Power supply system protection method, system and device and storage medium

Abstract

The invention discloses a power supply system protection method, system and device and a storage medium, and the key points of the technical scheme are that the current health degree of each protection device is calculated according to the current health parameter and historical health degree of each protection device; determining a dynamic threshold according to the current health degrees of all the protection devices, and comparing the current health degree of the main protection device with the dynamic threshold to obtain a comparison result; wherein the main protection device is a protection device which is preset to take effect in advance in all protection devices; and controlling the calculation of the comprehensive score of each protection action of each protection device according to the comparison result, determining an optimal protection action and a target protection device according to all the comprehensive scores, and controlling the target protection device to execute the optimal protection action. According to the invention, misjudgment can be reduced, and the fault detection precision and the system reliability are improved.

Description

Technical Field

[0001] This invention belongs to the field of power supply system technology, and specifically relates to a power supply system protection method, system, device and storage medium. Background Technology

[0002] The power supply system is a system that provides electrical energy to various devices and buildings, and is an important infrastructure to ensure the normal operation of daily life and work.

[0003] To ensure the safe operation of the power supply system, protective devices such as relays are usually installed on the power supply system. Their core function is to quickly and accurately isolate the fault area when a fault occurs, so as to prevent equipment damage, power grid collapse and large-scale power outages.

[0004] However, current power supply system protection typically relies on static thresholds or simple rules to determine the health status of protection devices, which makes it difficult to reflect the health status of devices in complex dynamic environments in real time. Summary of the Invention

[0005] The purpose of this invention is to provide a power supply system protection method, system, device and storage medium that can reduce false alarms and improve the accuracy of fault detection and system reliability.

[0006] The first aspect of this invention provides a power supply system protection method, comprising:

[0007] The current health status of each protection device is calculated based on its current health parameters and historical health status.

[0008] A dynamic threshold is determined based on the current health status of all protection devices. The current health status of the main protection device is compared with the dynamic threshold to obtain a comparison result. The main protection device is the protection device that is predetermined to take effect first among all protection devices.

[0009] Based on the comparison results, the comprehensive score of each protection action of each protection device is calculated, the optimal protection action and the target protection device are determined based on all comprehensive scores, and the target protection device is controlled to execute the optimal protection action.

[0010] In some implementations, the current health parameters include: current current deviation, current temperature deviation, and current vibration spectrum characteristics; the calculation of the current health status of each protection device based on its current health parameters and historical health status includes:

[0011] The current reliability score of each of the protection devices is calculated based on the current current deviation, current temperature deviation, and current vibration spectrum characteristics of each protection device.

[0012] The attenuation factor and the historical health factor are calculated according to the historical health degrees of all the protection devices;

[0013] The current health degree is calculated according to the historical health degrees, the current reliability score, the current attenuation factor, the historical health factor and the cycle time of each protection device.

[0014] In some embodiments, the current attenuation factor is calculated according to the historical health degrees of all the protection devices, comprising:

[0015] A regulation threshold between all the protection devices is preset;

[0016] The difference between the historical health degrees of all the protection devices is calculated, and a current adjacency matrix is established according to the difference between the historical health degrees of all the protection devices;

[0017] The current attenuation factor is calculated according to all the historical health degrees, all the regulation thresholds and the current adjacency matrix.

[0018] In some embodiments, the dynamic threshold is determined according to the current health degrees of all the protection devices, comprising:

[0019] A health degree mean value is calculated according to the current health degrees of all the protection devices;

[0020] The health degree mean value is multiplied by a preset dynamic adjustment coefficient to obtain the dynamic threshold.

[0021] In some embodiments, the calculation of the comprehensive score of each protection action of each protection device is controlled according to the comparison result, comprising:

[0022] In the case that the comparison result is that the current health degree of the primary protection device is less than the dynamic threshold, the current risk urgency degree of each protection device is calculated according to the current health degree of each protection device;

[0023] Each protection action probability of each protection device is calculated according to the current health degree of each protection device based on a proximal policy optimization algorithm;

[0024] The comprehensive score of each protection action of each protection device is calculated according to the current risk urgency degree of each protection device and the protection action probability of each protection device.

[0025] In some embodiments, the current risk urgency degree of each protection device is calculated according to the current health degree of each protection device, comprising:

[0026] The difference between the current health degree of each protection device and a preset reference value is calculated to obtain a first risk term;

[0027] a second risk term is calculated according to a preset adjustment threshold, current health degrees of all protection devices, and an adjacency matrix of the protection devices;

[0028] a current risk emergency degree of each protection device is obtained by weighted calculation of the first risk term and the second risk term of each protection device.

[0029] In some embodiments, the calculation of each protection action probability of each protection device based on the proximal policy optimization algorithm according to the current health degree of each protection device comprises:

[0030] the strategy network parameters of the proximal policy optimization algorithm and the training health degree of the protection device are initialized;

[0031] a training action score is calculated according to the training health degree of the protection device, the training action score is normalized to obtain a training action probability, the probability ratio of the proximal policy optimization algorithm is updated according to the training action probability, the loss function and the strategy network parameters are updated according to the updated probability ratio, and the algorithm is iterated for multiple rounds until convergence;

[0032] each protection action probability of each protection device is calculated by inputting the current health degree of each protection device into the strategy network constructed by the converged proximal policy optimization algorithm.

[0033] The second aspect of the present application provides a power supply system protection system, comprising:

[0034] a health degree calculation module configured to calculate the current health degree of each protection device according to the current health parameter and the historical health degree of each protection device;

[0035] a determination comparison module configured to determine a dynamic threshold according to the current health degrees of all protection devices, compare the current health degree of the main protection device with the dynamic threshold to obtain a comparison result, wherein the main protection device is a protection device that is predetermined to act first among all protection devices;

[0036] a calculation control module configured to control the calculation of the comprehensive score of each protection action of each protection device according to the comparison result, determine an optimal protection action and a target protection device according to all comprehensive scores, and control the target protection device to perform the optimal protection action.

[0037] The third aspect of the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0038] The fourth aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the method.

[0039] The technical solution provided by the present application has the following advantages and effects: the current health degree of the protection device is calculated in real time by fusing the current health parameter and the historical health degree, so as to ensure that the evaluation result accurately reflects the device state, thereby improving the accuracy of fault detection and system reliability. The dynamic threshold is adjusted by the current health degree of all protection devices, and the dynamic threshold is automatically adjusted according to the current health degree, which can reduce false positives compared with the fixed threshold. Through the setting of the dynamic threshold, no matter how the health degree of the group protection device changes, it can be determined whether the main protection device deviates significantly from the health degree of the group protection device. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 is a flowchart of the power supply system protection method provided by the present application;

[0041] Figure 2 is a structural block diagram of the power supply system protection system provided by the present application;

[0042] Figure 3 is an internal structure diagram of the computer device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to facilitate the understanding of the present application, specific embodiments of the present application will be described in more detail below with reference to the accompanying drawings.

[0044] Unless specifically stated or defined otherwise, the terms "first, second,..." used herein are merely used for distinguishing names and do not represent a specific number or order.

[0045] Unless specifically stated or defined otherwise, the term "and / or" used herein includes any and all combinations of one or more related listed items.

[0046] It should be noted that "fixed to" and "connected to" in this paper can be directly fixed or connected to an element, or indirectly fixed or connected to an element.

[0047] As shown in the following Figure 1 The present embodiment provides a power supply system protection method, which includes the following steps S1-S4:

[0048] Step S1, the current health degree of each protection device is calculated according to the current health parameter and the historical health degree of each protection device.

[0049] In practical applications, the current health parameters include: a current current deviation, a current temperature deviation and a current vibration spectrum feature, the current current deviation is a difference between a current collected real-time current value and a reference current value, the current temperature deviation is a difference between a current collected real-time temperature value and a reference temperature value, the current vibration spectrum feature is calculated through a vibration spectrum feature calculation formula, and the vibration spectrum feature calculation formula is:

[0050]

[0051] wherein FFT(f) represents a collected real-time spectrum, FFT ref represents a reference spectrum. A relay in a power supply system is taken as a protection device, a current collected real-time spectrum is substituted into a vibration spectrum feature calculation formula to calculate a current vibration spectrum feature, in other embodiments, the protection device can be a circuit breaker, a fuse, etc., an actual current value of the protection device is measured through a current sensor, a real-time temperature value of the protection device is measured through a temperature sensor, and a real-time spectrum of the protection device is measured through a vibration sensor. To prevent the protection device failure from affecting the safety of the power supply system, the real-time current value, the real-time temperature value and the real-time spectrum of the protection device are usually periodically sampled.

[0052] Specifically, the current health degree of each protection device is calculated according to the current health parameters and the historical health degree of each protection device, including:

[0053] The current reliability score of each protection device is calculated according to the current current deviation, the current temperature deviation and the current vibration spectrum feature of each protection device;

[0054] A current attenuation factor and a historical health factor are constructed according to the historical health degrees of all protection devices;

[0055] The current health degree is calculated according to the historical health degree, the current reliability score, the current attenuation factor, the historical health factor and a cycle time of each protection device.

[0056] In practical applications, the calculation formula of the current health degree is:

[0057] S i (i)(t)=S i (i)(t-Δt)+(γ1R i (i)(t)-γ2Q i (i)(t)+γ3H i (i)(t))·Δt

[0058] wherein S i (i)(t) represents the current health degree of the i th protection device, S i (i)(t-Δt) represents the historical health degree of the i th protection device at the last moment, Ri (t) represents the current reliability score of the i-th protection device, Q i (t) represents the current attenuation factor of the i-th protection device, H i (t) represents the historical health factor of the i-th protection device, which is obtained by querying historical health data, γ1 represents a reliability weight coefficient, γ2 represents an attenuation weight coefficient, and γ3 represents a historical weight coefficient. By fusing the multi-dimensional parameters such as the current reliability score, the current attenuation factor, and the historical health factor with the historical health degree, the current health degree of the protection device is calculated in real time, so as to ensure that the evaluation result accurately reflects the device state, thereby improving the accuracy of fault detection and system reliability.

[0059] The calculation formula of the current reliability score is:

[0060]

[0061] Among them, the R i represents the current reliability score of the i-th protection device, I idev represents the current current deviation of the i-th protection device, T idev represents the current temperature deviation of the i-th protection device, V ibiscore represents the current vibration spectrum feature of the i-th protection device, α1 represents a first reliability coefficient, α2 represents a second reliability coefficient, α3 represents a third reliability coefficient, β1 represents a current deviation factor, and β2 represents a temperature deviation factor. The first reliability coefficient, the second reliability coefficient, and the third reliability coefficient are 0.6, 0.3, and 0.1 respectively, the current deviation factor is 0.5, and the temperature deviation factor is 0.2. In other embodiments, the first reliability coefficient, the second reliability coefficient, the third reliability coefficient, the current deviation factor, and the temperature deviation factor can be adjusted according to actual conditions. The contact arc erosion and coil insulation aging of the protection device can be detected through the current deviation, the coil insulation aging, mechanical jamming, and spring fatigue of the protection device can be detected through the temperature deviation, and the potential faults such as mechanical jamming, spring fatigue, and particle impact of the protection device can be detected through the vibration spectrum feature. Through the multi-parameter fusion of the current deviation, the temperature deviation, and the vibration spectrum feature, most failure modes can be covered, and accurate scoring of the reliability of the protection device can be realized.

[0062] Specifically, the current attenuation factor is calculated according to the historical health degrees of all protection devices, including:

[0063] A preset adjustment threshold is set between all protection devices.

[0064] The difference between the historical health degrees of all protection devices is calculated, and a current adjacency matrix is established according to the difference between the historical health degrees of all protection devices.

[0065] The current attenuation factor is calculated according to all historical health degrees, all adjustment thresholds and the current adjacency matrix.

[0066] Specifically, the calculation formula of the current attenuation factor is:

[0067]

[0068] Wherein, the K ij represents the adjustment threshold between the i th protection device and the j th protection device, the nonlinear influence of propagation, A ij (t) represents the current adjacency matrix, which is a protection device collaborative network topology in the form of a matrix, for the power supply system to select a backup protection device. The adjacency matrix establishes the relationship between the protection devices through the difference between the historical health degree of the i th protection device and the historical health degree of the j th protection device. In this embodiment, δ is 0.15, and in other embodiments, δ can be adjusted according to actual conditions. The adjacency matrix is updated through the difference between the historical health degrees of the protection devices, realizing adaptive adjustment of the protection device collaborative network topology, preventing switching delay and misjudgment.

[0069] Step S2, determining a dynamic threshold according to the current health degree of all protection devices, comparing the current health degree of the main protection device with the dynamic threshold to obtain a comparison result; wherein the main protection device is a protection device that is predetermined to act first among all protection devices, and the protection devices other than the main protection device are backup protection devices.

[0070] In actual application, the dynamic threshold is determined through the current health degree of the protection device, realizing dynamic evaluation of the health degree of the protection device, having the ability of adaptive adjustment according to the current health degree of the protection device and the dynamic change of the network topology, preventing switching delay or misjudgment.

[0071] Specifically, the determination of the dynamic threshold according to the current health degree of all protection devices comprises:

[0072] The health degree mean value is calculated according to the current health degree of all protection devices;

[0073] The health degree mean value is multiplied by a preset dynamic adjustment coefficient to obtain the dynamic threshold.

[0074] Specifically, the calculation formula of the dynamic threshold is:

[0075]

[0076] Wherein, sigma represents a dynamic threshold value, mu represents a dynamic adjustment coefficient, N represents the number of protection devices, the dynamic adjustment coefficient can be set to 0.7, the dynamic threshold value is adjusted through the health degree average, when the health degree average decreases, the dynamic threshold value automatically decreases, and an alarm is given in the case that the main protection device is significantly deteriorated at the group level. Compared with the fixed threshold value, the false alarm can be reduced. Through the setting of the dynamic threshold value, whether the health degree of the group protection device, the main protection device can be determined whether to significantly deviate from the health degree of the group protection device.

[0077] Step S3, according to the comparison result, control the calculation of the comprehensive score of each protection action of each protection device, determine the optimal protection action and the target protection device according to all comprehensive scores, and control the target protection device to execute the optimal protection action.

[0078] Specifically, the control of the calculation of the comprehensive score of each protection action of each protection device according to the comparison result comprises:

[0079] In the case that the comparison result is that the current health degree of the main protection device is less than the dynamic threshold value, the current risk emergency degree of each protection device is calculated according to the current health degree of each protection device;

[0080] According to the current health degree of each protection device, the protection action probability of each protection device is calculated based on the proximal strategy optimization algorithm;

[0081] According to the current risk emergency degree of each protection device and the protection action probability of each protection device, the comprehensive score of each protection action of each protection device is calculated.

[0082] In actual application, in the case that the current health degree of the main protection device is lower than the dynamic threshold value, it means that the power supply system is likely to have problems, and it is necessary to check comprehensively and make relevant response, that is, to trigger the action condition. After triggering the action condition, the current risk emergency degree of each protection device and the protection action probability corresponding to each protection device need to be calculated, the current risk emergency degree of each protection device and the protection action probability corresponding to each protection device are multiplied to obtain the comprehensive score of each protection action of each protection device. The protection action with the highest comprehensive score is taken as the optimal protection action, and the protection device corresponding to the optimal protection action is taken as the target protection device. The target protection device can be the main protection device or the standby protection device, and the target protection device is controlled to execute the optimal protection action to realize the protection of the power supply system.

[0083] Specifically, the calculation of the current risk emergency degree of each protection device according to the current health degree of each protection device comprises:

[0084] a difference between the preset reference value and the current health degree of each protection device is calculated to obtain a first risk term;

[0085] a second risk term is calculated according to a preset adjustment threshold, the current health degrees of all protection devices and the adjacency matrix of the protection devices;

[0086] the first risk term and the second risk term of each protection device are weighted and calculated to obtain a current risk emergency degree of each protection device.

[0087] In practical applications, the calculation formula of the current risk emergency degree is:

[0088]

[0089] wherein, P i (i) represents the current risk emergency degree of the ithprotection device, (1-S i (i)) represents the first risk term, ω1 represents the weight coefficient of the first risk term, and ω2 represents the weight coefficient of the second risk term. In this embodiment, the preset reference value is 1, which means that the worse the current health of the ithprotection device is, the higher the risk priority is. The second risk term means that the worse the health state of the protection devices related to the ithprotection device is, the higher the risk priority of the ithprotection device is. The weight coefficient of the first risk term and the weight coefficient of the second risk term are adjusted according to actual situations.

[0090] Specifically, the calculation of each protection action probability of each protection device based on the proximal policy optimization algorithm according to the current health degree of each protection device comprises:

[0091] initializing the strategy network parameters of the proximal policy optimization algorithm and the training health degree of the protection device;

[0092] calculating a training action score according to the training health degree of the protection device, performing normalization processing on the training action score to obtain a training action probability, updating the probability ratio of the proximal policy optimization algorithm according to the training action probability, updating the loss function and the strategy network parameters according to the updated probability ratio, and performing multiple rounds of iteration until the algorithm converges;

[0093] inputting the current health degree of each protection device into the strategy network constructed by the converged proximal policy optimization algorithm to calculate each protection action probability of each protection device.

[0094] In practical applications, the z i = f θ (S i(t)) Calculate the training action score, in this embodiment, the relay general state only has two actions of opening and closing, but other possibilities are not excluded, so the specific score setting needs to be set according to the actual needs. Then the training action score is normalized to a probability distribution by the Softmax function to obtain the training action probability of each action. The loss function of the proximal policy algorithm is the clipping loss function:

[0095]

[0096] Where, θ represents the policy network parameters, r t (θ) represents the probability ratio, and ε represents the clipping range, such as 0.2. The probability ratio is the output result of the Softmax function, and the probability ratio is represented as:

[0097]

[0098] Where, represents the training action probability of the last round of iteration corresponding to the training action, and π θ (a t |S t ) represents the training action probability of the current round of iteration corresponding to the training action. The probability ratio is updated using the training action probability of the current round, and the loss function and the policy network parameters are updated using the updated probability ratio. In the next round of iteration, the training action probability is updated, thereby changing the output of the Softmax function.

[0099] After the proximal policy optimization algorithm converges, the current health degree of each protection device is substituted into the policy network constructed by the converged proximal policy optimization algorithm to obtain the action score of each protection action of each protection device. The action score is normalized to an action probability by the Softmax function, and the calculation formula of the normalization processing is:

[0100]

[0101] Where, π i represents the action probability of each protection action of the i-th protection device, a k represents the k-th action, S i (t) represents the current health degree of the i-th protection device, z i,k represents the action score of the k-th action of the i-th protection device. For example, the current health degree of the i-th relay is obtained by the policy network to obtain the action score:

[0102] z i = [2.1, 0.5, -1.0, 1.3]

[0103] According to the action score, the action probability is calculated:

[0104]

[0105] After calculation, we get:

[0106] π i =[0.55,0.15,0.03,0.27]

[0107] According to the calculation result, it is known that the relay currently selects the first action (may be "off") with the highest probability of 55%, and selects the fourth action (such as "load shedding") with a probability of 27%. The system traverses all protection device nodes i, calculates the probability of each protection action, selects the action with the highest protection action probability and the corresponding protection device, and executes the action. By selecting the protection device with the highest probability, the most reliable protection of the power supply system is realized.

[0108] As shown in Figure 2 , the embodiment of the present application further provides a power supply system protection system, comprising:

[0109] a health degree calculation module 10, configured to calculate the current health degree of each protection device according to the current health parameter and the historical health degree of each protection device;

[0110] a determination comparison module 20, configured to determine a dynamic threshold value according to the current health degree of all protection devices, compare the current health degree of the main protection device with the dynamic threshold value, and obtain a comparison result; wherein the main protection device is a protection device that is predetermined to act first among all protection devices;

[0111] a calculation control module 30, configured to control the calculation of the comprehensive score of each protection action of each protection device according to the comparison result, determine an optimal protection action and a target protection device according to all comprehensive scores, and control the target protection device to execute the optimal protection action.

[0112] The above-mentioned modules of the power supply system protection system can be realized by software, hardware and combinations thereof in whole or in part. The above-mentioned modules and units can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned modules.

[0113] As shown in Figure 3 , the embodiment of the present application discloses a computer device, comprising a memory and a processor, wherein the memory stores a computer program;

[0114] The computer device can be a server, and its internal structure diagram can be as shown in Figure 3As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. 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, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The network interface of the computer device is used to communicate with the external terminal through the network connection. The computer program is executed by the processor to implement the power supply system protection method described in the above embodiments.

[0115] Those skilled in the art can understand that, Figure 3 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0116] The embodiment of the present application also discloses a computer readable storage medium storing a computer program, wherein the computer program causes a computer to execute the power supply system protection method described in the above embodiments.

[0117] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. 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 above-mentioned embodiment methods. Any reference to memory, storage, database or other medium used in the embodiments provided by the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM) and the like.

[0118] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described, however, any combination of the technical features is considered to be within the scope of the present specification.

Claims

1. A method of protecting a power supply system, characterized by, The application relates to a method for determining an optimal protection action of a plurality of protection devices, comprising the following steps: calculating a current health degree of each protection device according to a current health parameter and a history health degree of each protection device; determining a dynamic threshold according to the current health degrees of all protection devices, comparing the current health degree of a main protection device with the dynamic threshold to obtain a comparison result, wherein the main protection device is a protection device that is scheduled to act first among all protection devices; controlling calculation of a comprehensive score of each protection action of each protection device according to the comparison result, determining an optimal protection action and a target protection device according to all comprehensive scores, and controlling the target protection device to execute the optimal protection action.

2. The power system protection method of claim 1, wherein, The current health parameter comprises a current current deviation, a current temperature deviation and a current vibration spectrum feature; the current health degree of each protection device is calculated according to the current health parameter and the history health degree of each protection device, and the calculation comprises the following steps: calculating a current reliability score of each protection device according to the current current deviation, the current temperature deviation and the current vibration spectrum feature of each protection device; calculating a decay factor and a history health factor according to the history health degrees of all protection devices; calculating the current health degree of each protection device according to the history health degree, the current reliability score, the current decay factor, the history health factor and a cycle time of each protection device.

3. The power system protection method of claim 2, wherein, The current decay factor is calculated according to the history health degrees of all protection devices, and the calculation comprises the following steps: presetting an adjustment threshold between all protection devices; calculating a difference between the history health degrees of all protection devices, and establishing a current adjacency matrix according to the difference between the history health degrees of all protection devices; calculating the current decay factor according to all history health degrees, all adjustment thresholds and the current adjacency matrix.

4. The power system protection method of claim 1, wherein, The dynamic threshold is determined according to the current health degrees of all protection devices, and the determination comprises the following steps: calculating a health degree mean value according to the current health degrees of all protection devices; multiplying the health degree mean value by a preset dynamic adjustment coefficient to obtain the dynamic threshold.

5. The power system protection method of claim 1, wherein, The calculation of the comprehensive score of each protection action of each protection device is controlled according to the comparison result, and the calculation comprises the following steps: in the case that the comparison result is that the current health degree of the main protection device is less than the dynamic threshold, calculating a current risk emergency degree of each protection device according to the current health degree of each protection device; calculating a protection action probability of each protection device of each protection device according to the current health degree of each protection device based on a near-end strategy optimization algorithm; calculating a comprehensive score of each protection action of each protection device according to the current risk emergency degree of each protection device and the protection action probability of each protection device of each protection device.

6. The power system protection method of claim 5, wherein, The current risk emergency degree of each protection device is calculated according to the current health degree of each protection device, and the calculation comprises the following steps: calculating a first risk item by subtracting a preset reference value from the current health degree of each protection device; calculating a second risk item according to a preset adjustment threshold, the current health degrees of all protection devices and an adjacency matrix of the protection device; performing weighted calculation on the first risk item and the second risk item of each protection device to obtain the current risk emergency degree of each protection device.

7. The power system protection method of any one of claims 1-6, wherein, The protection action probability of each protection device is calculated based on a proximal policy optimization algorithm according to the current health degree of each protection device, and the protection action probability of each protection device is calculated based on the proximal policy optimization algorithm according to the current health degree of each protection device, including: The strategy network parameters of the proximal policy optimization algorithm and the training health degree of the protection device are initialized; The training action score is calculated according to the training health degree of the protection device, the training action probability is obtained by normalizing the training action score, the probability ratio of the proximal policy optimization algorithm is updated according to the training action probability, the loss function and the strategy network parameters are updated according to the updated probability ratio, and the algorithm converges after multiple iterations; The current health degree of each protection device is input into the strategy network constructed by the converged proximal policy optimization algorithm, and the protection action probability of each protection device is calculated.

8. A power system protection system characterized by, Including: The health degree calculation module is configured to calculate the current health degree of each protection device according to the current health parameter and the historical health degree of each protection device; The determination and comparison module is configured to determine a dynamic threshold according to the current health degree of all protection devices, compare the current health degree of the main protection device with the dynamic threshold to obtain a comparison result, and determine the comparison result; wherein the main protection device is a protection device that is predetermined to act first among all protection devices; The calculation control module is configured to control the calculation of the comprehensive score of each protection action of each protection device according to the comparison result, determine the optimal protection action and the target protection device according to all comprehensive scores, and control the target protection device to execute the optimal protection action.

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1-7.

10. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1-7.

Images