Electric power information system operation stability analysis method and system

Abstract

The invention discloses an electric power information system operation stability analysis method. The method comprises the steps that historical data information of a target electric power information system is acquired and preprocessed to obtain a training data set; constructing an electric power information system operation stability analysis primary model based on a BP neural network, and training to obtain an electric power information system operation stability analysis model; operation data information of the target electric power information system is obtained in real time, and operation stability analysis is conducted through the electric power information system operation stability analysis model. The invention also discloses a system for realizing the method for analyzing the operation stability of the electric power information system. According to the method, the historical data of the target electric power information system is combined and analyzed to extract the data with high correlation, and the stability analysis system is constructed and trained based on the BP neural network, so that the operation stability analysis of the electric power information system can be realized, the reliability is higher, the accuracy is better, and the timeliness is better.

Description

Technical Field

[0001] The present invention belongs to the field of electrical automation, and in particular relates to a method and system for analyzing the operation stability of an electric power information system. Background Art

[0002] With the development of economy and technology and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and life, bringing endless convenience to people's production and life. Therefore, ensuring the stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] At present, with the widespread adoption of power information systems, the safe and stable operation of power information systems is of vital importance to power systems. During the operation of power information systems, due to the existence of hidden dangers such as program and architecture design defects and unreasonable configuration of software and hardware resources, external conditions such as a surge in business access can easily lead to system operation failures, resulting in data loss or business interruption and other adverse effects. Therefore, it is of great significance to analyze the operational stability of power information systems in a timely and accurate manner.

[0004] At present, the operation stability analysis scheme adopted by the power information system is mainly implemented by monitoring the performance indicators of the software and hardware deployed in the system, threshold alarms, manual inspections, etc. However, such existing schemes are often based on data under the condition that the monitoring indicators have reached or exceeded the alarm threshold, that is, the data under the condition that the system fault has been or is infinitely close to being triggered; in this case, the analysis and disposal based on the monitored data are poor in effect, and there is a significant lag in the discovery and location of system faults. Summary of the invention

[0005] One of the purposes of the present invention is to provide a method for analyzing the operation stability of an electric power information system with high reliability, good accuracy and good timeliness.

[0006] A second objective of the present invention is to provide a system for implementing the method for analyzing the operation stability of the power information system.

[0007] The method for analyzing the operation stability of a power information system provided by the present invention comprises the following steps:

[0008] S1. Obtain historical data information of the target power information system;

[0009] S2. Preprocess the data information obtained in step S1 to obtain a training data set;

[0010] S3. Based on BP neural network, construct a primary model for power information system operation stability analysis;

[0011] S4. Using the training data set obtained in step S2, the primary model for power information system operation stability analysis constructed in step S3 is trained to obtain a power information system operation stability analysis model;

[0012] S5. Acquire the operation data information of the target power information system in real time, and use the power information system operation stability analysis model obtained in step S4 to perform operation stability analysis on the target power information system.

[0013] The acquisition of historical data information of the target power information system described in step S1 specifically includes the following steps:

[0014] Obtain historical data information of the target power information system;

[0015] The data information includes business access data information, operation status data information and external weather data information.

[0016] The business access volume data information specifically includes the number of active users accessing within a set time interval, the page access frequency and the data growth; through bypass monitoring and network mirror backup analysis, the core switch network mirror backup of the network where the target power information system is located is diverted by the mirror diversion device according to the IP address, and the user network access information is collected; the user network access information includes request messages, response messages, IP information, response time and request URL address; after data analysis, service analysis and processing, the three indicator data of the real-time active user number, page access frequency and average service response time of the target power information system are generated at a fixed frequency; at the target power information system database end, several core business data tables are selected, the data volume growth in the database is queried and counted, and the indicator results of business data growth are obtained;

[0017] The operation status data information specifically includes the operation status indicators of the application server and the database server in the target power information system; at the same time, the operation status indicators of the application server and the database backend in the target power information system are obtained at the same frequency as the operation status indicators of the application server and the database server; the operation status indicators include CPU usage, memory usage, call success rate of important interfaces, average database response time, number of slow SQL statements in the database, and average service response time;

[0018] The external weather data information specifically includes external weather information and external temperature information.

[0019] The preprocessing of the data information obtained in step S1 described in step S2 specifically includes the following steps:

[0020] Verify and correct data;

[0021] One-hot encoding is used to encode time, week, month, weather and temperature data; among them, time is encoded as a 24-dimensional vector to correspond to 0:00-23:00, week is encoded as a 7-dimensional vector to correspond to Monday-Sunday; month is encoded as a 12-dimensional vector to correspond to January-December; weather is encoded as a 5-dimensional vector to correspond to sunny, thunderstorm, heavy snow, hail and other weather; temperature is encoded as a 7-dimensional vector to correspond to below zero, 0-10℃, 10-20℃, 20-30℃, 30-35℃, 35-40℃ and above 40℃; after encoding, time, week, month, weather and temperature data are converted into a feature vector set of high-dimensional sparse structure;

[0022] The following formula is used to calculate the correlation between the vectors in the feature vector set and the service access volume data information and the operation status data information:

[0023]

[0024] Where ρ xy is the correlation between x and y, where x is a vector in the feature vector set, and y is any indicator in the service access volume data information and the operation status data information; μ x is the expected value of x; μ y is the expected value of y; σ x is the variance of x; σ y is the variance of y; E((x-μ x )(y-μ y )) is the deviation of x from μ x The value of y deviates from μ y The expected value of the product of the values ​​of ;

[0025] According to the obtained correlation data, vectors in the feature vector set whose correlation is greater than a set value are selected as candidate training data;

[0026] The following formula is used to normalize the business access data and operation status data:

[0027]

[0028] Where X norm is the data after normalization; X is the data before normalization; X min is the minimum value in the corresponding sample data set before normalization; X max is the maximum value in the corresponding sample data set before normalization;

[0029] The normalized service access volume data information and operation status data information are used as candidate training data;

[0030] Finally, all candidate training data are aggregated to construct a training dataset.

[0031] The step S3 described in which a primary model for analyzing the operation stability of the power information system is constructed based on the BP neural network specifically includes the following steps:

[0032] The constructed primary model for the operation stability analysis of the power information system includes input layer, hidden layer and output layer;

[0033] Merge the data of the constructed training data set to obtain the input vector, which is used as the output of the input layer;

[0034] The hidden layer constructed is a BP neural network with 3 single hidden layers;

[0035] In the output layer, three output nodes are set, corresponding to the stable state, fault warning state and fault state of the power information system respectively.

[0036] The construction of the primary model for analyzing the operation stability of the power information system also includes the following steps:

[0037] The number of nodes N in the hidden layer is calculated using the following formula: h :

[0038]

[0039] Where N s is the sample size; N i is the number of nodes in the input layer; N o is the number of nodes in the output layer; α is a set integer variable;

[0040] When calculating, the value of α is set from the minimum integer value α min Start, and gradually increase to the set maximum integer value α max , and record the number of nodes N' of all hidden layers obtained during the increase of α h ; Finally, select the number of nodes N' in the hidden layer corresponding to the highest prediction accuracy of the model h , as the final hidden layer node number N h .

[0041] The construction of the primary model for analyzing the operation stability of the power information system also includes the following steps:

[0042] In the constructed model, the sigmoid function is used as the activation function;

[0043] The following formula is used as the hidden layer output function:

[0044] O j =f(∑W ij ×Xi -q j )

[0045] Where O j is the output function value of the hidden layer; W ij is the weight coefficient connecting the i-th input node and the j-th hidden node; X i is the input vector of the i-th input layer node; q j is the jth hidden node threshold; f() is the activation function;

[0046] The following formula is used as the output function of the output layer:

[0047] Y k =f(T jk ×O j -q k )

[0048] Where Y k is the output function value of the output layer; T jk is the weight coefficient connecting the jth input node and the kth output node; j is the input of the jth node in the hidden layer; q k is the threshold of the kth output node; f() is the activation function.

[0049] The training described in step S4 specifically includes the following steps:

[0050] The following formula is used as the loss function:

[0051]

[0052] Where L is the loss function value; t pi is the expected output value of node i; pi is the calculated output value of the i node;

[0053] Use the following formula to calculate W ij To make corrections:

[0054] ΔW ij (nn+1)=h×φ i ×O j +a×ΔW ij (nn)

[0055] where ΔW ij (nn+1) is W at the nn+1th iteration ij The correction value of; h is the set learning factor; φ i is the calculation error of output node i; j is the calculated output of output node j; a is the set momentum factor; ΔW ij(nn) is the nnth iteration W ij Correction value.

[0056] The present invention also provides a system for implementing the method for analyzing the operation stability of the power information system, comprising a data acquisition module, a data processing module, a model construction module, a model training module and a stability analysis module; the data acquisition module, the data processing module, the model construction module, the model training module and the stability analysis module are connected in series in sequence; the data acquisition module is used to acquire historical data information of the target power information system, and upload the data information to the data processing module; the data processing module is used to pre-process the acquired data information according to the received data information to obtain a training data set, and upload the data information to the model construction module; the model construction module is used to construct a primary model for the operation stability analysis of the power information system based on the BP neural network according to the received data information, and upload the data information to the model training module; the model training module is used to train the constructed primary model for the operation stability analysis of the power information system according to the received data information, using the obtained training data set, to obtain the operation stability analysis model of the power information system, and upload the data information to the stability analysis module; the stability analysis module is used to acquire the operation data information of the target power information system in real time according to the received data information, and use the obtained operation stability analysis model of the power information system to perform operation stability analysis.

[0057] The method and system for analyzing the operational stability of an electric power information system provided by the present invention combines and analyzes the historical data of a target electric power information system to extract highly relevant data, and constructs and trains a stability analysis system based on a BP neural network. Therefore, the present invention can not only realize the operational stability analysis of an electric power information system, but also has higher reliability, better accuracy, and better timeliness. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 The figure is a schematic diagram of the method flow of the present invention.

[0059] Figure 2 Schematic diagram of the functional modules of the system of the present invention. DETAILED DESCRIPTION

[0060] like Figure 1 The method flow chart of the method of the present invention is shown as follows: The method for analyzing the operation stability of the power information system disclosed in the present invention comprises the following steps:

[0061] S1. Obtaining historical data information of the target power information system; specifically comprising the following steps:

[0062] Obtain historical data information of the target power information system;

[0063] The data information includes business access data information, operation data information and external weather data information;

[0064] Among them, the business access volume data information specifically includes the number of active users accessing within a set time interval, the page access frequency and the data growth; through bypass monitoring and network mirror backup analysis, the core switch network mirror backup of the network where the target power information system is located is diverted by the mirror diversion device according to the IP address, and the user network access information is collected; the user network access information includes request messages, response messages, IP information, response time and request URL address; after data analysis, service analysis and processing, the three indicator data of the real-time active user number, page access frequency and average service response time of the target power information system are generated at a fixed frequency; at the target power information system database end, several core business data tables are selected, the data volume growth in the database is queried and counted, and the indicator results of business data growth are obtained; in specific implementation, it is preferred to obtain the above data at a frequency of 5min / time;

[0065] The operation status data information specifically includes the operation status indicators of the application server and the database server in the target power information system; at the same time, the operation status indicators of the application server and the database backend in the target power information system are obtained at the same frequency as the operation status indicators of the application server and the database server; the operation status indicators include CPU usage, memory usage, call success rate of important interfaces, average database response time, number of slow SQLs in the database and average service response time; in specific implementation, the above data is preferably obtained at a frequency of 5 minutes / time;

[0066] The external weather data information specifically includes external weather information and external temperature information;

[0067] S2. Preprocess the data information obtained in step S1 to obtain a training data set; specifically, the steps include:

[0068] Verify and correct the data; during specific implementation, perform data quality verification, identify and mark missing data and abnormal fluctuation data, and supplement and correct abnormal data by system interpolation or manual supplementation;

[0069] One-hot encoding is used to encode time, week, month, weather and temperature data; among them, time is encoded as a 24-dimensional vector to correspond to 0:00-23:00, week is encoded as a 7-dimensional vector to correspond to Monday-Sunday; month is encoded as a 12-dimensional vector to correspond to January-December; weather is encoded as a 5-dimensional vector to correspond to sunny, thunderstorm, heavy snow, hail and other weather; temperature is encoded as a 7-dimensional vector to correspond to below zero, 0-10℃, 10-20℃, 20-30℃, 30-35℃, 35-40℃ and above 40℃; after encoding, time, week, month, weather and temperature data are converted into a feature vector set of high-dimensional sparse structure;

[0070] The following formula is used to calculate the correlation between the vectors in the feature vector set and the service access volume data information and the operation status data information:

[0071]

[0072] Where ρ xy is the correlation between x and y, where x is a vector in the feature vector set, and y is any indicator in the service access volume data information and the operation status data information; μ x is the expected value of x, and μ y is the expected value of y, and σ x is the variance of x, and σ y is the variance of y, and E((x-μ x )(y-μ y )) is the deviation of x from μ x The value of y deviates from μ y The expected value of the product of the values ​​of ;

[0073] According to the obtained correlation data, vectors in the feature vector set whose correlation is greater than a set value (preferably 0.5) are selected as candidate training data;

[0074] The following formula is used to normalize the business access data and operation status data:

[0075]

[0076] Where X norm is the data after normalization; X is the data before normalization; X min is the minimum value in the corresponding sample data set before normalization; X max is the maximum value in the corresponding sample data set before normalization;

[0077] The normalized service access volume data information and operation status data information are used as candidate training data;

[0078] The detected interruption alarm of the healthy operation time indicator of the target power information system (the alarm state indicates that the system has page access requests that are unreachable and database connections that are interrupted) is used as a unified standard for judging whether the system has failed. The samples are classified according to the time range of the alarm and the sample collection time. Among them, the samples collected in the alarm state are fault state samples, the samples collected within 1 hour before the alarm occurs are fault warning state samples, and the others are stable state samples; samples are selected from the stable state, fault warning state, and fault state samples to form a training sample set, and all samples are classified and marked;

[0079] Finally, all candidate training data are aggregated to construct a training dataset;

[0080] S3. Based on the BP neural network, a primary model for analyzing the operation stability of the power information system is constructed; the specific steps include:

[0081] The constructed primary model for the operation stability analysis of the power information system includes input layer, hidden layer and output layer;

[0082] Merge the data of the constructed training data set to obtain the input vector, which is used as the output of the input layer;

[0083] The hidden layer constructed is a BP neural network with 3 single hidden layers;

[0084] In the output layer, three output nodes are set, corresponding to the stable state, fault warning state and fault state of the power information system;

[0085] After the input vector enters the BP neural network architecture through the input layer nodes, it first undergoes preliminary calculation processing through the hidden layer nodes, and then undergoes further calculation and adjustment in the output layer to complete the data forward propagation process in the evaluation model. Then, the error feedback back propagation from the output layer to the input layer is used to complete the parameter adjustment to fit the curve to achieve the expected effect.

[0086] The number of hidden layers determines the quality of the nonlinear mapping between the input and output of the neural network. Theoretically, when the two-layer BP network has enough hidden layer nodes, any nonlinear mapping of input and output can be realized. However, in practical applications, it is found that when the number of hidden layers is too large, it will lead to overfitting of the model, increase the difficulty of training, and make it difficult for the model to converge. Therefore, in the specific implementation, the number of nodes N in the hidden layer is calculated using the following formula: h :

[0087]

[0088] Where N s is the sample size; N i is the number of nodes in the input layer; N o is the number of nodes in the output layer; α is a set integer variable;

[0089] When calculating, the value of α is set from the minimum integer value α min (preferably 2), and gradually increase to the set maximum integer value α max (The preferred value is 10), and record the number of nodes N' of all hidden layers obtained during the increase of α h ; Finally, select the number of nodes N' in the hidden layer corresponding to the highest prediction accuracy of the model h , as the final hidden layer node number N h ;

[0090] At the same time, in the constructed model, the sigmoid function is used as the activation function;

[0091] The following formula is used as the hidden layer output function:

[0092] O j =f(ΣW ij ×X i -q j )

[0093] Where O j is the output function value of the hidden layer; W ij is the weight coefficient connecting the i-th input node and the j-th hidden node; X i is the input vector of the i-th input layer node; q j is the jth hidden node threshold; f() is the activation function;

[0094] The following formula is used as the output function of the output layer:

[0095] Y k =f(T jk ×O j -q k )

[0096] Where Y k is the output function value of the output layer; T jk is the weight coefficient connecting the jth input node and the kth output node; j is the input of the jth node in the hidden layer; q k is the threshold of the kth output node; f() is the activation function;

[0097] S4. Using the training data set obtained in step S2, the primary model for power information system operation stability analysis constructed in step S3 is trained to obtain a power information system operation stability analysis model;

[0098] During training, the following formula is used as the loss function:

[0099]

[0100] Where L is the loss function value; t pi is the expected output value of node i; pi is the calculated output value of the i node;

[0101] Use the following formula to calculate W ij To make corrections:

[0102] ΔW ij (nn+1)=h×φ i ×O j +a×ΔW ij (nn)

[0103] where ΔW ij (nn+1) is W at the nn+1th iteration ij The correction value of; h is the set learning factor; φ i is the calculation error of output node i; j is the calculated output of output node j; a is the set momentum factor; ΔW ij (nn) is the nnth iteration W ij Correction value of

[0104] S5. Acquire the operation data information of the target power information system in real time, and use the power information system operation stability analysis model obtained in step S4 to perform operation stability analysis on the target power information system.

[0105] like Figure 2The functional module schematic diagram of the system of the present invention is shown as follows: the system for realizing the operation stability analysis method of the power information system disclosed in the present invention comprises a data acquisition module, a data processing module, a model construction module, a model training module and a stability analysis module; the data acquisition module, the data processing module, the model construction module, the model training module and the stability analysis module are connected in series in sequence; the data acquisition module is used to acquire the historical data information of the target power information system, and upload the data information to the data processing module; the data processing module is used to pre-process the acquired data information according to the received data information to obtain a training data set, and upload the data information to the model construction module; the model construction module is used to construct a primary model for operation stability analysis of the power information system based on the BP neural network according to the received data information, and upload the data information to the model training module; the model training module is used to train the constructed primary model for operation stability analysis of the power information system according to the received data information, using the obtained training data set, to obtain the operation stability analysis model of the power information system, and upload the data information to the stability analysis module; the stability analysis module is used to acquire the operation data information of the target power information system in real time according to the received data information, and use the obtained operation stability analysis model of the power information system to perform operation stability analysis.

Claims

1. A method for analyzing the operation stability of a power information system, comprising the following steps: S1. Obtain historical data information of the target power information system; S2. Preprocess the data information obtained in step S1 to obtain a training data set; S3. Based on BP neural network, construct a primary model for power information system operation stability analysis; S4. Using the training data set obtained in step S2, the primary model for power information system operation stability analysis constructed in step S3 is trained to obtain a power information system operation stability analysis model; S5. Acquire the operation data information of the target power information system in real time, and use the power information system operation stability analysis model obtained in step S4 to perform operation stability analysis on the target power information system.

2. The method for analyzing the operation stability of an electric power information system according to claim 1, characterized in that The acquisition of historical data information of the target power information system described in step S1 specifically includes the following steps: Obtain historical data information of the target power information system; The data information includes business access data information, operation status data information and external weather data information.

3. The method for analyzing the operation stability of an electric power information system according to claim 2 is characterized in that The business access data information specifically includes the number of active users, page access frequency and data growth within a set time interval; By means of bypass monitoring and network mirror backup analysis, the network mirror backup of the core switch of the network where the target power information system is located is diverted by the mirror diversion device according to the IP address, and the user network access information is collected; the user network access information includes request messages, response messages, IP information, response time and request URL address; after data analysis, service analysis and processing, the three indicator data of the real-time active users, page access frequency and average service response time of the target power information system are generated at a fixed frequency; On the target power information system database side, select several core business data tables, query and count the data volume growth in the database, and obtain the indicator results of business data growth; The operation status data information specifically includes the operation status indicators of the application server and the database server in the target power information system; at the same time, the operation status indicators of the application server and the database server backend in the target power information system are obtained at the same frequency as the operation status indicators of the application server and the database server; The operating indicators include CPU usage, memory usage, call success rate of important interfaces, average database response time, number of slow SQL statements in the database, and average service response time; The external weather data information specifically includes external weather information and external temperature information.

4. The method for analyzing the operation stability of an electric power information system according to claim 3 is characterized in that The preprocessing of the data information obtained in step S1 described in step S2 specifically includes the following steps: Verify and correct data; One-hot encoding is used to encode time, week, month, weather and temperature data; among them, time is encoded as a 24-dimensional vector to correspond to 0:00-23:00, week is encoded as a 7-dimensional vector to correspond to Monday-Sunday; month is encoded as a 12-dimensional vector to correspond to January-December; weather is encoded as a 5-dimensional vector to correspond to sunny, thunderstorm, heavy snow, hail and other weather; temperature is encoded as a 7-dimensional vector to correspond to below zero, 0-10℃, 10-20℃, 20-30℃, 30-35℃, 35-40℃ and above 40℃; after encoding, time, week, month, weather and temperature data are converted into a feature vector set of high-dimensional sparse structure; The following formula is used to calculate the correlation between the vectors in the feature vector set and the service access volume data information and the operation status data information: Where ρ xy is the correlation between x and y, where x is a vector in the feature vector set, and y is any indicator in the service access volume data information and the operation status data information; μ x is the expected value of x; μ y is the expected value of y; σ x is the variance of x; σ y is the variance of y; E((x-μ x )(y-μ y )) is the deviation of x from μ x The value of y deviates from μ y The expected value of the product of the values ​​of ; According to the obtained correlation data, vectors in the feature vector set whose correlation is greater than a set value are selected as candidate training data; The following formula is used to normalize the business access data and operation status data: Where X norm is the data after normalization; X is the data before normalization; X min is the minimum value in the corresponding sample data set before normalization; X max is the maximum value in the corresponding sample data set before normalization; The normalized service access volume data information and operation status data information are used as candidate training data; Finally, all candidate training data are aggregated to construct a training dataset.

5. The method for analyzing the operation stability of an electric power information system according to claim 4, characterized in that The step S3 described in which a primary model for analyzing the operation stability of the power information system is constructed based on the BP neural network specifically includes the following steps: The constructed primary model for the operation stability analysis of the power information system includes input layer, hidden layer and output layer; Merge the data of the constructed training data set to obtain the input vector, which is used as the output of the input layer; The hidden layer constructed is a BP neural network with 3 single hidden layers; In the output layer, three output nodes are set, corresponding to the stable state, fault warning state and fault state of the power information system respectively.

6. The method for analyzing the operation stability of an electric power information system according to claim 5, characterized in that The construction of the primary model for analyzing the operation stability of the power information system also includes the following steps: The number of nodes N in the hidden layer is calculated using the following formula: h : Where N s is the sample size; N i is the number of nodes in the input layer; N o is the number of nodes in the output layer; α is a set integer variable; When calculating, the value of α is set from the minimum integer value α min Start, and gradually increase to the set maximum integer value α max , and record the number of nodes N' of all hidden layers obtained during the increase of α h ; Finally, select the number of nodes N' in the hidden layer corresponding to the highest prediction accuracy of the model h , as the final hidden layer node number N h .

7. The method for analyzing the operation stability of an electric power information system according to claim 6, characterized in that The construction of the primary model for analyzing the operation stability of the power information system also includes the following steps: In the constructed model, the sigmoid function is used as the activation function; The following formula is used as the hidden layer output function: O j =f(∑W ij ×X i -q j ) Where O j is the output function value of the hidden layer; W ij is the weight coefficient connecting the i-th input node and the j-th hidden node; X i is the input vector of the i-th input layer node; q j is the jth hidden node threshold; f() is the activation function; The following formula is used as the output function of the output layer: Y k =f(T jk ×O j -q k ) Where Y k is the output function value of the output layer; T jk is the weight coefficient connecting the jth input node and the kth output node; j is the input of the jth node in the hidden layer; q k is the threshold of the kth output node; f() is the activation function.

8. The method for analyzing the operation stability of an electric power information system according to claim 7, characterized in that The training described in step S4 specifically includes the following steps: The following formula is used as the loss function: Where L is the loss function value; t pi is the expected output value of node i; pi is the calculated output value of the i node; Use the following formula to calculate W ij To make corrections: ΔW ij (nn+1)=h×φ i ×O j +a×ΔW ij (nn) where ΔW ij (nn+1) is W at the nn+1th iteration ij The correction value of ; h is the set learning factor; φ i is the calculation error of output node i; j is the calculated output of output node j; a is the set momentum factor; ΔW ij (nn) is the nnth iteration W ij Correction value.

9. A system for implementing the method for analyzing the operation stability of a power information system according to any one of claims 1 to 8, characterized in that It includes a data acquisition module, a data processing module, a model building module, a model training module and a stability analysis module; the data acquisition module, the data processing module, the model building module, the model training module and the stability analysis module are connected in series in sequence; the data acquisition module is used to obtain the historical data information of the target power information system and upload the data information to the data processing module; The data processing module is used to pre-process the acquired data information according to the received data information to obtain a training data set, and upload the data information to the model construction module; The model building module is used to build a primary model for analyzing the operation stability of the power information system based on the BP neural network according to the received data information, and upload the data information to the model training module; The model training module is used to train the constructed primary model for power information system operation stability analysis using the obtained training data set according to the received data information, obtain the power information system operation stability analysis model, and upload the data information to the stability analysis module; The stability analysis module is used to obtain the operation data information of the target power information system in real time according to the received data information, and use the obtained power information system operation stability analysis model to perform operation stability analysis.

Images