Substation equipment state discrimination and exception handling intelligent decision-making method

Abstract

The invention discloses a substation equipment state discrimination and exception handling intelligent decision-making method, which comprises the steps of establishing an equipment state management and control intelligent decision-making library based on historical data and expert experience, processing and analyzing equipment detection parameters based on an LSTM-CNN technology, and outputting an equipment state active early warning classification result. Equipment operation and maintenance, overhaul and test strategies are automatically put forward according to the type, content and level of active early warning of the equipment state, when an early warning signal occurs to the power transformation equipment, the system carries out intelligent decision analysis and judgment through the equipment state management and control decision library, and if equipment overhaul is judged, an overhaul strategy is automatically pushed; if the inspection is judged, pushing an operation and maintenance strategy; if the equipment test is needed, a test strategy is automatically pushed; the method can improve the operation and maintenance quality, improve the power supply reliability, improve the operation level of the transformer substation, improve the operation stability of the transformer substation, reduce the power-off maintenance times of the transformer substation, and improve the satisfaction degree of power utilization service.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent monitoring of substation equipment, and in particular relates to an intelligent decision-making method for state identification and abnormality processing of substation equipment. Background Art

[0002] With the continuous development of my country's economy and society, the demand for electricity in people's production and life is also rising rapidly. Against this background, the speed of my country's power industry construction has also been improved. Due to the imbalance of my country's current energy resources and economic development, the State Grid Corporation of China has proposed the important mission of building a smart grid to promote people's production and life. Smart grid is an intelligent power grid, which is built on the basis of an integrated high-speed two-way communication network combined with advanced technology and control methods. In order to meet the current public demand for power quality, the construction of smart grid is also imperative. According to the smart grid construction plan of the State Grid Corporation of China, the construction of smart grid has entered the third stage. In this stage, a unified strong smart grid will be built in an all-round way, and the technology and equipment will reach the international advanced level.

[0003] At present, the monitoring systems of the main equipment and various auxiliary equipment of substations are independent, the interfaces are scattered, the equipment coverage is incomplete, the information transmission is incomplete, the information of the main and auxiliary equipment cannot be effectively shared, and the substation operation and maintenance personnel lack effective monitoring means for the equipment under their jurisdiction; the equipment operation data obtained through various monitoring methods cannot be effectively utilized, and accurate diagnosis and active early warning of the equipment health status cannot be achieved; a large amount of work relies on manual control, and when the equipment produces abnormal conditions or failures, fault judgment and processing mainly rely on human experience, with a long response time, low processing efficiency and poor results; the collection and entry of operation and inspection data relies on manual completion, with a high error rate and the problem of repeated entry.

[0004] At present, the State Grid's mature systems for equipment status identification are mainly concentrated in online monitoring and fault diagnosis. The more representative online monitoring systems for power equipment include XJ's CBS-2000 online monitoring and analysis system for power equipment and GD NS-8000 online monitoring and fault diagnosis system for substation equipment status. Although these systems have realized the networked and intelligent comprehensive monitoring of common power equipment, they are very lacking in decision-making processing and are far from meeting the requirements for realizing the self-decision-making function of smart grids.

[0005] Typical systems abroad include Canada's AIM system, the PM (predictive maintenance) system developed by Entech in the United States, the substation maintenance management system developed by Finland's IVO Transmission Services, and the Main man maintenance management system and CMMS-IAMS (intelligent control-maintenance-management integrated system) developed in the Netherlands. However, these systems also mainly focus on power equipment fault diagnosis.

[0006] With the full implementation of integrated control, the field of power grid control has entered the era of big data. Under the background of big data, the data sources for equipment status judgment are more diverse and the data scale is larger. The processing and analysis results often require multi-faceted confirmation by operation and maintenance personnel. There are also problems of false alarms, wrong reports, and non-reports. The actual application situation that is finally presented is to rely on the experience of operation and maintenance personnel to deal with problems, and it is inevitably difficult to prepare scientific judgment results.

[0007] In order to solve the above problems, it is necessary to develop an intelligent decision-making method for substation equipment status identification and abnormality processing. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the prior art and to provide a method for intelligent decision-making on the status of substation equipment and abnormality processing, which can improve the quality of operation and inspection, improve the reliability of power supply, enhance the operation level of substations, improve the stability of substation operation, reduce the number of power outages for maintenance in substations, and improve the satisfaction with electricity services.

[0009] The object of the present invention is achieved by: a method for intelligent decision-making on the state identification and abnormality handling of substation equipment, comprising the following steps: S1, based on historical data and expert experience, build an intelligent decision-making library for substation equipment status management and control, including the category, content and level of active warning of equipment status and the corresponding intelligent decision-making results; S2, real-time detection of substation equipment status through the equipment monitoring system, and sending the real-time detection parameters of the equipment status to the remote monitoring center through wireless communication technology; S3, in the remote monitoring center, based on LSTM-CNN technology, the real-time status detection parameters of substation equipment are processed and analyzed. First, the various real-time detection signal data embedding layers are converted into vectors, and then input into LSTM for feature extraction. The output of LSTM is then used as the input of CNN for further feature extraction. Finally, the real-time classification results of active warning of equipment status are obtained, including category, content and level; S4, compare and analyze the obtained real-time classification results of the active warning of equipment status with the intelligent decision-making library of substation equipment status control for intelligent decision-making. If it is determined to be equipment maintenance, the maintenance strategy is automatically pushed; if it is determined to be inspection, the operation and maintenance strategy is automatically pushed; if equipment testing is required, the test strategy is automatically pushed; S5, in the remote monitoring center, the intelligent decision results are displayed, including the judgment results and their automatic push strategies, and the intelligent decision results are automatically sent to the relevant responsible persons.

[0010] Preferably, the step S1 includes the following steps: S11, based on historical data and LSTM-CNN technology, processes and analyzes the historical status parameters of substation equipment. First, various historical signal data embedding layers are converted into vectors, and then input into LSTM for feature extraction. The output of LSTM is then used as the input of CINN for further feature extraction. Finally, the classification results of active warning signals of equipment status are obtained, including categories, contents and levels and their corresponding intelligent decision-making results. S12, based on expert experience, supplement and correct the classification result of the active warning signal of the equipment status obtained in step S11 to make it closer to actual operation experience; S13, based on the new historical data accumulated during the continued operation of the equipment, repeat steps S11 and S12, optimize and continuously improve the intelligent decision-making library for substation equipment status management and control, and gradually weaken the expert experience to make it closer to the equipment's own status operation characteristics.

[0011] Preferably, in step S11, before processing and analyzing the historical status parameters of the substation equipment, the historical equipment status operating parameters are first formed into time series data in chronological order.

[0012] Preferably, in step S3, for the LSTM network part, the depth of the model is increased by adding LSTM network units to improve the prediction ability of the model, a random deactivation method is used between each LSTM network layer to prevent the model from overfitting, and a vector of a specified format is output in the fully connected layer (Dense).

[0013] Preferably, in step S4, the real-time classification results of the active warning of the equipment status and the intelligent decision-making comparison analysis and judgment of the intelligent decision-making library of the substation equipment status management and control are completed through automatic scanning.

[0014] Due to the adoption of the above technical solution, the beneficial effects of the present invention are: (1) The present invention establishes an intelligent decision-making database for equipment status management and control, and automatically proposes equipment operation and maintenance, inspection and testing strategies based on the category, content and level of active warning of equipment status, thereby reducing the occurrence rate of failures and improving resource utilization efficiency. When an actual failure occurs, the location of the failure can be determined in a timely and effective manner, and the cause and solution measures can be proposed, thereby greatly reducing economic losses and achieving effective utilization of resources and sustainable development. (2) The present invention combines the hybrid model short-term warning method of LSTM and CNN, first converts the various signal data embedding layers into vectors, then inputs them into LSTM for feature extraction, and then uses the output of LSTM as the input of CNN for further feature extraction, and finally obtains the classification result, which can effectively improve the active warning accuracy of substation equipment; (3) The present invention constructs and applies a knowledge base for identifying the state of substation equipment and handling abnormalities, thereby achieving active early warning of abnormal equipment and improving the accuracy and scientificity of intelligent decision-making, thereby assisting operation and maintenance personnel to timely improve the operating status of equipment, timely control and eliminate defects, quickly and accurately judge and handle equipment failures, and shorten response time; In general, the present invention can improve the quality of operation and inspection, improve the reliability of power supply, and enhance the operation level of substations, thus adding impetus to the construction of smart substations of State Grid Corporation. At the same time, the present invention can improve the stability of substation operation, reduce the number of power outages and maintenance of substations, and improve the satisfaction of electricity services. It provides convenience for operation and maintenance work, comprehensively enhances the technical support capability of smart substations for the large-scale operation and maintenance system of State Grid Corporation, effectively supports the construction of my country's ultra-high voltage interconnected power grid, enhances the international core competitiveness of my country's electrical equipment manufacturing enterprises, and promotes the development of smart grids. It has significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is the intelligent decision-making flow chart of equipment status control of the present invention.

[0016] Figure 2 It is a structural diagram of the LSTM-CNN model of the present invention.

[0017] Figure 3 It is an active early warning flow chart of a specific embodiment (circuit breaker) of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0019] like Figure 1 and Figure 2 As shown, the present invention provides a method for intelligent decision-making on substation equipment status identification and abnormality processing, comprising the following steps: S1, based on historical data and expert experience, builds an intelligent decision-making library for substation equipment status management and control, including the category, content and level of active warning of equipment status and the corresponding intelligent decision-making results.

[0020] The construction of the intelligent decision-making library for substation equipment status control includes the following steps: S11, based on historical data and LSTM-CNN technology, processes and analyzes the historical status parameters of substation equipment. First, various historical signal data embedding layers are converted into vectors, and then input into LSTM for feature extraction. The output of LSTM is then used as the input of CINN for further feature extraction. Finally, the classification results of active warning signals of equipment status are obtained, including categories, contents and levels and their corresponding intelligent decision-making results. S12, based on expert experience, supplement and correct the classification result of the active warning signal of the equipment status obtained in step S11 to make it closer to actual operation experience; S13, based on the new historical data accumulated during the continued operation of the equipment, repeat steps S11 and S12, optimize and continuously improve the intelligent decision-making library for substation equipment status management and control, and gradually weaken the expert experience to make it closer to the equipment's own status operation characteristics.

[0021] Before processing and analyzing the historical status parameters of the substation equipment, the historical equipment status operating parameters are first formed into time series data in chronological order.

[0022] S2, real-time detection of substation equipment status through the equipment monitoring system, and sending the real-time detection parameters of the equipment status to the remote monitoring center through wireless communication technology.

[0023] In specific implementation, due to the different characteristics of the equipment itself, it is necessary to determine the status detection parameters according to its different own characteristics, so as to better and intelligently decide the active warning content of the equipment.

[0024] S3, in the remote monitoring center, based on the LSTM-CNN technology, the real-time status detection parameters of the substation equipment are processed and analyzed. First, the various real-time detection signal data embedding layers are converted into vectors, and then input into the LSTM for feature extraction. The output of the LSTM is used as the input of the CNN for further feature extraction. Finally, the real-time classification results of the active warning of the equipment status are obtained, including category, content and level.

[0025] Among them, for the LSTM network part, the depth of the model can be increased by adding LSTM network units to improve the prediction ability of the model. The random deactivation method is used between each LSTM network layer to prevent the model from overfitting, and the vector of the specified format is output in the fully connected layer (Dense).

[0026] The LSTM-CNN model is a deep learning model that combines the advantages of long short-term memory network (LSTM) and convolutional neural network (CNN).

[0027] LSTM is mainly used to process sequential data, such as time series or text, which has a chronological order. It can learn long-term dependencies in sequences and control the transmission and update of information through special gating mechanisms (including input gate, forget gate, and output gate).

[0028] CNN is good at processing data with grid structures, such as images, audio spectra, etc. It uses convolution kernels in the convolution layer to extract local features in the data, and downsamples through the pooling layer to reduce the data dimension while retaining important features.

[0029] Power load data has the characteristics of time series and nonlinearity. In the LSTM-CNN model, the CNN part is usually used to extract the spatial features of the data. For example, when processing text, it can extract the local semantic associations between word vectors; and the LSTM part is used to further process the extracted features in sequence order and mine the time series features. For example, in the text sentiment analysis task, combining the two can better understand the impact of text semantics and its sequential changes on sentiment, thereby improving prediction accuracy.

[0030] S4, compares the real-time classification results of the active warning of the equipment status and the intelligent decision-making library of the substation equipment status management and control through automatic scanning for intelligent decision-making comparison analysis and judgment. If it is determined to be equipment maintenance, the maintenance strategy is automatically pushed; if it is determined to be an inspection, the operation and maintenance strategy is automatically pushed; if equipment testing is required, the test strategy is automatically pushed.

[0031] Taking circuit breakers as an example, their active warning levels are generally divided into three levels. According to the nature of equipment status defect, amplitude, change trend and logical relationship between each other, the severity and urgency of the impact on the safe operation of the equipment, the active warning signal colors are divided into blue warning, yellow warning and red warning.

[0032] (1) Blue warning: indicates that a single or partial status quantity of the equipment is abnormal, which does not affect the operation of the equipment for the time being. The equipment is in a state that only requires attention or tracking analysis, and is judged as equipment testing.

[0033] (2) Yellow warning: Indicates that the trend of the equipment status quantity change is developing in the direction of the standard limit, but has not exceeded the standard limit. The equipment may have hidden dangers and requires enhanced monitoring or further analysis and evaluation. The corresponding judgment is inspection.

[0034] (3) Red warning: indicates that the equipment status exceeds the relevant standard limit, or the change trend is obvious. The equipment may have defects and may develop into a failure. It is urgent to take corresponding measures, and the corresponding judgment is equipment maintenance.

[0035] like Figure 3 As shown, the active warning forms of the circuit breaker are divided into single warning of a single state quantity and comprehensive warning of multiple state quantities. The state warning process is set as follows: if the state information quantity exceeds the required value of the single state quantity warning rule, a red warning is issued; if it does not exceed the required value of the single state quantity warning rule, a multi-state comprehensive warning is issued. When the comprehensive warning exceeds the requirement, warnings of different colors are issued according to the situation. When the comprehensive warning does not exceed the requirement, the circuit breaker is in a normal state.

[0036] S5, in the remote monitoring center, displays the intelligent decision results, including the judgment results and their automatic push strategies, and automatically sends the intelligent decision results to the relevant responsible persons to remind them to pay attention and handle them in time.

[0037] In summary, the present invention automatically extracts equipment operation and maintenance, inspection, and test strategies by establishing an intelligent decision-making library for equipment status control and management, and according to the category, content, and level of active warning of equipment status. When a warning signal appears in the substation equipment, the system performs intelligent decision-making analysis and judgment through the equipment status control and management decision library. If it is determined to be equipment maintenance, the inspection strategy is automatically pushed; if it is determined to be an inspection, the operation and maintenance strategy is pushed; if equipment testing is required, the test strategy is automatically pushed.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A method for intelligent decision-making on substation equipment status identification and abnormality handling, characterized in that: The steps include: S1, based on historical data and expert experience, build an intelligent decision-making library for substation equipment status management and control, including the category, content and level of active warning of equipment status and the corresponding intelligent decision-making results; S2, real-time detection of substation equipment status through the equipment monitoring system, and sending the real-time detection parameters of the equipment status to the remote monitoring center through wireless communication technology; S3, in the remote monitoring center, based on LSTM-CNN technology, the real-time status detection parameters of substation equipment are processed and analyzed. First, the various real-time detection signal data embedding layers are converted into vectors, and then input into LSTM for feature extraction. The output of LSTM is then used as the input of CNN for further feature extraction. Finally, the real-time classification results of active warning of equipment status are obtained, including category, content and level; S4, compare and analyze the obtained real-time classification results of the active warning of equipment status with the intelligent decision-making library of substation equipment status control for intelligent decision-making. If it is determined to be equipment maintenance, the maintenance strategy is automatically pushed; if it is determined to be inspection, the operation and maintenance strategy is automatically pushed; if equipment testing is required, the test strategy is automatically pushed; S5, in the remote monitoring center, the intelligent decision results are displayed, including the judgment results and their automatic push strategies, and the intelligent decision results are automatically sent to the relevant responsible persons.

2. The intelligent decision-making method for substation equipment status identification and abnormality processing according to claim 1 is characterized in that: The step S1 includes the following steps: S11, based on historical data and LSTM-CNN technology, processes and analyzes the historical status parameters of substation equipment. First, various historical signal data embedding layers are converted into vectors, and then input into LSTM for feature extraction. The output of LSTM is then used as the input of CINN for further feature extraction. Finally, the classification results of active warning signals of equipment status are obtained, including categories, contents and levels and their corresponding intelligent decision-making results. S12, based on expert experience, supplement and correct the classification result of the active warning signal of the equipment status obtained in step S11 to make it closer to actual operation experience; S13, based on the new historical data accumulated during the continued operation of the equipment, repeat steps S11 and S12, optimize and continuously improve the intelligent decision-making library for substation equipment status management and control, and gradually weaken the expert experience to make it closer to the equipment's own status operation characteristics.

3. The intelligent decision-making method for substation equipment status identification and abnormality processing according to claim 2 is characterized by: In the step S11, before processing and analyzing the historical status parameters of the substation equipment, the historical equipment status operating parameters are first formed into time series data in chronological order.

4. The intelligent decision-making method for substation equipment status identification and abnormality processing according to claim 1 is characterized by: In the step S3, for the LSTM network part, the depth of the model is increased by adding LSTM network units to improve the prediction ability of the model, a random deactivation method is used between each LSTM network layer to prevent the model from overfitting, and a vector of a specified format is output in the fully connected layer (Dense).

5. The intelligent decision-making method for substation equipment status identification and abnormality processing according to claim 1 is characterized by: In step S4, the real-time classification results of the active warning of the equipment status and the intelligent decision-making comparison analysis and judgment of the intelligent decision-making library of the substation equipment status management and control are completed through automatic scanning.

Images