Key section identification method and device in power system, equipment, storage medium and program product

Abstract

The invention discloses a key section identification method and device in a power system, equipment, a storage medium and a program product. The method comprises the following steps: acquiring real-time monitoring data and historical extreme event information of each piece of equipment in the power system; based on the similarity between each piece of real-time monitoring data and historical extreme event information, at least one piece of fault equipment is determined from all the equipment, each piece of equipment serves as a section, and each piece of fault equipment serves as a target section; inputting the real-time monitoring data of each target section into the trained key section recognition model for recognition to obtain the probability that each target section is a key section; and determining a weight score of each target section, and based on the probability of each target section and the weight score, identifying a key section of the power system from each target section. By adopting the method, the accuracy of key section identification in the power system can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of electric power, and particularly to a method, device, equipment, storage medium and program product for identifying critical sections in a power system. Background Art

[0002] A modern power system is a highly complex network composed of power plants, transmission lines, substations and user terminals. With the development of social economy and the progress of technology, the power demand is continuously increasing, the scale of the power grid is expanding day by day, and the structure is becoming more complex. This not only increases the difficulty of operation and management of the power system, but also puts forward higher requirements for the safety and stability of the system. In a power system, certain equipment or areas (i.e., "sections") are crucial for maintaining the safe and stable operation of the entire system. If these critical sections fail or are overloaded, it may trigger a chain reaction, resulting in large-scale power outages. Therefore, identifying and monitoring these critical sections is one of the core tasks to ensure the reliability of the power system.

[0003] Currently, the main method used for critical section identification is based on machine learning. However, the training of the model in this method depends on a large amount of sample data to learn patterns and rules. When it comes to extreme events (such as large-scale power outages, power grid failures caused by natural disasters, etc.), the model may not be able to comprehensively capture the occurrence rules of extreme events, equipment failures and other rare but high-risk situations, thus resulting in low accuracy in critical section identification.

[0004] Therefore, how to improve the accuracy of critical section identification in a power system has become an urgent problem to be solved. Summary of the Invention

[0005] The embodiments of the present application provide a method, device, equipment, storage medium and program product for identifying critical sections in a power system, which can improve the accuracy of critical section identification in the power system.

[0006] In a first aspect, the embodiments of the present application provide a method for identifying critical sections in a power system, the method comprising:

[0007] Obtain the real-time monitoring data of each device in the power system and historical extreme event information;

[0008] Based on the similarity between each real-time monitoring data and the historical extreme event information, determine at least one faulty device from each device, each device serving as a section, and each faulty device serving as a target section;

[0009] Input the real-time monitoring data of each target section into a trained critical section identification model for identification, and obtain the probability that each target section is a critical section;

[0010] Determine the weight score of each target section, and based on the probability and weight score of each target section, identify the key sections of the power system from each target section.

[0011] In one embodiment, determining the weight score of each target section includes: obtaining the historical key section prediction results corresponding to at least one historical target section in the power system during a historical time period, and based on each historical key section prediction result, determining the weight score of each historical target section at the previous moment; the historical key section prediction result includes that the historical target section is a historical key section, or the historical target section is a historical non-key section; determining the average value of the real-time operation data included in the real-time monitoring data corresponding to each of the multiple target sections; based on the weight score of each historical target section at the previous moment, the real-time operation data corresponding to each target section, the average value, and a preset time period, using a preset weight score determination function to determine the weight score of each target section.

[0012] In one embodiment, based on each historical key section prediction result, determining the weight score of each historical target section at the previous moment includes: obtaining the occurrence times of each historical target section; based on the historical key section prediction result corresponding to each historical target section, determining the number of times that each historical target section is a historical key section; based on the occurrence times of each historical target section and the number of times that the historical target section is a historical key section, determining the weight score of each historical target section at the previous moment.

[0013] In one embodiment, before determining at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information, the method further includes: performing feature transformation on the historical extreme event information to obtain a first feature vector corresponding to the historical extreme event information; performing feature transformation on the real-time monitoring data of each device to obtain a second feature vector corresponding to each device; determining the similarity between the first feature vector and each second feature vector as the similarity between each real-time monitoring data and the historical extreme event information.

[0014] In one embodiment, based on the probability and weight score of each target section, identifying the key sections of the power system from each target section includes: based on the probability and weight score of each target section, determining the final probability that each target section is a key section; selecting at least one final probability greater than a preset probability threshold from the multiple final probabilities, and determining the target sections corresponding to the at least one final probability as the key sections of the power system.

[0015] In one embodiment, the method further includes: obtaining historical non-extreme event information and historical extreme event information of the power system, and simulating different types of extreme events by using a pre-constructed power system simulation model to generate simulated extreme event information; constructing a training data set based on the historical non-extreme event information, the historical extreme event information, and the simulated extreme event information; wherein, the label corresponding to the faulty device in the simulated extreme event information and the historical extreme event information is the historical critical section, and the label corresponding to the faulty device in the historical non-extreme event information is the historical non-critical section; performing feature transformation on each data in the training data set to obtain the transformed training data set; using the transformed training data set as the input variable to train the initialized critical section identification model, and obtaining the trained critical section identification model when it is determined that the training stop condition is satisfied.

[0016] In a second aspect, the present application provides a critical section identification device in a power system, which is applied to each working node in a distributed cluster; the device includes:

[0017] An acquisition module, configured to acquire real-time monitoring data of each device in the power system and historical extreme event information;

[0018] A determination module, configured to determine at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information, and each faulty device is used as a target section;

[0019] An identification module, configured to input the real-time monitoring data of each target section into the trained critical section identification model for identification to obtain the probability that each target section is a critical section;

[0020] The determination module is further configured to determine the weight score of each target section, and identify the critical section of the power system from each target section based on the probability and the weight score of each target section.

[0021] In a third aspect, the present application provides a computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0022] Acquire real-time monitoring data of each device in the power system and historical extreme event information;

[0023] Determine at least one faulty device from each device based on the similarity between the real-time monitoring data and the historical extreme event information, and each faulty device is used as a target section;

[0024] Input the real-time monitoring data of each target section into the trained critical section identification model for identification to obtain the probability that each target section is a critical section;

[0025] Determine the weight scores of each target section, and based on the probabilities and weight scores of each target section, identify the key sections of the power system from each target section.

[0026] Fourthly, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0027] Obtain the real-time monitoring data and historical extreme event information of each device in the power system;

[0028] Based on the similarity between the real-time monitoring data and the historical extreme event information, determine at least one faulty device from each device, and each faulty device serves as a target section;

[0029] Input the real-time monitoring data of each target section into the trained key section identification model for identification to obtain the probability that each target section is a key section;

[0030] Determine the weight scores of each target section, and based on the probabilities and weight scores of each target section, identify the key sections of the power system from each target section.

[0031] Fifthly, the present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, the following steps are implemented:

[0032] Obtain the real-time monitoring data and historical extreme event information of each device in the power system;

[0033] Based on the similarity between the real-time monitoring data and the historical extreme event information, determine at least one faulty device from each device, and each faulty device serves as a target section;

[0034] Input the real-time monitoring data of each target section into the trained key section identification model for identification to obtain the probability that each target section is a key section;

[0035] Determine the weight scores of each target section, and based on the probabilities and weight scores of each target section, identify the key sections of the power system from each target section.

[0036] The above-mentioned method, device, equipment, storage medium and program product for identifying key sections in a power system. A computer device can obtain real-time monitoring data of each device in the power system and historical extreme event information. Based on the similarity between the real-time monitoring data and the historical extreme event information, at least one faulty device is determined from each device, and each faulty device serves as a target section. The real-time monitoring data of each target section is input into a trained key section identification model for identification to obtain the probability that each target section is a key section. Determine the weight score of each target section, and based on the probability and weight score of each target section, identify the key sections of the power system from each target section. By using this method, first, based on the similarity between the real-time monitoring data of each device in the power system and the historical extreme event information, the occurrence rules of rare but high-risk situations such as extreme events and equipment failures can be captured more comprehensively, so as to identify multiple target sections (or potential key sections) in advance. Then, for each target section, through a pre-trained key section identification model, the probability that each target section is a key section can be determined more accurately. After that, by calculating the weight score of each target section and based on the weight score of each target section and the probability that the target section is a key section, the key sections can be identified more accurately from multiple target sections, that is, the accuracy of identifying key sections in the power system can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0038] Figure 1 FIG. is a schematic diagram of an application scenario of a method for identifying key sections in a power system provided by an embodiment of the present application;

[0039] Figure 2 FIG. is a schematic flowchart of a method for identifying key sections in a power system provided by an embodiment of the present application;

[0040] Figure 3 FIG. is a schematic flowchart of another method for identifying key sections in a power system provided by an embodiment of the present application;

[0041] Figure 4 FIG. is a schematic structural diagram of a device for identifying key sections in a power system provided by an embodiment of the present application;

[0042] Figure 5 FIG. is a schematic structural diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0043] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0044] The application scenario of the critical section identification method in the power system provided by the embodiments of the present application will be introduced below.

[0045] Please refer to Figure 1 , Figure 1 which is a schematic diagram of an application scenario of a critical section identification method in a power system provided by an embodiment of the present application. As Figure 1 shown, it includes a computer device 101 and a database server 102. Among them, data transmission is carried out between the computer device 101 and the database server 102 through a network.

[0046] The computer device 101 can first obtain the real-time monitoring data and historical extreme event information of each device in the power system from the database server 102; then, based on the similarity between the real-time monitoring data and the historical extreme event information, at least one faulty device is determined from each device, and each faulty device is used as a target section; the real-time monitoring data of each target section is input into the trained critical section identification model for identification to obtain the probability that each target section is a critical section; the weight score of each target section is determined, and based on the probability and weight score of each target section, the critical section of the power system is identified from each target section. By using this method, first, based on the similarity between the real-time monitoring data of each device in the power system and the historical extreme event information, the occurrence rules of rare but high-risk situations such as extreme events and equipment failures can be more comprehensively captured, so as to identify multiple target sections (or potential critical sections) in advance. Then, for each target section, through the pre-trained critical section identification model, the probability that each target section is a critical section can be more accurately determined; after that, by calculating the weight score of each target section and based on the weight score of each target section and the probability that the target section is a critical section, the critical section can be more accurately identified from multiple target sections, that is, the accuracy of critical section identification in the power system can be improved.

[0047] Optionally, the computer device 101 can be a terminal device or a server. Among them, the terminal device mentioned here can include but is not limited to: smart phones, tablet computers, laptop computers, desktop computers, smart watches, smart TVs, smart vehicle terminals, etc. The server mentioned here can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, etc.

[0048] The database server 102 can be an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, etc., which is not limited here.

[0049] Please refer to Figure 2 , Figure 2 which is a schematic flowchart of a method for identifying critical sections in a power system provided by an embodiment of the present application. This method can be executed by a computer device (such as the computer device 101 in Figure 1 ). As Figure 2 shown, the method for identifying critical sections in this power system may include but is not limited to the following steps:

[0050] S201. Obtain the real-time monitoring data and historical extreme event information of each device in the power system.

[0051] Optionally, the historical extreme event information may include but is not limited to basic extreme event information, grid operation data corresponding to extreme events, environmental data corresponding to extreme events, event consequence data corresponding to extreme events, etc., which is not limited here.

[0052] Among them, the basic extreme event information may include but is not limited to event identification, event type, event time, event location, event influence range, etc. Among them, the event identification refers to the unique identifier corresponding to the extreme event, which is used to distinguish different extreme events; the event type is used to describe the nature of the extreme event, and the event type is, for example, typhoon, rainstorm, thunderstorm, ice disaster, earthquake, etc.; the event time refers to the specific time when the extreme event occurs, which includes the start time and end time of the extreme event, and can be accurate to minutes or seconds; the event location refers to the geographical location where the extreme event occurs, and the event location is, for example, a certain substation, transmission line or a certain area, etc.; the event influence range refers to the grid area or equipment range affected by the extreme event, and the event influence range is, for example, a certain section, a certain substation or the entire grid, etc.

[0053] Among them, the power grid operation data corresponding to extreme events may include, but are not limited to, voltage, current, power, frequency, load distribution, equipment status, and fault information, etc. during the occurrence of extreme events. Among them, voltage refers to the voltage values of each target section and substation during the occurrence of extreme events, which may include, but are not limited to, phase voltage, line voltage, etc.; current refers to the current values of each target section and substation during the occurrence of extreme events, which may include, but are not limited to, phase current, line current, etc.; power refers to the power of each section and substation during the occurrence of extreme events, which may include active power and reactive power; frequency refers to the frequency fluctuation of the power grid during the occurrence of extreme events; load distribution refers to the load conditions of each region in the power grid during the occurrence of extreme events, which may include, but are not limited to, load factor, maximum load, minimum load, etc.; equipment status refers to the operating status of each equipment in the power system during the occurrence of extreme events, which may include, but are not limited to, transformer temperature, switch status, protection device operation conditions, etc.; fault information refers to the specific fault type, fault location, and fault time, etc. corresponding to the situation where certain equipment in the power system fails or trips due to extreme events.

[0054] Among them, the environmental data corresponding to extreme events may include, but are not limited to, temperature, humidity, wind speed, precipitation, air pressure, wind direction, weather warning information, geographic information system data, etc. Among them, temperature refers to the air temperature during the occurrence of extreme events, such as extreme high temperature or extreme low temperature; humidity refers to the air humidity during the occurrence of extreme events, especially high humidity may cause insulation problems; wind speed refers to the wind speed during the occurrence of extreme events, for example, strong wind may cause transmission line galloping or tower collapse; precipitation includes rainfall or snowfall during the occurrence of extreme events, for example, heavy rain may cause floods, and heavy snow may cause transmission lines to be crushed by snow accumulation; air pressure refers to the change in atmospheric pressure during the occurrence of extreme events, for example, a low-pressure system may cause storm surges; wind direction refers to the wind direction during the occurrence of extreme events, for example, strong wind may cause transmission line galloping; weather warning information refers to the meteorological warning information before the occurrence of extreme events, such as typhoon warning, heavy rain warning, thunderstorm warning, etc.; geographic information system data refers to information such as the terrain, landform, and sea wave height of the location where extreme events occur.

[0055] Among them, the event consequence data corresponding to extreme events may include, but are not limited to, power outage scope, restoration time, economic loss, casualties, emergency response measures, etc. Among them, the power outage scope includes the power outage area and power outage time caused by extreme events; the restoration time refers to the time from the occurrence of extreme events to the power grid returning to normal operation, which reflects the severity of extreme events and the difficulty of restoration; economic loss refers to the direct economic loss caused by extreme events, including equipment damage, power supply interruption, etc.; casualties refer to the number of casualties and the injury situation caused by extreme events; emergency response measures refer to the emergency measures taken after the occurrence of extreme events, such as the dispatch of repair teams, the activation of standby power sources, etc.

[0056] Optionally, the real-time monitoring data of each device may include, but is not limited to, the real-time operation data and real-time environment data of each device.

[0057] Among them, the real-time operation data of each device may include, but is not limited to, the real-time voltage, real-time current, real-time power, real-time frequency, real-time load distribution, real-time device status, etc. corresponding to each device; the real-time environment data of each device may include, but is not limited to, the real-time temperature, real-time humidity, real-time wind speed, real-time precipitation, real-time air pressure, real-time wind direction, real-time weather warning information, real-time geographic information system data, etc. corresponding to each device.

[0058] S202. Determine at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information. Each device serves as a section, and each faulty device serves as a target section.

[0059] In an alternative embodiment, the similarity between the real-time monitoring data and the historical extreme event information may be the cosine distance between the real-time monitoring data and the historical extreme event information, or the Euclidean distance between the real-time monitoring data and the historical extreme event information, etc., which is not limited here.

[0060] Optionally, the target section may also be referred to as a potential critical section.

[0061] S203. Input the real-time monitoring data of each target section into the trained critical section identification model for identification to obtain the probability that each target section is a critical section.

[0062] In an alternative embodiment, the critical section identification model may be obtained by a computer device training a pre-constructed initialized critical section identification model using the historical monitoring data and historical extreme event information of each device in the power system.

[0063] In this embodiment, the historical monitoring data may include historical operation data and historical environment data. The historical extreme event information may be obtained by the computer device through analyzing the historical operation data and historical environment data. That is, the computer device may use the historical operation data and historical environment data of each device in the power system to train the initialized critical section identification model to obtain the trained critical section identification model.

[0064] Optionally, the computer device uses the historical operation data and historical environment data of each device in the power system to train the initialized critical section identification model, and obtains the trained critical section identification model, which may include: obtaining the historical operation data and historical environment data of each device in the power system; analyzing the historical operation data and historical environment data of each device to obtain historical extreme event information; performing data cleaning processing on the historical extreme event information, the historical operation data of each device, and the historical environment data of each device respectively to obtain the processed historical extreme event information, the processed historical operation data of each device, and the processed historical environment data of each device; constructing a training data set based on the processed historical extreme event information, the processed historical operation data of each device, and the processed historical environment data of each device; training the initialized critical section identification model with the training data set to obtain the predicted critical section result corresponding to each device; training the initialized critical section identification model in the direction of reducing the difference between the predicted critical section result and the actual critical section result corresponding to each device until the stop training condition is met, and obtaining the trained critical section identification model.

[0065] S204. Determine the weight score of each target section, and based on the probability and weight score of each target section, identify the critical section of the power system from each target section.

[0066] In the embodiment of the present application, the computer device can obtain the real-time monitoring data and historical extreme event information of each device in the power system; determine at least one faulty device from each device based on the similarity between the real-time monitoring data and the historical extreme event information, and each faulty device is used as a target section; input the real-time monitoring data of each target section into the trained critical section identification model for identification to obtain the probability that each target section is a critical section; determine the weight score of each target section, and based on the probability and weight score of each target section, identify the critical section of the power system from each target section. By using this method, first, based on the similarity between the real-time monitoring data of each device in the power system and the historical extreme event information, the occurrence rules of rare but high-risk situations such as extreme events and equipment failures can be captured more comprehensively, so as to identify multiple target sections (or potential critical sections) in advance. Then, for each target section, through the pre-trained critical section identification model, the probability that each target section is a critical section can be determined more accurately; afterwards, by calculating the weight score of each target section and based on the weight score of each target section and the probability that the target section is a critical section, the critical section can be identified more accurately from multiple target sections, that is, the accuracy of critical section identification in the power system can be improved.

[0067] In an alternative embodiment,Figure 2 In the key section identification method in the power system shown, before the computer device determines at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information, the historical extreme event information may be subjected to feature transformation to obtain a first feature vector corresponding to the historical extreme event information; the real-time monitoring data of each device is subjected to feature transformation to obtain a second feature vector corresponding to each device; the similarity between the first feature vector and each second feature vector is determined as the similarity between each real-time monitoring data and each historical extreme event information.

[0068] In this embodiment, the computer device performs feature transformation on the real-time monitoring data of each device to obtain a second feature vector corresponding to each device, which may include: determining the real-time monitoring value of each data in the real-time monitoring data of each device; determining the mean value of each data in the historical monitoring data of each device; determining the transformation rate of each data based on the mean value and the real-time monitoring value of each data; and determining the second feature vector corresponding to each device based on the change rate of each data.

[0069] In this embodiment, when the computer device determines the similarity between the first feature vector and each second feature vector, the following formula (1) may be used.

[0070] (1)

[0071] In formula (4), S nm represents the similarity between the real-time monitoring data corresponding to the nth section (device) and the historical extreme event information m; v m represents the first feature vector corresponding to the historical extreme event information m; v 0n represents the second feature vector corresponding to the nth section.

[0072] In some embodiments, the computer device determines at least one faulty device from the respective devices based on the similarity between each real-time monitoring data and the historical extreme event information, which may include: performing normalization processing on the similarity between each real-time monitoring data and the historical extreme event information respectively to obtain a plurality of normalized similarities; and determining the device corresponding to the normalized similarity greater than the preset similarity threshold among the plurality of normalized similarities as the faulty device.

[0073] Exemplarily, assume that the number of devices is 3, denoted as Device 1, Device 2, and Device 3. The normalized similarity between Device 1 and the historical extreme event information is 0.91, the normalized similarity between Device 2 and the historical extreme event information is 0.92, and the normalized similarity between Device 3 and the historical extreme event information is 0.80. And assume that the preset similarity threshold is 0.90. In this case, the computer device can determine that 0.91 and 0.92 are greater than the preset similarity threshold of 0.90. In this situation, the computer device can determine Device 1 and Device 2 as faulty devices.

[0074] Adopting this implementation manner, the computer device determines the similarity between the real-time monitoring data of each device and the historical extreme event, which is beneficial to selecting the target section (or potential key section) from each device (or section) based on multiple similarities. Thus, the computational complexity of subsequent identification of the key section can be reduced, and the efficiency of key section identification can be improved.

[0075] In an alternative implementation manner, Figure 2 In the method for identifying the key section in the power system shown, the computer device determines the weight score of each target section, which may include: obtaining the historical key section prediction results corresponding to at least one historical target section in the power system during the historical time period, and determining the weight score of each historical target section at the previous moment based on each historical key section prediction result; the historical key section prediction result includes that the historical target section is a historical key section, or the historical target section is a historical non-key section; determining the average value of the real-time operation data included in the real-time monitoring data corresponding to each of the multiple target sections; and determining the weight score of each target section by using a preset weight score determination function based on the weight score of each historical target section at the previous moment, the real-time operation data corresponding to each target section, the average value, and the preset time period.

[0076] Among them, the historical target section refers to the section where a device fails in the power system during the historical time period, which includes the historical key section, that is, the key section caused by an extreme event.

[0077] In this implementation manner, the computer device determines the weight score of each historical target section at the previous moment based on each historical key section prediction result, which may include: obtaining the occurrence times of each historical target section; determining the number of times each historical target section is a historical key section based on the historical key section prediction result corresponding to each historical target section; and determining the weight score of each historical target section at the previous moment based on the occurrence times of each historical target section and the number of times the historical target section is a historical key section.

[0078] Optionally, when determining the weight score of the moment before each historical target section based on the occurrence times of each historical target section and the number of times that the historical target section is a historical key section, the computer device may adopt the following formula (2).

[0079] (2)

[0080] In formula (2), represents the weight score of the moment before the historical target section (i.e., the moment at t-1).

[0081] Optionally, when the prediction results of the historical key sections of the historical target section are all historical non-key sections, the weight score of the moment before the historical target section can be determined by the following formula (3).

[0082] (3)

[0083] In this embodiment, the preset weight score determination function can be shown as the following formula (4).

[0084] (4)

[0085] In formula (4), represents the weight score of the i-th target section; represents the smoothing coefficient, and the value range of can be [0.4, 0.7]; represents the weight score of the moment before the historical target section (i.e., the moment at t-1); x i represents the real-time operation data corresponding to the i-th target section; I represents the total number of target sections; represents the average value of the real-time operation data corresponding to I target sections; T represents the preset time period; S im represents the similarity between the real-time monitoring data corresponding to the i-th target section and the historical extreme event information m. Optionally, the determination method of S im can refer to the aforementioned formula (1), which will not be elaborated here.

[0086] In some embodiments, when the target section occurs for the first time, since the risk of the section with a first-time failure is smaller than that of the section with multiple failures, the computer device may assign 0.1 to the weight score of the target section. That is to say, when the target section occurs for the first time, the computer device may determine that the weight score of the target section is 0.1.

[0087] With this embodiment, the computer device can determine the weight score of each target section. In this way, it is beneficial to adjust the probability that each target section is a critical section based on the weight score of each target section. Thus, it is beneficial to improve the accuracy of critical section identification.

[0088] In an alternative embodiment, Figure 2 In the method for identifying critical sections in the power system shown, the computer device identifies the critical sections of the power system from each target section based on the probability and weight score of each target section, which may include: determining the final probability that each target section is a critical section based on the probability and weight score of each target section; selecting at least one final probability greater than the preset probability threshold from multiple final probabilities, and determining the target sections corresponding to the at least one final probability as the critical sections of the power system.

[0089] Optionally, when the computer device determines the final probability that each target section is a critical section based on the probability and weight score of each target section, the following formula (5) may be used.

[0090] (5)

[0091] In formula (5), represents the final probability that the i-th target section is a critical section; represents the probability that the i-th target section is a critical section; represents the weight score of the i-th target section, which can be determined by the above formula (4).

[0092] Exemplarily, assume there are 3 target sections, denoted as target section 1, target section 2, and target section 3. Among them, the final probability that target section 1 is a critical section is 0.85, the final probability that target section 2 is a critical section is 0.87, the final probability that target section 3 is a critical section is 0.89, and assume the preset probability threshold is 0.86. Then the computer device can select the final probabilities greater than the preset probability threshold 0.86 from the three final probabilities 0.85, 0.87, and 0.89, that is, 0.87 and 0.89, and determine the target section 2 corresponding to 0.87 and the target section 3 corresponding to 0.89 as the critical sections of the power system.

[0093] With this embodiment, the computer device can adjust the probability that each target section is a critical section based on the weight score of each target section to obtain the final probability that each target section is a critical section. Thus, based on multiple final probabilities, the critical sections in the power system can be identified more accurately, that is, the accuracy of critical section identification in the power system is improved.

[0094] Next, in combination withFigure 3 This section elaborates on the overall process of the critical section identification method in the power system provided by the embodiments of this application. Please refer to Figure 3 , Figure 3 FIG. Figure 3 shows a schematic flowchart of another critical section identification method in the power system provided by the embodiments of this application. As

[0095] S301. Obtain the real-time monitoring data and historical extreme event information of each device in the power system.

[0096] In an alternative embodiment, the related description of step 301 can refer to the description in the aforementioned step S201, and will not be elaborated here.

[0097] S302. Perform feature transformation on the historical extreme event information to obtain the first feature vector corresponding to the historical extreme event information, and perform feature transformation on the real-time monitoring data of each device to obtain the second feature vector corresponding to each device.

[0098] In an alternative embodiment, the feature transformation may include standardization feature transformation and non-linear feature transformation.

[0099] Optionally, for the historical extreme event information and the real-time monitoring data of each device, the computer device may select different feature transformation methods based on the data distribution type of each data to perform feature transformation processing on each data.

[0100] Exemplarily, the computer device may use logarithmic transformation to perform feature transformation processing on data with a long-tailed distribution, such as power data, current data, etc., to compress large values and make their distribution more uniform. Among them, the expression of the logarithmic transformation can be shown as the following formula (6).

[0101] (6)

[0102] In formula (6), represents the transformed data obtained after performing feature transformation on the data X with a long-tailed distribution type; represents the scaling factor of the log function; represents the slope of the exp function; represents the offset.

[0103] Exemplarily, the computer device may use sine transformation to perform feature transformation processing on data with a periodic distribution, such as temperature data, humidity data, etc., so that the transformed data is more conducive to model learning. Among them, the expression of the sine transformation can be shown as the following formula (7).

[0104] (7)

[0105] In formula (7), represents the transformed data obtained after the feature transformation of the data Y with a periodic distribution type of data distribution; represents the amplitude of the sin function; represents the control frequency of the sin function; represents the phase shift.

[0106] S303. Determine the similarity between the first eigenvector and each second eigenvector as the similarity between each real-time monitoring data and the historical extreme event information.

[0107] In an alternative embodiment, the computer device may use the aforementioned formula (1) to determine the similarity between the first eigenvector and each second eigenvector.

[0108] S304. Based on the similarity between each real-time monitoring data and the historical extreme event information, determine at least one faulty device from each device. Each device serves as a section, and each faulty device serves as a target section.

[0109] In an alternative embodiment, the computer device determines at least one faulty device from each of the devices based on the similarity between each real-time monitoring data and the historical extreme event information, which may include: performing normalization processing on the similarity between each real-time monitoring data and the historical extreme event information respectively to obtain multiple normalized similarities; determining the device corresponding to the normalized similarity greater than the preset similarity threshold among the multiple normalized similarities as the faulty device.

[0110] S305. Input the real-time monitoring data of each target section into the trained critical section recognition model for recognition to obtain the probability that each target section is a critical section.

[0111] In an alternative embodiment, the relevant description of step S305 can refer to the description in the aforementioned step 203, and will not be elaborated here.

[0112] S306. Determine the weight score of each target section.

[0113] In an alternative embodiment, the computer device may use the aforementioned formulas (2) to (4) to determine the weight score of each target section.

[0114] S307. Based on the probability and weight score of each target section, determine the final probability that each target section is a critical section.

[0115] In an alternative embodiment, when the computer device determines the final probability of each target section being a critical section based on the probability of each target section and the weight score, the aforementioned formula (5) can be adopted.

[0116] S308. Select at least one final probability greater than a preset probability threshold from multiple final probabilities, and determine the target sections corresponding to the at least one final probability as the critical sections of the power system.

[0117] In the embodiment of the present application, based on the similarity between the real-time monitoring data of each device in the power system and the historical extreme event information, the computer device can more comprehensively capture the occurrence rules of rare but high-risk situations such as extreme events and equipment failures, so as to identify multiple target sections (or potential critical sections) in advance. Then, for each target section, through the pre-trained critical section identification model, the probability of each target section being a critical section can be determined more accurately; after that, by calculating the weight score of each target section and based on the weight score of each target section and the probability of this target section being a critical section, the critical section can be more accurately identified from multiple target sections, that is, the accuracy of critical section identification in the power system can be improved.

[0118] In an alternative embodiment, Figure 2 and Figure 3 In the critical section identification method in the power system shown, the trained critical section identification model can be determined by the computer device in the following manner: Obtain the historical non-extreme event information and historical extreme event information of the power system, and use the pre-constructed power system simulation model to simulate different types of extreme events to generate simulation extreme event information; Based on the historical non-extreme event information, historical extreme event information, and simulation extreme event information, construct a training data set; Among them, the labels of the faulty devices in the simulation extreme event information and historical extreme event information are historical critical sections, and the labels of the faulty devices in the historical non-extreme event information are historical non-critical sections; Perform feature transformation on each data in the training data set to obtain the transformed training data set; Use the transformed training data set as the input variable to train the initialized critical section identification model, and obtain the trained critical section identification model when it is determined that the stop training condition is met.

[0119] Among them, the historical non-extreme event information refers to the historical operation data of a preset proportion in the remaining historical operation data except the historical extreme event information in the historical operation data, and the historical non-extreme event information includes non-extreme event operation data and non-critical section data. Optionally, the preset proportion can be 2 / 3.

[0120] In this embodiment, the historical extreme event information can be obtained by a computer device through analyzing the historical monitoring data of each device in the power system, where the historical monitoring data includes historical operation data and historical environmental data.

[0121] In this embodiment, optionally, the label corresponding to the historical critical section can be 1, and the label corresponding to the historical non-critical section can be -1, which is not limited herein.

[0122] In this embodiment, before a computer device uses a pre-constructed power system simulation model to simulate different types of extreme events and generate simulation extreme event information, it can also construct a power system simulation model, which includes a power plant component, a transmission line component, a substation component, a load component, and a distributed energy component; determine the type of the extreme event to be simulated and the parameters corresponding to each extreme event. Optionally, the types of extreme events to be simulated can include, but are not limited to, natural disasters, human factors, other extreme events, etc. Among them, natural disasters are, for example, typhoons, heavy rains, thunderstorms, ice disasters, earthquakes, etc.; human factors are, for example, equipment failures, cyberattacks, etc.; other extreme events are, for example, large-scale power outages, frequency fluctuations, etc.

[0123] In this embodiment, a computer device uses a pre-constructed power system simulation model to simulate different types of extreme events and generate simulation extreme event information, which may include: using the pre-constructed power system simulation model to simulate different types of extreme events to obtain the simulation results corresponding to each type of extreme event, and the simulation results at least include the power grid operation state and the impacts caused by the extreme events; generating simulation extreme event information based on the simulation results corresponding to each type of extreme event. Optionally, the simulation extreme event information may include, but is not limited to, simulation extreme event identifiers, simulation extreme event types, simulation extreme event times, simulation extreme event locations, simulation extreme event impact ranges, power grid operation states, simulation extreme event impacts, etc.

[0124] In this embodiment, a computer device constructs a training dataset based on historical non-extreme event information, historical extreme event information, and simulation extreme event information, which may include: performing data cleaning processing on the historical non-extreme event information, historical extreme event information, and simulation extreme event information respectively to obtain the processed historical non-extreme event information, processed historical extreme event information, and processed simulation extreme event information; using the set composed of the processed historical non-extreme event information, processed historical extreme event information, and processed simulation extreme event information as the training dataset.

[0125] In this embodiment, the computer device performs feature transformation on each data in the training dataset to obtain the transformed training dataset, which may include: performing feature transformation on each data based on the data distribution type of each data in the training dataset to obtain each transformed data; and obtaining the transformed training dataset based on each transformed data.

[0126] Optionally, the feature transformation includes a normalization feature transformation and a non-linear feature transformation. The computer device can scale all the data to the same scale. For example, through Z-score normalization or Min-Max normalization processing, to ensure the comparability between different data.

[0127] Optionally, for the data with a long-tailed distribution type in the training dataset, such as power data, current data, etc., the computer device can use logarithmic transformation to perform feature transformation on the long-tailed distribution data to compress large values and make its distribution more uniform. Among them, the expression of the logarithmic transformation is as shown in the above formula (6). For the data with a periodic distribution type, such as temperature data, humidity data, etc., the computer device can adopt the sine transformation method to perform feature transformation on the periodic distribution data, so that the transformed data is more conducive to model learning, so that the model can better understand the variation law of the data. Among them, the expression of the sine transformation is as shown in the above formula (7).

[0128] Adopting this embodiment, during the training of the critical section identification model, the computer device introduces the generation of simulation extreme event information based on the power system simulation model, which can increase the sample quantity of the extreme event information, so that the model can fully learn the characteristics of the extreme event, so as to obtain a critical section identification model with better critical section identification accuracy. Therefore, it is beneficial to more accurately determine the probability that each target section is a critical section.

[0129] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indication of the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same moment, but can be executed at different moments. The execution order of these steps or stages does not necessarily need to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0130] Based on the same inventive concept, an embodiment of the present application further provides a critical section identification device in a power system for implementing the critical section identification method in the above-mentioned power system. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the critical section identification device in the power system provided below can refer to the limitations on the critical section identification method in the power system above, and will not be repeated here.

[0131] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of a critical section identification device in a power system provided by an embodiment of the present application. As Figure 4 shown, the critical section identification device in the power system may include, but is not limited to:

[0132] An acquisition module 401, configured to acquire real-time monitoring data of each device in the power system and historical extreme event information;

[0133] A determination module 402, configured to determine at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information, and each faulty device serves as a target section;

[0134] An identification module 403, configured to input the real-time monitoring data of each target section into a trained critical section identification model for identification, and obtain the probability that each target section is a critical section;

[0135] The determination module 402 is further configured to determine the weight score of each target section, and based on the probability and weight score of each target section, identify the critical section of the power system from each target section.

[0136] In one embodiment, when the determination module 402 is used to determine the weight score of each target section, it is specifically configured to: acquire the historical critical section prediction results corresponding to at least one historical target section in the power system during a historical time period, and based on each historical critical section prediction result, determine the weight score of the previous moment of each historical target section; the historical critical section prediction result includes that the historical target section is a historical critical section, or the historical target section is a historical non-critical section; determine the average value of the real-time operation data included in the real-time monitoring data corresponding to multiple target sections respectively; based on the weight score of the previous moment of each historical target section, the real-time operation data corresponding to each target section, the average value, and a preset time period, use a preset weight score determination function to determine the weight score of each target section.

[0137] In one embodiment, when determining module 402 is used to determine the weight score at the previous moment of each historical target section based on the prediction results of each historical key section, it is specifically used for: obtaining the occurrence times of each historical target section; determining the number of times each historical target section is a historical key section based on the prediction results of the historical key sections corresponding to each historical target section; and determining the weight score at the previous moment of each historical target section based on the occurrence times of each historical target section and the number of times the historical target section is a historical key section.

[0138] In one embodiment, the device may further include a processing module. Before the determining module 402 is used to determine at least one faulty device from each device based on the similarity between each real-time monitoring data and the historical extreme event information, the processing module is used to perform feature transformation on the historical extreme event information to obtain a first feature vector corresponding to the historical extreme event information; perform feature transformation on the real-time monitoring data of each device to obtain a second feature vector corresponding to each device, and the determining module 402 is further used to determine the similarity between the first feature vector and each second feature vector as the similarity between each real-time monitoring data and the historical extreme event information.

[0139] In one embodiment, when determining module 403 is used to identify the key sections of the power system from each target section based on the probability and weight score of each target section, it is specifically used for: determining the final probability of each target section being a key section based on the probability and weight score of each target section; selecting at least one final probability greater than the preset probability threshold from multiple final probabilities, and determining the target sections corresponding to the at least one final probability as the key sections of the power system.

[0140] In one embodiment, the device may further include a training module, and the training module is used for: obtaining the historical non-extreme event information and historical extreme event information of the power system, and using a pre-constructed power system simulation model to simulate different types of extreme events to generate simulation extreme event information; constructing a training data set based on the historical non-extreme event information, historical extreme event information, and simulation extreme event information; where the labels of the faulty devices in the simulation extreme event information and historical extreme event information are historical key sections, and the labels of the faulty devices in the historical non-extreme event information are historical non-key sections; performing feature transformation on each data in the training data set to obtain a transformed training data set; using the transformed training data set as an input variable to train an initialized key section identification model, and obtaining a trained key section identification model when it is determined that the stop training condition is met.

[0141] Each module in the above-mentioned key section identification device in the power system can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above modules can be embedded in the processor in the terminal device in the form of hardware or be independent of it, or be stored in the memory in the terminal device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0142] In an exemplary embodiment, the embodiment of the present application provides a computer device, which can be a terminal device, and its internal structure diagram can be as Figure 5 shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communicating with external terminals in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for identifying key sections in a power system. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0143] Those skilled in the art can understand that Figure 5 the structure shown in

[0144] is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0145] In an exemplary embodiment, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned key section identification method in the power system are implemented.

[0146] In an exemplary embodiment, the present application provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps in the above-mentioned key section identification method in the power system are implemented.

[0147] It should be noted that the data involved in the present application (including but not limited to real-time monitoring data of each device, historical extreme event information, similarity, probability that each target section is a key section, etc.) are all information and data authorized by users or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant regulations.

[0148] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0149] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0150] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for identifying a critical section in a power system, characterized in that: The method comprises: Obtain real-time monitoring data and historical extreme event information of each device in the power system; Based on the similarity between each of the real-time monitoring data and the historical extreme event information, at least one faulty device is determined from the various devices, wherein each of the devices is used as a section and each of the faulty devices is used as a target section; The real-time monitoring data of each target section is input into the trained key section identification model for identification, and the probability of each target section being a key section is obtained; A weight score of each target section is determined, and based on the probability of each target section and the weight score, a critical section of the power system is identified from each target section.

2. The method according to claim 1, characterized in that: Determining the weight score of each target section includes: Obtain historical critical section prediction results corresponding to at least one historical target section in the power system within a historical time period, and determine the weight score of each historical target section at the previous moment based on each historical critical section prediction result; the historical critical section prediction result includes that the historical target section is a historical critical section, or that the historical target section is a historical non-critical section; Determine an average value of the real-time operation data included in the real-time monitoring data corresponding to the plurality of target sections respectively; Based on the weight score of each of the historical target sections at the previous moment, the real-time operating data corresponding to each of the target sections, the average value and the preset time period, a preset weight score determination function is used to determine the weight score of each of the target sections.

3. The method according to claim 2, characterized in that Determining the weight score of each of the historical target sections at the previous moment based on the prediction results of each of the historical key sections includes: Obtaining the number of occurrences of each of the historical target sections; Based on the prediction results of the historical key sections corresponding to each of the historical target sections, determining the number of times each of the historical target sections is the historical key section; Based on the number of occurrences of each of the historical target sections and the number of times the historical target section is the historical key section, a weight score of each of the historical target sections at a previous moment is determined.

4. The method according to claim 1, characterized in that: Before determining at least one faulty device from each of the devices based on the similarity between each of the real-time monitoring data and the historical extreme event information, the method further includes: Performing feature transformation on the historical extreme event information to obtain a first feature vector corresponding to the historical extreme event information; Performing feature transformation on the real-time monitoring data of each device to obtain a second feature vector corresponding to each device; The similarity between the first feature vector and each of the second feature vectors is determined as the similarity between each of the real-time monitoring data and the historical extreme event information.

5. The method according to claim 1, characterized in that The step of identifying a key section of the power system from each of the target sections based on the probability and the weight score of each of the target sections comprises: Based on the probability of each target section and the weight score, determining a final probability that each target section is the key section; At least one final probability greater than a preset probability threshold is selected from the multiple final probabilities, and the target section corresponding to the at least one final probability is determined as the critical section of the power system.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Acquiring historical non-extreme event information and historical extreme event information of the power system, and simulating different types of extreme events using a pre-built power system simulation model to generate simulated extreme event information; Based on the historical non-extreme event information, the historical extreme event information and the simulated extreme event information, a training data set is constructed; wherein the labels corresponding to the faulty equipment in the simulated extreme event information and the historical extreme event information are historical critical sections, and the labels corresponding to the faulty equipment in the historical non-extreme event information are historical non-critical sections; Performing feature transformation on each data in the training data set to obtain a transformed training data set; The transformed training data set is used as an input variable to train the initialized critical section recognition model, and when it is determined that a stop training condition is met, a trained critical section recognition model is obtained.

7. A critical section identification device in a power system, characterized in that: The device comprises: The acquisition module is used to obtain real-time monitoring data of various equipment in the power system and historical extreme event information; A determination module, configured to determine at least one faulty device from the devices based on the similarity between each of the real-time monitoring data and the historical extreme event information, wherein each of the devices is used as a section and each faulty device is used as a target section; The recognition module is used to input the real-time monitoring data of each target section into the trained key section recognition model for recognition, and obtain the probability that each target section is a key section; The determination module is further used to determine a weight score of each target section, and identify a critical section of the power system from each target section based on the probability and the weight score of each target section.

8. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images