer to obtain the initial health assessment value of the indicator; performing weighted average calculation on the final weight values ​​and the initial health assessment values ​​of all indicators under the same equipment layer to obtain the initial health assessment value of the equipment; and using the temperature coefficient of the equipment to correct the initial health assessment value of the equipment to obtain the final health assessment value of the equipment.

[0061] Specifically, in order to calculate the initial health assessment value of the device, it is necessary to calculate the initial health assessment value of each parameter in the parameter layer based on the real-time parameter data of the device and the preset parameter threshold. The parameter threshold includes two categories: one is the optimal parameter value to ensure the normal operation of the device, and the other is the highest parameter value at which the device cannot operate normally. It is calculated by the following formula:

[0062]

[0063] Where X is the initial health assessment value of the parameter; i is the real-time parameter data; x0 is the optimal parameter value; x max is the highest parameter value.

[0064] For example, assuming that the threshold of the operating life of wind power equipment is 20 years, the initial value is 0 years, and during actual testing, it is found that the service life is 6 years. The initial health assessment value of the operating life of the equipment is 0.7 obtained through the above calculation formula.

[0065] Furthermore, after obtaining the initial health assessment value of each parameter, the initial health assessment value of each indicator in the indicator layer is calculated by weighted average based on the final weight value of each parameter obtained in the above calculation process. The specific calculation formula is as follows:

[0066]

[0067] Where Z is the initial health assessment value of each structure except the parameters, W k is the final weight value of the kth structure.

[0068] Substituting the parameters under the same indicator into the formula, we can get the initial health assessment value of each indicator in the indicator layer, and then use the initial health assessment value of each indicator as X k Substituting the formula again and repeating the above calculation process with its corresponding final weight value, the initial health assessment value of each device can be calculated.

[0069] Furthermore, considering that the operating state of the power system is also easily affected by the temperature environment of its internal lines, in addition to considering the parameters of each device itself, it is also necessary to determine the final score of each device based on the temperature coefficient determined by the heat absorption power and heat dissipation power of the system's power transmission lines.

[0070] Specifically, the heat absorption power of the power system transmission line includes the Joule heat power generated by the conductor's own resistance and the sunlight absorption power during daytime operation, while the heat dissipation power includes convection heat dissipation power and radiation heat dissipation power.

[0071] The heat absorption power Q1 is calculated by the following formula:

[0072] Q1(t)=q I (t)+q S (t)

[0073] q I (t) = I 2 R ref [1+α(T(t)-T d )]

[0074] q S (t) = S × A S ×D

[0075] In the formula, q I(t) is the Joule heat power; I is the line current; T(t) is the average operating temperature of the line; T d is the rated ambient temperature; R ref is the resistance per unit length at rated ambient temperature; α is the resistance temperature coefficient of the transmission line; q S (t) is the sunlight absorption power; S is the solar radiation power density; A S is the surface heat absorption coefficient of the transmission line; D is the diameter of the transmission line.

[0076] The heat dissipation power Q2 is calculated by the following formula

[0077] Q2(t)=q c (t)+q r (t)

[0078] q c (t) = A c (T(t)-T a (t))

[0079] q r (t) = A r [(273+T(t)) 4 -(273+T a (t)) 4 ]

[0080] Where qc(t) is the convective heat dissipation power; qr(t) is the radiative heat dissipation power; A c is the convective heat transfer coefficient; Ta(t) is the ambient temperature of the transmission line; A r is the radiation heat dissipation coefficient.

[0081] The temperature coefficient A is obtained from this T for:

[0082]

[0083] Finally, the final health assessment value Z of the equipment is calculated according to the following formula f :

[0084] Z f =A T ·Z.

[0085] For step 104, the health level to which each final health assessment value belongs is determined according to the preset health score table, and the real-time parameter data of the equipment with a health level lower than the good level is preprocessed and input into the preset abnormal diagnosis model to output the abnormal diagnosis result of the equipment.

[0086] In an embodiment of the present invention, after calculating the final health assessment value of each device, the value can be compared with a preset health score table to determine the current health status of the device. Among them, since each device is used to perform different functions, the division of its health status is also different. The specific division standard can be determined based on historical usage data and the maintenance experience of relevant technicians. For example, based on experience, wind power generation equipment can be divided into the following health levels: the final health assessment value is 1-0.9 for a good state; 0.9-0.75 for a reminder state; 0.75-0.55 for an abnormal state; 0.55-0 for a fault state. When the final health assessment value of the device is lower than 0.9, it needs to be detected and diagnosed. When multiple devices need to be detected and diagnosed at the same time, they are processed from low to high according to the health level.

[0087] Furthermore, after determining the health level of the equipment, it is necessary to use the trained abnormal diagnosis model to detect and diagnose the equipment whose level is lower than the good state; the abnormal diagnosis model is obtained by training the preset neural network through the abnormal data samples of the equipment as input and the abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation and normalization processing on the historical abnormal data set.

[0088] Specifically, in order to further improve the generalization ability of the neural network model and reduce the overfitting phenomenon of the neural network model, it is necessary to perform K-fold cross-validation on the historical anomaly data set, where the historical anomaly data set can be divided into a training set and a test set, and then the training set is evenly divided into K groups, namely O1, O2, ... O K , training and validation are performed for each set of data in turn. One set is set as the validation set, and the remaining K-1 sets of data are set as the training set. Generally, the K value is 5 to 10. Repeat the training K times, and take the optimal value of the K-times evaluation index as the final evaluation index. Finally, the model is trained again on the entire training set using the optimal hyperparameters, and the test set is used for performance evaluation.

[0089] Furthermore, the minimum-maximum normalization method is used to linearly map the historical anomaly data set to the range of [0,1], and each data in the original sample set is normalized to obtain the anomaly data samples used to train the neural network model and the anomaly results corresponding to the anomaly data samples.

[0090] Furthermore, in order to avoid abnormal data samples from taking too long to train the neural network, the training process needs to be optimized. The specific training process includes the following steps: randomly generating uniformly distributed population individual values ​​as the weights and thresholds of the neural network for position initialization, and calculating the initial fitness of each population individual; using a preset fitness function to verify the initial fitness to determine whether the fitness meets the standard value; if so, assigning the weight and threshold corresponding to the fitness to the neural network to obtain the preset neural network, otherwise updating the fitness of the population individuals and repeating the verification until the fitness meets the standard value or the number of iterations reaches a maximum, assigning the weight and threshold with the highest fitness value to the neural network to obtain the preset neural network; using abnormal data samples to train the preset neural network to obtain an abnormal diagnosis model.

[0091] Specifically, the updating process of the population individuals is to perform global optimization and position update according to the distribution characteristics of the arithmetic operators, and use the error function as the fitness function to check whether the obtained individuals meet the optimal standards. If they meet the standards, the optimal individuals are used as the weights and thresholds of the BP neural network, and the training is repeated several times. The neural network parameters are adjusted to the set accuracy range or until the training times are met, thereby obtaining the preset neural network.

[0092] After the abnormal diagnosis model is obtained through the above process, the model can be used to diagnose the equipment that does not meet the good level. The specific diagnosis process includes the following steps: feature extraction is performed on the preprocessed real-time parameter data to obtain key features for characterizing the equipment working mode and abnormal state; wherein the preprocessing includes data cleaning processing, data synchronization processing, feature normalization processing and missing value filling processing; pattern recognition is performed on the key features to obtain the abnormal identification results of the equipment; the abnormal diagnosis model is used to classify the key features and the abnormal identification results to obtain the abnormal type and abnormal location of the equipment; according to the abnormal type and the abnormal location, the abnormal diagnosis result of the equipment is determined.

[0093] Furthermore, a rule engine is used to convert the abnormality type and abnormality location after summary analysis to obtain abnormality diagnosis information of the equipment; wherein the rule engine is composed of an expert knowledge base and historical maintenance data; the abnormality diagnosis information is reconstructed to obtain the abnormality diagnosis result.

[0094] Specifically, feature extraction is to extract those features that can reveal important patterns and anomalies in the data, such as frequency peaks, periodic changes in time series, etc., and then identify them to obtain behaviors and patterns hidden in the feature data, such as analyzing long-term dependencies and periodic changes in time series data. Finally, the abnormal diagnosis model will classify and locate these abnormal data based on the relationship between the training data and the fault, so as to determine the specific part of the problem in the power system. For example, if the abnormal diagnosis model identifies abnormal fluctuations in current or voltage data, it may indicate problems in the motor, such as insulation failure or poor line contact; for fault classification: if the abnormal diagnosis model identifies continuous high-frequency peaks and increased vibration amplitude in the fluctuating data, it will be classified as a signal that the transformer is about to fail.

[0095] Finally, after obtaining the anomaly type and location, the logic and rule engine can be used to convert the technical analysis results into specific diagnostic information, and then the information can be reconstructed using technologies such as natural language generation to obtain abnormal diagnosis results of the power system that can be directly read.

[0096] like Figure 3 , Figure 4 As shown, an embodiment of the present invention provides an abnormality diagnosis device for a power system. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. From the hardware level, Figure 3 As shown, it is a hardware architecture diagram of an electronic device in which an abnormality diagnosis device of a power system provided by an embodiment of the present invention is located, except Figure 3 In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing messages, etc. Taking software implementation as an example, Figure 4 As shown, as a device in a logical sense, the CPU of the electronic device in which it is located reads the corresponding computer program in the non-volatile memory into the internal memory and runs it. This embodiment provides an abnormality diagnosis device for a power system, including:

[0097] The determination module 400 is used to determine the importance matrix of all structures that affect the operation state of the power system according to the hierarchical structure tree of the power system; wherein the hierarchical structure tree includes a device layer, an indicator layer and a parameter layer which are layered from top to bottom, and the importance matrix is ​​used to characterize the influence of different structures of the power system on the normal operation of the system;

[0098] A calculation module 402 is used to calculate the final health assessment value of each device according to the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device; wherein the final weight value is obtained by iteratively calculating the importance matrix, and the temperature coefficient is determined by the heat absorption power and heat dissipation power of the system power transmission line;

[0099] The diagnosis unit 404 is used to determine the health level to which each final health assessment value belongs according to a preset health score table, and input the real-time parameter data of the equipment whose health level is lower than the good level into a preset abnormal diagnosis model after preprocessing, and output the abnormal diagnosis result of the equipment; wherein the abnormal diagnosis model is obtained by training a preset neural network with the abnormal data samples of the equipment as input and the abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation processing and normalization processing on the historical abnormal data set.

[0100] In an embodiment of the present invention, iterative calculation of the importance matrix to obtain the final weight value of the structure includes: calculating the row data sum of each row of data and the column data sum of each column of data in the importance matrix; fuzzy processing all row data sums and all column data sums to obtain a fuzzy matrix that meets consistency requirements; hierarchical weight sorting processing of the fuzzy matrix to obtain an initial weight value for each parameter; iteratively updating the initial weight value using the fuzzy matrix after reciprocal transformation to obtain a final weight value that meets the iteration conditions.

[0101] In an embodiment of the present invention, when the calculation module 402 calculates the final health assessment value of each device based on the final weight value and initial health assessment value of each structure and the temperature coefficient of the device, it is specifically used to perform the following operations: normalize the real-time parameter data of the power system and the preset parameter threshold to obtain the initial health assessment value of each parameter; perform weighted average calculation on the final weight values ​​and initial health assessment values ​​of all parameters under the same indicator layer to obtain the initial health assessment value of the indicator; perform weighted average calculation on the final weight values ​​and initial health assessment values ​​of all indicators under the same equipment layer to obtain the initial health assessment value of the device; and use the temperature coefficient of the device to correct the initial health assessment value of the device to obtain the final health assessment value of the device.

[0102] In the embodiment of the present invention, the common diagnosis model is specifically obtained by training in the following manner: randomly generating uniformly distributed population individual values ​​as the weights and thresholds of the neural network for position initialization processing, and calculating the initial fitness of each population individual; using a preset fitness function to verify the initial fitness to determine whether the fitness meets the standard value; if so, assigning the weights and thresholds corresponding to the fitness to the neural network to obtain the preset neural network, otherwise updating the fitness of the population individuals and repeating the verification until the fitness meets the standard value or the number of iterations reaches a maximum, assigning the weights and thresholds with the highest fitness values ​​to the neural network to obtain the preset neural network; using abnormal data samples to train the preset neural network to obtain the abnormal diagnosis model.

[0103] In the embodiment of the present invention, the preprocessing includes data cleaning processing, data synchronization processing, feature normalization processing and missing value filling processing.

[0104] In an embodiment of the present invention, the abnormal diagnosis result of the equipment is specifically obtained in the following manner: feature extraction is performed on the preprocessed real-time parameter data to obtain key features for characterizing the working mode and abnormal state of the equipment; pattern recognition is performed on the key features to obtain the abnormal identification result of the equipment; the key features and the abnormal identification result are classified and processed using the abnormal diagnosis model to obtain the abnormal type and abnormal location of the equipment; and the abnormal diagnosis result of the equipment is determined based on the abnormal type and the abnormal location.

[0105] In an embodiment of the present invention, the abnormality diagnosis result of the equipment is determined according to the abnormality type and the abnormality location, including: using a rule engine to convert the abnormality type and the abnormality location after summary analysis to obtain abnormality diagnosis information of the equipment; wherein the rule engine is composed of an expert knowledge base and historical maintenance data; and reconstructing the abnormality diagnosis information to obtain the abnormality diagnosis result.

[0106] It is to be understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on an abnormality diagnosis device for an electric power system. In other embodiments of the present invention, an abnormality diagnosis device for an electric power system may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0107] The information interaction, execution process and other contents between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention. For the specific contents, please refer to the description in the embodiment of the method of the present invention, and no further description is given here.

[0108] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, an abnormality diagnosis method for a power system in any embodiment of the present invention is implemented.

[0109] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the processor executes an abnormality diagnosis method for a power system in any embodiment of the present invention.

[0110] Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above-mentioned embodiments is stored, and a computer (or CPU or MPU) of the system or device can be enabled to read and execute the program code stored in the storage medium.

[0111] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0112] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0113] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0114] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or expansion module is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0115] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical factors in the process, method, article or device including the elements.

[0116] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for abnormality diagnosis of a power system, characterized in that: include: According to the hierarchical structure tree of the power system, the importance matrix of all structures that affect the operation status of the power system is determined; wherein the hierarchical structure tree includes the equipment layer, the indicator layer and the parameter layer which are layered from top to bottom, and the importance matrix is ​​used to characterize the influence of different structures of the power system on the normal operation of the system; The final health assessment value of each device is calculated based on the final weight value and initial health assessment value of each structure and the temperature coefficient of the device; wherein the final weight value is obtained by iteratively calculating the importance matrix, and the temperature coefficient is determined by the heat absorption power and heat dissipation power of the system power transmission line; Determine the health level to which each final health assessment value belongs according to a preset health score table, and input the real-time parameter data of the equipment whose health level is lower than the good level into a preset abnormal diagnosis model after preprocessing, and output the abnormal diagnosis result of the equipment; wherein the abnormal diagnosis model is obtained by training a preset neural network with abnormal data samples of the equipment as input and abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation processing and normalization processing on historical abnormal data sets; The heat absorption power includes the Joule heat power generated by the resistance of the wire itself and the sunlight absorption power during daytime operation, and the heat dissipation power includes convection heat dissipation power and radiation heat dissipation power; Heat absorption power Q 1 is calculated by the following formula: In the formula, is the Joule heat power; I is the line current; is the average operating temperature of the line; is the rated ambient temperature; is the resistance per unit length at rated ambient temperature; is the resistance temperature coefficient of the transmission line; is the sunlight absorption power; S is the solar radiation power density; A S is the surface heat absorption coefficient of the transmission line; D is the diameter of the transmission line; Heat dissipation power Q 2 is calculated by the following formula: In the formula, It is the heat dissipation power through convection; is the radiation heat dissipation power; is the convective heat transfer coefficient; is the ambient temperature of the transmission line; is the radiation heat dissipation coefficient; Temperature coefficient A T for: 。 2. The method according to claim 1, characterized in that: The final weight values ​​of the structure obtained by iterative calculation of the importance matrix include: Calculate the row data sum of each row of data and the column data sum of each column of data of the importance matrix; Perform fuzzy processing on all row data and all column data to obtain a fuzzy matrix that meets the consistency requirements; Perform hierarchical weight sorting on the fuzzy matrix to obtain the initial weight value of each parameter; The initial weight value is iteratively updated using the fuzzy matrix after reciprocal transformation to obtain the final weight value that meets the iteration conditions.

3. The method according to claim 1, characterized in that The final health assessment value of the device is calculated based on the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device, including: Normalize the real-time parameter data of the power system and the preset parameter thresholds to obtain the initial health assessment value of each parameter; The final weight values ​​and initial health assessment values ​​of all parameters under the same indicator layer are weighted averaged to obtain the initial health assessment value of the indicator; The final weight values ​​and initial health assessment values ​​of all indicators under the same device layer are weighted averaged to obtain the initial health assessment value of the device; The initial health assessment value of the device is corrected using the temperature coefficient of the device to obtain the final health assessment value of the device.

4. The method according to claim 1, characterized in that The abnormal diagnosis model is specifically obtained by training in the following manner: The randomly generated uniformly distributed population individual values ​​are used as the weights and thresholds of the neural network for position initialization processing, and the initial fitness of each population individual is calculated; The initial fitness is verified using a preset fitness function to determine whether the fitness meets the standard value; if so, the weight and threshold corresponding to the fitness are assigned to the neural network to obtain the preset neural network; otherwise, the fitness of the individuals in the population is updated and calculated, and the verification is repeated until the fitness meets the standard value or the number of iterations reaches the maximum, and the weight and threshold with the highest fitness value are assigned to the neural network to obtain the preset neural network; The preset neural network is trained using abnormal data samples to obtain the abnormal diagnosis model.

5. The method according to claim 1, characterized in that The preprocessing includes data cleaning processing, data synchronization processing, feature normalization processing and missing value filling processing.

6. The method according to claim 5, characterized in that The abnormal diagnosis result of the device is specifically obtained in the following manner: Perform feature extraction on the preprocessed real-time parameter data to obtain key features for characterizing the equipment working mode and abnormal state; Performing pattern recognition on the key features to obtain an abnormality recognition result of the device; Using the abnormality diagnosis model to classify the key features and the abnormality identification results to obtain the abnormality type and abnormality location of the device; An abnormality diagnosis result of the device is determined according to the abnormality type and the abnormality location.

7. The method according to claim 6, characterized in that The determining, according to the abnormality type and the abnormality location, an abnormality diagnosis result of the device includes: The rule engine is used to convert the abnormality type and abnormality location after summary analysis to obtain abnormal diagnosis information of the equipment; wherein the rule engine is composed of an expert knowledge base and historical maintenance data; The abnormality diagnosis information is reconstructed to obtain the abnormality diagnosis result.

8. An abnormality diagnosis device for a power system, applied to the abnormality diagnosis method for a power system as claimed in any one of claims 1 to 7, characterized in that: include: A determination module is used to determine the importance matrix of all structures that affect the operation status of the power system according to the hierarchical structure tree of the power system; wherein the hierarchical structure tree includes a device layer, an indicator layer and a parameter layer which are layered from top to bottom, and the importance matrix is ​​used to characterize the influence of different structures of the power system on the normal operation of the system; A calculation module, used to calculate the final health assessment value of each device according to the final weight value and the initial health assessment value of each structure and the temperature coefficient of the device; wherein the final weight value is obtained by iteratively calculating the importance matrix, and the temperature coefficient is determined by the heat absorption power and heat dissipation power of the system power transmission line; A diagnosis unit is used to determine the health level to which each final health assessment value belongs according to a preset health score table, and input the real-time parameter data of the equipment whose health level is lower than the good level into a preset abnormal diagnosis model after preprocessing, and output the abnormal diagnosis result of the equipment; wherein the abnormal diagnosis model is obtained by training a preset neural network with abnormal data samples of the equipment as input and abnormal results corresponding to the abnormal data samples as output, and the abnormal data samples are obtained by performing K-fold cross validation processing and normalization processing on historical abnormal data sets.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed in a computer, the computer is caused to execute the method according to any one of claims 1 to 7.

Images