CN113591393B

CN113591393B - Fault diagnosis method, device, equipment and storage medium of intelligent substation

Info

Publication number: CN113591393B
Application number: CN202110913435.5A
Authority: CN
Inventors: 李铁成; 任江波; 刘清泉; 耿少博; 孙利强; 王献志; 徐岩; 王鸣誉
Original assignee: State Grid Corp of China SGCC; Electric Power Research Institute of State Grid Hebei Electric Power Co Ltd; State Grid Hebei Energy Technology Service Co Ltd
Current assignee: State Grid Corp of China SGCC; Electric Power Research Institute of State Grid Hebei Electric Power Co Ltd; State Grid Hebei Energy Technology Service Co Ltd
Priority date: 2021-08-10
Filing date: 2021-08-10
Publication date: 2024-05-31
Anticipated expiration: 2041-08-10
Also published as: CN113591393A

Abstract

The invention is applicable to the technical field of power grid safety, and provides a fault diagnosis method, device, equipment and storage medium of an intelligent substation, wherein the fault diagnosis method of the intelligent substation comprises the following steps: acquiring real-time alarm information of a secondary system; carrying out preset processing on the real-time alarm information to obtain pre-processed alarm information; comparing all preset classified alarm information sets in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain sub-classified alarm information sets of the real-time alarm information; constructing a relevance rule of the sub-classified alarm information set based on a P-growth algorithm; based on an ant lion algorithm, processing preset association indexes of the association rules to obtain association indexes of real-time alarm information; and obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information. By adopting the invention, the accuracy of fault diagnosis can be improved.

Description

Fault diagnosis method, device, equipment and storage medium of intelligent substation

Technical Field

The invention belongs to the technical field of power grid safety, and particularly relates to a fault diagnosis method, device, equipment and storage medium of an intelligent substation.

Background

The intelligent transformer substation adopts advanced, reliable, integrated and environment-friendly intelligent equipment to automatically complete basic functions such as information acquisition, measurement, control, protection, metering, detection and the like. The intelligent substation mainly comprises a primary system and a secondary system, wherein the primary system is a system formed by equipment such as a generator, a transmission line, a transformer, a circuit breaker and the like; the secondary system is a system composed of a measuring meter, an insulation monitoring device, a relay protection device, an automatic device and other devices. The secondary system is used as an auxiliary system for monitoring, measuring, controlling, protecting and adjusting the primary system in the intelligent substation, and the position in the intelligent substation is important. If the secondary system fails in operation, the primary system cannot normally operate, so that the safety of the whole power grid is affected.

At present, the fault diagnosis of the secondary system of the traditional intelligent substation is mainly carried out by a dispatcher according to the alarm information displayed by the secondary system, and the fault diagnosis is carried out by relying on working experience, so that an effective diagnosis means is lacking, and the fault analysis of the secondary system is inaccurate, and the diagnosis accuracy is lower.

In addition, with the continuous development of artificial intelligence, algorithms such as a neural network, a support vector machine, association rule mining and the like are widely applied to fault diagnosis of a secondary system, however, the algorithms are usually optimized on the complexity of fault diagnosis, and the complexity of fault diagnosis is reduced, but the diagnosis accuracy is difficult to improve and the diagnosis accuracy is lower due to the fact that more diagnosis data are difficult to screen.

Disclosure of Invention

In view of the above, the embodiments of the present invention provide a fault diagnosis method, apparatus, device and storage medium for an intelligent substation, so as to solve the problem of low diagnosis accuracy existing in the current fault diagnosis method.

A first aspect of an embodiment of the present invention provides a fault diagnosis method for an intelligent substation, including:

acquiring real-time alarm information of a secondary system of the intelligent substation;

carrying out preset processing on the real-time alarm information to obtain pre-processed alarm information;

Comparing a preset classified alarm information set in a preset C4.5 fault decision tree model with the preprocessed alarm information to obtain a sub-classified alarm information set of real-time alarm information;

Constructing a relevance rule of the sub-classified alarm information set based on a P-growth algorithm;

based on an ant lion algorithm, processing preset association indexes of the association rules to obtain association indexes of real-time alarm information;

And obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information.

In one possible implementation, comparing the preset classified alarm information set in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain the sub-classified alarm information set of the real-time alarm information, including:

Acquiring historical alarm information of a secondary system, wherein the historical alarm information is divided into training information and test information according to preset rules;

Training the training information through a C4.5 decision tree algorithm to obtain a decision tree model;

Verifying and pruning the decision tree model by using the test information to obtain a preset C4.5 fault decision tree model; the preset C4.5 fault decision tree model consists of a plurality of preset classified alarm information sets with different fault types;

And comparing all preset classified alarm information sets in the preset C4.5 fault decision tree model with the preset alarm information, and determining the preset classified alarm information set with the highest matching degree as a sub-classified alarm information set of the real-time alarm information.

In one possible implementation, before training the training information through the C4.5 decision tree algorithm to obtain the decision tree model, the method further includes:

constructing a fault attribute database of historical alarm information; the history alarm information comprises secondary equipment alarm information, GOOSE abnormal messages and SV abnormal messages; the fault attribute database comprises device locking, self-checking alarm information, GOOSE alarm information, SV alarm information and fault types; the historical alarm information is stored in a fault attribute database according to a preset rule;

and constructing a decision tree model based on the C4.5 decision tree algorithm and the fault attribute database.

In one possible implementation, the C4.5 decision tree algorithm uses a Taylor series instead of a logarithmic operation;

Wherein, the information gain rate of the C4.5 decision tree algorithm is that

Gainratio (a) is the information Gain rate of a certain attribute a selected during classification, gain (a) is the information Gain of attribute a, splitInfo _A (T) is the split information measure of attribute a in the whole preset classified alarm information set T, T is the total number of data of the j-th preset classified alarm information set in the whole preset classified alarm information set T, TC _i is the total number of data belonging to the sub-classified alarm information set C _i in the whole preset classified alarm information set T, and TC _ij represents the total number of data belonging to the sub-classified alarm information set C _i in the j-th preset classified alarm information set in the whole preset classified alarm information set T.

In one possible implementation, constructing the relevance rule of the sub-classified alarm information set based on the P-growth algorithm includes:

performing iterative processing on the preset minimum support of the relevance rule based on the ant lion algorithm to obtain the minimum support of the sub-classification alarm information set;

deleting target alarm information with less than the minimum support degree in the sub-classified alarm information set to obtain a sub-classified alarm associated information set; wherein, the target alarm information is any alarm information in the sub-classified alarm information set;

constructing a relevance rule of the sub-classification alarm relevance information set based on a P-growth algorithm;

Wherein the support degree is A is target alarm information, N is the number of all target alarm information, and sup (A) is the support degree of the target alarm information A.

In one possible implementation, the association indicator includes a confidence level and an imbalance ratio;

Based on ant lion algorithm, processing the preset association index of the association rule to obtain the association index of the real-time alarm information, comprising:

And carrying out iterative processing on the preset first confidence coefficient and the first unbalance ratio based on an ant lion algorithm to obtain a second confidence coefficient and a second unbalance ratio of the real-time alarm information.

In one possible implementation manner, obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information includes:

when the confidence coefficient of the target association rule is larger than the second confidence coefficient and the unbalance ratio of the target association rule is smaller than the second unbalance ratio, determining the front piece of the target association rule as root alarm information in the real-time alarm information; wherein the target association rule is any one of association rules;

Wherein the confidence is A and B are respectively any different target alarm information in the sub-classified alarm information set,/>On the premise that the alarm information of the target A occurs, the probability of occurrence of the alarm information of the target B;

unbalance ratio of Sup (A) is the support degree of the target alarm information A, sup (B) is the support degree of the target alarm information B, and Sup (A-B) is the support degree of the target alarm information containing both A and B.

A second aspect of an embodiment of the present invention provides a fault diagnosis apparatus for an intelligent substation, including:

the information acquisition module is used for acquiring real-time alarm information of a secondary system of the intelligent substation;

the preset processing module is used for carrying out preset processing on the real-time alarm information to obtain the preset alarm information;

The information comparison module is used for comparing a preset classified alarm information set in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain a sub-classified alarm information set of the real-time alarm information;

The construction rule module is used for constructing the relevance rule of the sub-classification alarm information set based on the P-growth algorithm;

the index optimization module is used for processing preset association indexes of the association rules based on the ant lion algorithm to obtain association indexes of the real-time alarm information;

And the information determining module is used for obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information.

In one possible implementation manner, the information comparison module is further configured to obtain historical alarm information of the secondary system, where the historical alarm information is divided into training information and test information according to a preset rule;

In one possible implementation manner, the information comparison module is further configured to train the training information through a C4.5 decision tree algorithm, and before obtaining the decision tree model, the method further includes:

In one possible implementation, the C4.5 decision tree algorithm uses a Taylor series instead of a logarithmic operation; wherein, the information gain rate of the C4.5 decision tree algorithm is that

In one possible implementation manner, the construction rule module is further used for carrying out iterative processing on the preset minimum support of the relevance rule based on the ant lion algorithm to obtain the minimum support of the sub-classification alarm information set;

correspondingly, the index optimization module is also used for,

In one possible implementation manner, the information determining module is further configured to determine, when the confidence coefficient of the target association rule is greater than the second confidence coefficient and the imbalance ratio of the target association rule is smaller than the second imbalance ratio, a front piece of the target association rule as root alert information in the real-time alert information; wherein the target association rule is any one of association rules;

A third aspect of an embodiment of the invention provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method according to the first aspect when executing the computer program.

A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method according to the first aspect.

Compared with the prior art, the embodiment of the invention has the beneficial effects that:

In the embodiment of the invention, the real-time alarm information of the secondary system of the intelligent substation is firstly obtained, and the real-time alarm information is subjected to preset processing to obtain the preprocessed alarm information; and then, comparing the preset classified alarm information set in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain the sub-classified alarm information set of the real-time alarm information. And then, constructing a relevance rule of the sub-classified alarm information set based on a P-growth algorithm, and processing preset relevance indexes of the relevance rule based on an ant lion algorithm to obtain the relevance indexes of the real-time alarm information. And finally, obtaining the root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information. Therefore, the root alarm information can be obtained after the real-time alarm information is processed by presetting a C4.5 fault decision tree model, a P-growth algorithm and an ant lion algorithm, and the diagnosis accuracy is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a step flowchart of a fault diagnosis method of an intelligent substation provided by an embodiment of the present invention;

FIG. 2 is a schematic diagram of a decision tree for protecting a device from failure according to an embodiment of the present invention;

FIG. 3 is a flowchart of a C4.5 decision tree algorithm provided by an embodiment of the present invention;

Fig. 4 is a flowchart of an ant lion algorithm provided in an embodiment of the present invention;

fig. 5 is a schematic diagram of a fault diagnosis device of an intelligent substation according to an embodiment of the present invention;

fig. 6 is a schematic diagram of an electronic device according to an embodiment of the present invention.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the following description will be made by way of specific embodiments with reference to the accompanying drawings.

At present, correlation analysis of alarm information is also used in fault diagnosis of a secondary system of an intelligent substation, but the traditional correlation analysis can generate a large amount of useless diagnosis data, so that the accuracy of diagnosis is low, and a dispatcher cannot obtain an accurate fault cause.

In order to solve the problems in the prior art, the embodiment of the invention provides a fault diagnosis method, device, equipment and storage medium of an intelligent substation. The fault diagnosis method of the intelligent substation provided by the embodiment of the invention is first described below.

The main body of execution of the fault diagnosis method may be a fault diagnosis apparatus, which may be an electronic device having a processor and a memory, such as a mobile electronic device or a non-mobile electronic device. The mobile electronic device may be a mobile phone, a tablet computer, a notebook computer, a palm computer, a vehicle-mounted electronic device, a wearable device, or the like, and the non-mobile electronic device may be a server, a network attached memory, a personal computer, or the like, which is not particularly limited.

As shown in fig. 1, an embodiment of the present invention provides a fault diagnosis method for an intelligent substation, which may include the following steps:

And step S110, acquiring real-time alarm information of a secondary system of the intelligent substation.

In some embodiments, the acquisition of real-time alarm information of the secondary system of the intelligent substation needs to be acquired through two parts of a dispatching end and station end equipment.

The master station system of the dispatching end is deployed at the provincial dispatching end. The station end equipment can acquire the alarm information of the secondary system in the intelligent substation by utilizing the network message analysis recording device. The station terminal equipment performs preprocessing such as network information collection and filtration, timely collects the network abnormal information when abnormal conditions are found, and immediately reports the network abnormal information to the main station of the dispatching terminal. The alarm information is uploaded through the D5000 platform of the dispatching system and displayed on the interface, and the time window of the same piece of alarm information is set to be 8s.

The intelligent substation secondary system mainly comprises the following information output by the intelligent substation secondary system according to relevant specifications: running, alarming and acting, wherein the running information mainly comprises self-checking information, fault information, message state information and the like which are output by the device; the action information mainly comprises the power supply working state, the measured temperature, the equipment working condition and the like of each device; the alarm information mainly refers to sampling of a secondary system, abnormal switching value alarm signals and abnormal device alarm signals, and the abnormal switching value alarm signals mainly comprise: input signal anomalies: device sampling abnormality, on/off amount abnormality, SV/GOOSE broken link, SV quality abnormality and the like, and device abnormality alarm signals mainly comprise device power failure alarm, device locking and the like.

And step S120, carrying out preset processing on the real-time alarm information to obtain the preprocessed alarm information.

In some embodiments, since the alert information in the scheduling system D5000 includes the following useless alert information, such as nonsensical alert information or strobe alert information, the nonsensical alert information must be removed. And carrying out preset processing on the alarm information, wherein the preset processing is to remove meaningless alarm information and merge stroboscopic alarm information, and the preset processing is carried out to obtain the preprocessed alarm information.

And S130, comparing all preset classified alarm information sets in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain sub-classified alarm information sets of the real-time alarm information.

In some embodiments, a preset C4.5 fault decision tree model needs to be built to implement classification of alarm information.

First, the history warning information of the secondary system is acquired. The history alarm information comprises secondary system alarm information, GOOSE abnormal messages, SV abnormal messages and the like; the secondary system alarm information mainly comprises self-checking abnormality, sampling error, power failure and the like; the GOOSE abnormal message mainly comprises a GOOSE broken link, a flow abnormality, a file configuration error and the like; the SV abnormal message mainly comprises SV broken link, asynchronous sampling data, message counting error and the like; there is also device lockout information, which is not described in detail herein.

And then, storing the historical alarm information into a fault attribute database according to a preset rule. The fault attribute database comprises attributes and classification results, wherein the attributes comprise 4 types of device locking, self-checking alarm information, GOOSE alarm information and SV alarm information, and the fault types are the classification results. The method comprises the steps of setting alarm information of a secondary system in the history alarm information under a self-checking alarm information attribute, setting device locking information under a device locking attribute, setting a GOOSE abnormal message under the GOOSE alarm information attribute, and setting an SV abnormal message under the SV alarm information attribute. The fault types are divided into 5 major categories, 19 types. The fault information is the result of manual diagnosis, wherein the fault type mainly comprises:

① A merging unit: I/O port failure, power failure, program failure, configuration failure, sampling DSP failure;

② The protection device comprises: I/O port failure, power failure, program failure, configuration failure;

③ And (3) an intelligent terminal: I/O port failure, power failure, program failure, configuration failure;

④ And the measurement and control device comprises: I/O port failure, power failure, program failure, configuration failure;

⑤ Communication link: switch failure, communication fiber failure;

Specifically, the fault attribute database may be constructed according to the format of the following table one, and it should be noted that, this database only exemplifies the format of the fault attribute database, and only puts part of alarm information and fault types into the table, but not all the information in the embodiment of the present invention.

List one

And then, dividing the historical alarm information into training information and test information according to preset rules. Specifically, three-quarter of the alarm information can be used as training information, and one-fourth of the alarm information can be used as test information. Training the training information through a C4.5 decision tree algorithm to obtain a decision tree model, wherein the fault type in the decision tree model is a leaf node, and the branches are device locking, self-checking alarm information, GOOSE alarm information and SV alarm information in a fault attribute database. And verifying and pruning the decision tree model by using the test information to obtain a preset C4.5 fault decision tree model. After the preset C4.5 fault decision tree model is constructed, each path from the root node to the leaf node corresponds to a fault classification mode. Specifically, the preset C4.5 fault decision tree model is composed of 19 preset classified alarm information sets with different fault types. Fig. 2 illustrates four faults of the protection device as an example, and a part of the built preset C4.5 fault decision tree model is set.

In addition, the c4.5 decision tree algorithm needs to use the information gain rate of the historical alarm information to select the classified attribute, classify the sample through the classified attribute, and calculate the information entropy firstly before calculating the information gain rate, wherein the information entropy reflects the uncertainty degree of the information source. Here, the calculation process is not described in detail. However, it should be noted that, when the conventional c4.5 decision tree algorithm calculates the information gain rate, the logarithmic calculation is involved, and when the sample size is large, the logarithmic calculation increases the time of data processing, so that the invention specifically introduces a taylor series to avoid this situation, and omits the logarithmic calculation. The following was demonstrated:

when x takes infinity, the formula can be reduced to:

Specifically, as shown in the flowchart of the decision tree algorithm in fig. 3, the specific steps are as follows:

Step l: a root node is created.

Step 2: if the training set is an empty set, the leaf nodes are empty sets.

Step 3: the information gain ratio of the attributes in the attribute list is calculated.

Step 4: and selecting the attribute with the highest information gain rate from the attributes as a classification standard, and marking the internal node as attribute A.

Step 5: if the samples in attribute A belong to the same label, the label is the leaf node of the current sample.

Step 6: if the samples in the attribute A do not belong to the same label, the data in the attribute A are further divided, the information gain rate in the attribute A is recalculated, and a decision tree subtree is established.

Step 7: and 5, repeating the steps 5 and 6 to classify all the data in the attribute A until all the data in the attribute A are classified.

Step 8: and classifying the attributes in each attribute list, and ending classification to form an initial decision tree if the attribute list of the secondary fault training data in the database is empty.

Step 9: and (3) performing bottom-up evaluation on the decision tree model by using the test set, and judging whether pruning is required for the leaf nodes.

Step 10: pruning is carried out on branches needing pruning, and a complete preset C4.5 fault decision tree model is obtained.

Wherein, the information gain rate of the C4.5 decision tree algorithm is that

And finally, comparing all preset classified alarm information sets in the preset C4.5 fault decision tree model with the preset alarm information, and determining the preset classified alarm information set with the highest matching degree as a sub-classified alarm information set of the real-time alarm information.

The invention constructs a preset C4.5 fault decision tree model by utilizing a C4.5 decision tree algorithm, classifies the historical alarm information according to fault types to form a preset classified alarm information set, and the path from the root node to the leaf node of each branch corresponds to a fault classification mode. Therefore, the sub-classified alarm information set of the real-time alarm information with the highest matching degree can be obtained by inputting the preset alarm information into the preset C4.5 fault decision tree model. Therefore, the diagnosis data can be reduced, judgment from a plurality of diagnosis data is not needed, and the diagnosis accuracy is improved.

And step 140, constructing a relevance rule of the sub-classified alarm information set based on a P-growth algorithm.

Before performing the association analysis, first, a minimum support is determined. Specifically, the support degree refers to the probability of occurrence of the event, namely the ratio of the alarm information in the total data in the fault analysis, the alarm data with the too low occurrence number is considered to have no research value, and the invention only extracts the information meeting the support degree counting index.

In some embodiments, firstly, based on ant lion algorithm, iterative processing is performed on the preset minimum support of the association rule in advance to obtain the minimum support of the sub-classification alarm information set. Specifically, the preset minimum support degree may be set to 0, and the ant lion algorithm is adopted to perform multiple iterations to obtain the optimal minimum support degree. Then, deleting the target alarm information with the sub-classification alarm information set less than the minimum support degree to obtain a sub-classification alarm associated information set; the target alarm information is any one alarm information in the sub-classified alarm information set. And part of useless target alarm information is removed before the relevance rule is constructed, so that the constructed relevance rule data is reduced, and the diagnosis accuracy is improved. Wherein the support degree isA is target alarm information, N is the number of all target alarm information, and sup (A) is the support degree of the target alarm information A.

And finally, constructing a relevance rule of the sub-classification alarm relevance information set based on a P-growth algorithm.

The principle of the P-growth algorithm is as follows: the first scanning records the two replacement sets of all the transactions to obtain a vertex set graph. After all vertex and two-term displacement sets are obtained, a correlation diagram is drawn. And connecting any two replacement sets, when the replacement occurs again, adding 1 to the connection count, and removing the frequent association according to the association graph and the support count by the algorithm to obtain the frequent association set. All two frequent item sets in the transaction set can be obtained according to the frequent association set. Three frequent item sets can be obtained according to the frequent association set.

And step S150, processing preset association indexes of the association rule based on the ant lion algorithm to obtain the association indexes of the real-time alarm information.

In some embodiments, the association index includes a confidence level and an imbalance ratio. Where confidence refers to the probability of one transaction occurring while another transaction occurs. In fault diagnosis, it can be understood that when one type of alarm information occurs, the probability of occurrence of another type of alarm signal may reflect the derivative relationship of one type of alarm signal and another type of alarm signal. So as to analyze which of the alert signals may reflect the root cause of the fault. In order to find the correlation degree between the two transactions, a concept of degree of promotion is introduced, which can reflect A, B the correlation between the two transactions, and in fault analysis, that is, whether the two alarm signals show positive, negative correlation or no correlation. But the degree of lift is greatly affected by zero transactions: when zero transactions irrelevant to both A and B are less, the lifting degree is higher; otherwise, the lifting degree is smaller. Besides the index, the invention introduces an unbalance ratio IR, has zero invariance and is suitable for databases with more data. The two indexes are matched to reflect the correlation between the transactions. The unbalance ratio index can better reflect the relation between the occurrence probability of two transactions in the rule, and the closer the unbalance ratio is to zero, the greater the occurrence probability of the back piece under the condition that the front piece occurs, and the reverse is also true. While a larger imbalance ratio indicates that the front piece is occurring and the rear piece is more likely to occur, but vice versa. Wherein the confidence is A and B are any different target alert information in the sub-category alert information set,On the premise that the alarm information of the target A occurs, the probability of occurrence of the alarm information of the target B; unbalance ratio ofSup (A) is the support degree of the target alarm information A, sup (B) is the support degree of the target alarm information B, and Sup (A-B) is the support degree of the target alarm information containing both A and B.

In the invention, first, initial values of confidence and unbalance ratio are preset, namely first confidence and first unbalance ratio respectively. As shown in fig. 4, the ant lion algorithm is adopted to perform iterative processing, wherein the input quantity is the support degree, the confidence degree and the unbalance ratio, the weighted sum of the data coverage rate and the average unbalance ratio of all rules is taken as the fitness function, and the output quantity is the support degree, the confidence degree and the unbalance ratio in the optimal state, and the fitness value at the moment.

The parameters of the ant lion algorithm set in the invention are as follows: the number of ants and ant lions was set to 10, the maximum number of iterations was 100, and parameters K1 and K2 were 0.9 and 0.1, respectively.

The fitness function set is the weighted sum of the data coverage and the average imbalance ratio of all rules:

f(x₁)＝-(K₁*CR-K₂*IR_ave)；

the data coverage ratio CR refers to the ratio of the number of alarm signals in all generated rules to the number of all alarm signals in the sample set, and is used to detect whether the result is successful in full coverage. IR _ave refers to the average imbalance ratio of all rules.

And after the preset first confidence coefficient and the first unbalance ratio are processed by the ant lion algorithm, obtaining the optimal second confidence coefficient and the second unbalance ratio of the real-time alarm information.

Subsequently, a correlation analysis may be performed.

And step S150, obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information.

In some embodiments, when the confidence level of the target association rule is greater than the second confidence level and the imbalance ratio of the target association rule is less than the second imbalance ratio, determining the front piece of the target association rule as the root alert information in the real-time alert information; wherein the target association rule is any one of association rules.

After screening out the association rule and the optimal association index of the real-time alarm information, the alarm signal meeting the following conditions can be found out in the association rule: the confidence coefficient is larger than or equal to the second confidence coefficient, the unbalance ratio is smaller than or equal to the second unbalance ratio, and the generated association rule number is the largest. The alarm signal meeting the above condition is the root alarm.

The diagnosis data are more difficult to screen in the conventional fault diagnosis, so that the diagnosis accuracy is lower. In the embodiment of the invention, the real-time alarm information of the secondary system is firstly obtained, and the real-time alarm information is subjected to preset processing to obtain the pre-processing alarm information; and then, comparing the preset classified alarm information set in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain the sub-classified alarm information set of the real-time alarm information. And then, constructing a relevance rule of the sub-classified alarm information set based on a P-growth algorithm, and processing preset relevance indexes of the relevance rule based on an ant lion algorithm to obtain the relevance indexes of the real-time alarm information. And finally, obtaining the root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information. Therefore, the root alarm information can be obtained after the real-time alarm information is processed by presetting a C4.5 fault decision tree model, a P-growth algorithm and an ant lion algorithm, and the diagnosis accuracy is improved.

For example: and a single-phase short circuit fault occurs in a certain intelligent station, so that protection is refused. When a fault occurs, the secondary system displays the following information: the merging unit sends out an abnormal alarm signal of the device, the merging unit samples and gives an abnormal alarm, the protection device samples and gives an alarm, and the merging unit communicates with the protection device and gives an abnormal state. The alarm signal of the fault is: the merging unit device 11-1 alarms abnormally, the merging unit 11-1 samples abnormally, the protection device 101SV alarms and the protection 101 samples alarms. According to the analysis result of the preset C4.5 fault decision tree model, the fault type can be determined as the combined unit DSP fault. And comparing the alarm signals with all preset classified alarm information sets of the fault types of the category, thereby determining the sub-classified alarm information sets of the fault types.

The preset minimum support degree is 0, the preset first confidence degree is 0, and the preset first unbalance ratio is 1. After being processed by the ant lion algorithm, the optimal minimum support degree is 0.05, the second confidence degree is 0.648, and the second unbalance ratio is 0.36.

Then, a P-growth algorithm is adopted to construct a relevance rule of the sub-classification alarm information set, and the result of data mining can be known: the front part is the most association rule which accords with the confidence requirement and the unbalance ratio requirement and is used for merging unit sampling abnormality. The confidence of the sampling abnormality of the merging unit 11-1 to the alarm abnormality of the merging unit device 11-1 is 0.97, and the unbalance ratio is 0.35; the confidence of the sampling abnormality of the merging unit 11-1 on the SV alarm of the protection device 101 is 0.72, and the unbalance ratio is 0.34; the confidence of the sampling abnormality of the merging unit 11-1 on the sampling alarm of the protection 101 is 0.75, and the unbalance ratio is 0.26.

The confidence of the sampling anomaly of the merging unit 11-1 on the other three alarm signals is larger than 0.648, and the kulc measurement is larger, so that the two alarm signals are in close connection, besides, the unbalance ratio is smaller than 0.36, and the situation that the former happens is proved to be more likely, and the reverse situation is also true. In summary, the merging unit 11-1 samples the root alarm signal of the fault, and the rest signals are derived alarm signals. The fault occurs in the merging unit sampling module.

If the association rules are not screened, up to 17 association rules are given which meet the requirements, so that a lot of unnecessary association rules are reduced after alarm correlation analysis is added. Therefore, the accuracy of fault diagnosis is improved, and the efficiency of fault analysis is further improved.

For example: when misoperation occurs to relay protection of an intelligent station for a certain time, communication faults occur to a merging unit-protection device loop, and a dispatching center receives the following alarm information: protection device 101GOOSE alarm, protection device 101SV alarm, intelligent terminal 101GOOSE alarm, switch detection line communication interruption. And comparing analysis results of the preset fault decision tree model, and determining that the fault type is the fault of the I/O port of the protection device. Thereby determining a sub-classified set of alert information for this type of failure of the real-time alert information.

The preset minimum support degree is 0, the preset first confidence degree is 0, and the preset first unbalance ratio is 1. After being processed by ant lion algorithm, the optimal minimum support degree is 0.03, the second confidence degree is 0.553, the second unbalance ratio is 0.29,

Then, a P-growth algorithm is adopted to construct relevance rules of the sub-classification alarm information set, and 6 relevance rules are respectively 1, line communication interruption, line protection device 101GOOSE alarm, 2, line communication interruption, line protection device 101SV alarm, 3, line communication interruption, intelligent terminal 101GOOSE alarm, 4, line protection device 101GOOSE alarm, line protection device 101SV alarm, 5, line protection device 101GOOSE alarm, intelligent terminal 101GOOSE alarm, 6, line protection device 101SV alarm, intelligent terminal 101GOOSE alarm. Wherein, the confidence coefficient of the relevance rule 1 is 0.578, and the unbalance ratio is 0.45; the confidence coefficient of the association rule 2 is 0.564, and the unbalance ratio is 0.46; the confidence coefficient of the association rule 3 is 0.655, and the unbalance ratio is 0.62; the confidence coefficient of the association rule 4 is 0.673, and the unbalance ratio is 0.02; the confidence coefficient of the association rule 5 is 0.603, and the unbalance ratio is 0.23; the confidence level of the association rule 6 is 0.592 and the imbalance ratio is 0.24.

The confidence coefficient of each relevance rule is larger than 0.553, but only the unbalance ratio of the relevance rules with the numbers of 4,5 and 6 meets the index requirement, so that only the protection device 101GOOSE alarm and the protection device SV alarm are root alarms.

And comparing the correlation analysis when the preset fault decision tree model is not classified, when the alarm information is not classified, the result of the correlation analysis is interfered by other types of fault alarm information when the confidence coefficient is calculated, for example, when a merging unit breaks down, the quantity of the protection device 101GOOSE and SV alarm signals is increased, so that the confidence coefficient of the correlation rule of the protection device 101GOOSE and SV alarm signals as the front part is reduced. So that the screening criteria are not met, resulting in no root alert being found.

Based on the fault diagnosis method of the intelligent substation provided by the embodiment, correspondingly, the invention further provides a specific implementation mode of the fault diagnosis device of the intelligent substation, which is applied to the fault diagnosis method of the intelligent substation. Please refer to the following examples.

As shown in fig. 5, there is provided a fault diagnosis apparatus of an intelligent substation, the apparatus comprising:

The information acquisition module 510 acquires real-time alarm information of the secondary system;

the preset processing module 520 performs preset processing on the real-time alarm information to obtain preset alarm information;

The information comparison module 530 compares the preset classified alarm information set in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain a sub-classified alarm information set of the real-time alarm information;

the construction rule module 540 constructs a relevance rule of the sub-classified alarm information set based on a P-growth algorithm;

The index optimization module 550 processes preset association indexes of the association rules based on the ant lion algorithm to obtain association indexes of the real-time alarm information;

The information determining module 560 obtains the root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information.

In a possible implementation manner, the information comparison module 530 is further configured to obtain historical alarm information of the secondary system, where the historical alarm information is divided into training information and test information according to a preset rule;

In one possible implementation, the information comparison module 530 is further configured to train the training information through a C4.5 decision tree algorithm, and before the training information is obtained from the decision tree model, the method further includes:

In a possible implementation manner, the construction rule module 540 is further configured to perform iterative processing on a preset minimum support of the association rule based on the ant lion algorithm, so as to obtain a minimum support of the sub-classification alarm information set;

deleting target alarm information with less than the minimum support degree in the sub-classified alarm information set to obtain a sub-classified alarm associated information set; wherein, the target alarm information is any alarm information in the sub-classified alarm information set; wherein the support degree is A is target alarm information, N is the number of all target alarm information, and sup (A) is the support degree of the target alarm information A.

And constructing a relevance rule of the sub-classification alarm relevance information set based on a P-growth algorithm.

Accordingly, the metrics optimization module 550 is also configured to,

In one possible implementation, the information determining module 560 is further configured to determine the front piece of the target relevance rule as the root alert information in the real-time alert information when the confidence level of the target relevance rule is greater than the second confidence level and the imbalance ratio of the target relevance rule is less than the second imbalance ratio; wherein the target association rule is any one of association rules;

Fig. 6 is a schematic diagram of an electronic device according to an embodiment of the present invention. As shown in fig. 6, the electronic device 6 of this embodiment includes: a processor 60, a memory 61 and a computer program 62 stored in said memory 61 and executable on said processor 60. The processor 60, when executing the computer program 62, implements the steps of the fault diagnosis method embodiment of each intelligent substation described above, such as steps 110 to 160 shown in fig. 1. Or the processor 60, when executing the computer program 62, performs the functions of the modules/units of the apparatus embodiments described above, such as the functions of the modules 510-560 shown in fig. 5.

Illustratively, the computer program 62 may be partitioned into one or more modules, which are stored in the memory 61 and executed by the processor 60 to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing the specified functions, which instruction segments describe the execution of the computer program 62 in the electronic device 6. For example, the computer program 62 may be partitioned into modules 510 through 560 shown in FIG. 5.

The electronic device 6 may be a computing device such as a desktop computer, a notebook computer, a palm computer, a cloud server, etc. The electronic device 6 may include, but is not limited to, a processor 60, a memory 61. It will be appreciated by those skilled in the art that fig. 6 is merely an example of the electronic device 6 and is not meant to be limiting as the electronic device 6, may include more or fewer components than shown, or may combine certain components, or different components, e.g., the terminal may further include an input-output device, a network access device, a bus, etc.

The Processor 60 may be a central processing unit (Central Processing Unit, CPU), other general purpose Processor, digital signal Processor (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 61 may be an internal storage unit of the electronic device 6, such as a hard disk or a memory of the electronic device 6. The memory 51 may also be an external storage device of the electronic device 6, such as a plug-in hard disk, a smart memory card (SMART MEDIA CARD, SMC), a Secure Digital (SD) card, a flash memory card (FLASH CARD) or the like, which are provided on the electronic device 6. Further, the memory 61 may also include both an internal storage unit and an external storage device of the electronic device 6. The memory 61 is used for storing the computer program and other programs and data required by the terminal. The memory 61 may also be used for temporarily storing data that has been output or is to be output.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, the specific names of the functional units and modules are only for distinguishing from each other, and are not used for limiting the protection scope of the present invention. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the embodiments provided in the present invention, it should be understood that the disclosed apparatus/terminal and method may be implemented in other manners. For example, the apparatus/terminal embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present invention may implement all or part of the procedures in the above-described embodiments of the method, or may be implemented by instructing related hardware by a computer program, where the computer program may be stored in a computer readable storage medium, and the computer program may implement the steps of the above-described embodiments of the fault diagnosis method of each intelligent substation when executed by a processor. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium may include content that is subject to appropriate increases and decreases as required by jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is not included as electrical carrier signals and telecommunication signals.

The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.

Claims

1. The fault diagnosis method of the intelligent substation is characterized by comprising the following steps of:

comparing all preset classified alarm information sets in a preset C4.5 fault decision tree model with the preprocessed alarm information to obtain sub-classified alarm information sets of the real-time alarm information;

processing preset association indexes of the association rules based on an ant lion algorithm to obtain association indexes of the real-time alarm information;

Obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information;

Wherein, comparing all preset classified alarm information sets in the preset C4.5 fault decision tree model with the preprocessed alarm information to obtain sub-classified alarm information sets of the real-time alarm information, comprising:

Acquiring historical alarm information of a secondary system, wherein the historical alarm information is divided into training information and test information according to preset rules; constructing a fault attribute database of the history alarm information; the history alarm information comprises secondary equipment alarm information, a GOOSE abnormal message and an SV abnormal message; the fault attribute database comprises device locking, self-checking alarm information, GOOSE alarm information, SV alarm information and fault types; the history alarm information is stored in the fault attribute database according to a preset rule; constructing a decision tree model based on the C4.5 decision tree algorithm and the fault attribute database; training the training information through a C4.5 decision tree algorithm to obtain a trained decision tree model; wherein, the C4.5 decision tree algorithm adopts Taylor series to replace logarithmic operation; verifying and pruning the decision tree model by using the test information to obtain the preset C4.5 fault decision tree model; the preset C4.5 fault decision tree model consists of a plurality of preset classified alarm information sets with different fault types; comparing all the preset classified alarm information sets in the preset C4.5 fault decision tree model with the preprocessed alarm information, and determining the preset classified alarm information set with the highest matching degree as a sub-classified alarm information set of the real-time alarm information;

the constructing the relevance rule of the sub-classified alarm information set based on the P-growth algorithm comprises the following steps:

Performing iterative processing on the preset minimum support of the relevance rule based on an ant lion algorithm to obtain the minimum support of the sub-classification alarm information set; deleting target alarm information smaller than the minimum support degree in the sub-classification alarm information set to obtain a sub-classification alarm associated information set; wherein, the target alarm information is any alarm information in the sub-classified alarm information set; constructing a relevance rule of the sub-classification alarm relevance information set based on a P-growth algorithm, wherein the method specifically comprises the steps of carrying out iterative processing on a preset first confidence coefficient and a first unbalance ratio based on an ant lion algorithm to obtain a second confidence coefficient and a second unbalance ratio of the real-time alarm information; wherein the associated metrics include confidence and imbalance ratio;

the obtaining the root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information comprises the following steps:

When the confidence coefficient of the target association rule is larger than the second confidence coefficient and the unbalance ratio of the target association rule is smaller than the second unbalance ratio, determining the front piece of the target association rule as root alarm information in the real-time alarm information; wherein the target association rule is any one of the association rules.

2. The fault diagnosis method of intelligent substation according to claim 1, wherein the information gain rate of the C4.5 decision tree algorithm is

Gainratio is the information Gain ratio of a certain attribute a selected during classification, gain is the information Gain of the attribute a, splitInfo _A is the split information measure of the attribute a in the whole preset classified alarm information set T, T is the whole preset classified alarm information set, T _j is the total data number of the j-th preset classified alarm information set in the whole preset classified alarm information set T, TC _i is the total data number of the sub-classified alarm information set C _i in the whole preset classified alarm information set T, and TC _ij represents the total data number of the j-th preset classified alarm information set in the whole preset classified alarm information set T belonging to the sub-classified alarm information set C _i.

3. The fault diagnosis method of intelligent substation according to claim 1, wherein the support degree isA is target alarm information, N is the number of all target alarm information, and sup (A) is the support degree of the target alarm information A.

4. The fault diagnosis method of intelligent substation according to claim 1, wherein the confidence isA and B are respectively any different target alarm information in the sub-classified alarm information set,/>On the premise that the alarm information of the target A occurs, the probability of occurrence of the alarm information of the target B;

the unbalance ratio is Sup (A) is the support degree of the target alarm information A, sup (B) is the support degree of the target alarm information B, and Sup (A-B) is the support degree of the target alarm information containing both A and B.

5. A fault diagnosis device of an intelligent substation, comprising:

The information comparison module is used for comparing a preset classified alarm information set in a preset C4.5 fault decision tree model with the preprocessed alarm information to obtain a sub-classified alarm information set of the real-time alarm information;

The construction rule module is used for constructing the relevance rule of the sub-classification alarm information set based on a P-growth algorithm;

the index optimization module is used for processing preset association indexes of the association rules based on an ant lion algorithm to obtain the association indexes of the real-time alarm information;

The information determining module is used for obtaining root alarm information in the real-time alarm information according to the association index and the association rule of the real-time alarm information;

The information comparison module is used for acquiring historical alarm information of the secondary system, wherein the historical alarm information is divided into training information and test information according to a preset rule; constructing a fault attribute database of the history alarm information; the history alarm information comprises secondary equipment alarm information, a GOOSE abnormal message and an SV abnormal message; the fault attribute database comprises device locking, self-checking alarm information, GOOSE alarm information, SV alarm information and fault types; the history alarm information is stored in the fault attribute database according to a preset rule; constructing a decision tree model based on the C4.5 decision tree algorithm and the fault attribute database; training the training information through a C4.5 decision tree algorithm to obtain a trained decision tree model; wherein, the C4.5 decision tree algorithm adopts Taylor series to replace logarithmic operation; verifying and pruning the decision tree model by using the test information to obtain the preset C4.5 fault decision tree model; the preset C4.5 fault decision tree model consists of a plurality of preset classified alarm information sets with different fault types; comparing all the preset classified alarm information sets in the preset C4.5 fault decision tree model with the preprocessed alarm information, and determining the preset classified alarm information set with the highest matching degree as a sub-classified alarm information set of the real-time alarm information;

The construction rule module is used for carrying out iterative processing on the preset minimum support of the relevance rule based on the ant lion algorithm to obtain the minimum support of the sub-classification alarm information set; deleting target alarm information smaller than the minimum support degree in the sub-classification alarm information set to obtain a sub-classification alarm associated information set; wherein, the target alarm information is any alarm information in the sub-classified alarm information set; constructing a relevance rule of the sub-classification alarm relevance information set based on a P-growth algorithm, wherein the method specifically comprises the steps of carrying out iterative processing on a preset first confidence coefficient and a first unbalance ratio based on an ant lion algorithm to obtain a second confidence coefficient and a second unbalance ratio of the real-time alarm information; wherein the associated metrics include confidence and imbalance ratio;

The information determining module is used for determining the front piece of the target relevance rule as root alarm information in the real-time alarm information when the confidence coefficient of the target relevance rule is larger than the second confidence coefficient and the unbalance ratio of the target relevance rule is smaller than the second unbalance ratio; wherein the target association rule is any one of the association rules.

6. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 4 when the computer program is executed.

7. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 4.