Protection device training platform method and system based on action logic optimization

Abstract

The invention discloses a protection device training platform method and system based on action logic optimization, and relates to the technical field of protection device training, and the method comprises the following steps: obtaining fault detection data and state data when a protection device is protected; constructing a protection action evaluation model based on a bp neural network according to the fault detection data and the state data to obtain a protection action evaluation coefficient; performing optimization adjustment on the action logic of the protection device through a genetic algorithm according to the protection action evaluation coefficient; and constructing a virtual simulation platform by using the model after action logic optimization, and performing training simulation through a plurality of typical fault scenes. The action logic of the protection device is optimized and adjusted through the genetic algorithm, and intelligent improvement of the protection logic is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of relay protection, and more specifically, to a protection device training platform method and system based on action logic optimization. Background Art

[0002] In the power system, relay protection devices play a vital role. They can act quickly when power equipment fails, cut off the faulty part, and protect the safe and stable operation of the power system. However, with the expansion of the scale of the power grid and the continuous updating of technology, the action logic of the relay protection device is becoming increasingly complex, which puts higher requirements on the technical level of relay protection professionals. The existing protection device training platform usually includes a training system for the principle and commissioning of the protection device, a device simulation training system, and an examination system for assessing the learning situation of trainees. These systems meet the training needs of relay protection staff to a certain extent, but still expose some problems in actual application.

[0003] In the training system of protection device principle and debugging, although multimedia technology is used, the training content often focuses on the introduction of theoretical knowledge and lacks in-depth analysis of the action logic of specific protection devices. This makes it difficult for trainees to understand the actual working principles and debugging techniques of protection devices and to apply the knowledge they have learned to actual work.

[0004] In terms of device simulation training systems, although they can simulate the action behavior of relay protection devices when a fault occurs in the power system, the precision and accuracy of the simulation model need to be improved. In addition, existing simulation systems often lack comprehensive coverage of the action logic of protection devices and cannot simulate all possible fault conditions and protection action processes, thus limiting trainees' in-depth understanding and mastery of the action logic of protection devices.

[0005] In summary, the existing protection device training platform has many deficiencies in technology and application, and needs to be optimized and improved. To address these problems, we proposed a protection device training platform method and system based on action logic optimization. This method and system aims to improve trainees' understanding and mastery of the working principle, debugging skills and fault handling ability of protection devices by deeply analyzing the action logic of protection devices, optimizing training content, simulation models and examination systems, and providing strong guarantees for the stable operation of power systems.

[0006] The above disclosed technical solutions have at least the following technical problems: the existing simulation systems often lack comprehensive coverage of the action logic of the protection device and cannot simulate all possible fault conditions and protection action processes, thereby limiting the trainees' in-depth understanding and mastery of the action logic of the protection device. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0007] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a protection device training platform method and system based on action logic optimization, which optimizes the training content and simulation model through in-depth analysis of the action logic of the protection device to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions: A protection device training platform method and system based on action logic optimization include the following steps: obtaining fault detection data and status data of the protection device when protection occurs; constructing a protection action evaluation model based on a BP neural network according to the fault detection data and status data to obtain a protection action evaluation coefficient; optimizing and adjusting the action logic of the protection device through a genetic algorithm according to the protection action evaluation coefficient; using the model after action logic optimization to construct a virtual simulation platform, and conducting training simulation through several typical fault scenarios.

[0009] In a preferred embodiment, the fault detection data includes a fault influence coefficient; the state data includes a current state coefficient and a switch state coefficient; the specific method for obtaining the fault influence coefficient is as follows: set several fault scenarios based on the actual operation of the power system and the action logic of the protection device; simulate several fault scenarios through simulation software or an experimental platform; record the fault detection data of several fault scenario protection devices in real time; divide the fault detection data into equal parts according to the time series to obtain several time-series-based fault detection data; construct a fault tree with several time-series-based fault detection data and use the fault influence coefficient as the top event of the fault tree; assign a probability value to each basic event of the fault tree and perform weighted calculation to obtain the fault influence coefficient of each fault scenario.

[0010] In a preferred embodiment, the specific method for obtaining the current state coefficient is as follows: real-time current data within time t is collected by a current sensor and the real-time current data is fitted into a current curve based on MATLAB; the current data is calibrated and abnormal current values ​​are removed; the calibrated current values ​​are integrated into a current set and the current mean, current standard value and peak current are extracted from the current set as current characteristic values; the current state coefficient is calculated based on the current characteristic values ​​based on a preset current state coefficient calculation formula.

[0011] In a preferred embodiment, the specific method for obtaining the switch state coefficient is as follows: obtain historical state data and perform data cleaning and denoising processing; observe the protection device in the power grid system, and when an abnormality is detected, measure and record the number of operations of the current switch, record the time interval from the detection of the fault to the tripping of the protection device, and calculate the tripping action coefficient; combine the fault current value and the tripping action coefficient to further calculate the switch state coefficient.

[0012] In a preferred embodiment, the action logic of the protection device is optimized and adjusted through a genetic algorithm according to the protection action evaluation coefficient, specifically: based on the protection action evaluation coefficient, the optimization target of the action logic of the protection device is obtained; the action logic of the protection device is decomposed into several parameters, and these parameters are encoded in binary string and real number string encoding form; a group of initial action logics are randomly generated as the initial population of the genetic algorithm; each action logic in the population is evaluated using the protection action evaluation coefficient to obtain its fitness value; according to the fitness value, the action logic with high fitness is selected for iteration to generate a new population and the action logic in the new population is crossover and mutation operations are performed to obtain a new gene combination and action logic; the evaluation and iteration steps are repeated until a predetermined number of iterations is reached or the action logic that meets the optimization target is found; in each iteration, the optimal action logic is selected according to the fitness value, and a new population is generated through crossover and mutation operations; the optimal action logic finally obtained is decoded into a specific action logic, and applied to the protection device.

[0013] In a preferred embodiment, a protection device training platform system based on action logic optimization includes the following modules: a data acquisition module, a protection action evaluation model construction module, an action logic optimization module and a training simulation module; the data acquisition module is used to obtain the fault detection data and status data when the protection device is in protection; the protection action evaluation model construction module is used to construct a protection action evaluation model based on the bp neural network according to the fault detection data and the status data, and obtain the protection action evaluation coefficient; the action logic optimization module is used to optimize and adjust the action logic of the protection device through a genetic algorithm according to the protection action evaluation coefficient; the training simulation module is used to build a virtual simulation platform based on the model after the action logic optimization, and perform training simulation through several typical fault scenarios.

[0014] The technical effects and advantages of the protection device training platform method and system based on action logic optimization of the present invention are as follows: 1. The present invention constructs a protection action evaluation model through a BP neural network, which can efficiently process complex nonlinear relationships, thereby obtaining accurate protection action evaluation coefficients. This method not only improves the accuracy of the evaluation, but also enhances the generalization ability of the model, making it able to adapt to the protection action evaluation under different fault conditions.

[0015] 2. The present invention optimizes and adjusts the action logic of the protection device through the genetic algorithm, thereby realizing the intelligent improvement of the protection logic. As a global optimization search algorithm, the genetic algorithm can quickly find the optimal or suboptimal protection action logic while ensuring convergence, thereby improving the performance and reliability of the protection device. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural schematic diagram of the protection device training platform method based on action logic optimization of the present invention.

[0017] Figure 2 This is a schematic diagram of the system structure of the protection device training platform based on action logic optimization of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Embodiment 1, Figure 1 The present invention provides a protection device training platform method based on action logic optimization, which includes the following steps: Data collection: Obtain fault detection data and status data when the protection device is in protection; Action logic modeling: Based on the fault detection data and status data, a protection action evaluation model is constructed based on the BP neural network to obtain the protection action evaluation coefficient; Action logic optimization: optimize and adjust the action logic of the protection device through genetic algorithm according to the protection action evaluation coefficient; Training simulation: Use the model with optimized action logic to build a virtual simulation platform and conduct training simulation through various typical fault scenarios.

[0020] S1, obtaining fault detection data and status data of the protection device when protection occurs; The fault detection data includes a fault influence coefficient; the state data includes a current state coefficient and a switch state coefficient; The fault impact coefficient is a key indicator used to quantitatively evaluate the impact of faults on protection devices and systems. It helps trainers better understand the impact mechanism of faults by analyzing the impact of faults on the action logic of protection devices, system stability and reliability, so as to more accurately simulate fault scenarios in simulation training, optimize the action logic of protection devices, and improve trainees' fault handling capabilities and the operating efficiency of protection devices in actual work.

[0021] The specific method for obtaining the fault influence coefficient is as follows: Set several fault scenarios based on the actual operation of the power system and the action logic of the protection device; Simulate several fault scenarios through simulation software or experimental platform; Real-time recording of fault detection data of protection devices for several fault scenarios; The fault detection data is divided into equal parts according to the time series to obtain a number of fault detection data based on the time series; Construct a fault tree based on several time series-based fault detection data and use the fault influence coefficient as the top event of the fault tree; A probability value is assigned to each basic event of the fault tree and a weighted calculation is performed to obtain the fault impact coefficient of each fault scenario.

[0022] The specific calculation formula of the fault influence coefficient is:

[0023] in, is the fault influence coefficient, n is the total number of basic events in the fault tree, is the ith basic event in the fault tree, For basic events The probability of occurrence, For basic events The weight in the fault tree, is the fault duration of the basic event.

[0024] The fault influence coefficient has the following benefits for the protection device training platform based on action logic optimization: Accurately simulate fault scenarios: By setting several fault scenarios and simulating based on the actual operation of the power system and the action logic of the protection device, the fault conditions in the real power system can be simulated more accurately. The introduction of the fault impact coefficient enables the simulation process to more accurately reflect the impact of the fault on the power system, thereby improving the authenticity and accuracy of the simulation.

[0025] Optimize the action logic of protection devices: As the top event of the fault tree, the fault impact coefficient can intuitively reflect the impact of the fault on the action logic of the protection device. By simulating the fault scenario through simulation software or experimental platform, the fault detection data of the protection device can be recorded in real time, and the action logic of the protection device can be adjusted and optimized accordingly to make it more in line with the needs of the actual power system.

[0026] Improve training efficiency: The training platform based on fault impact coefficient can more quickly identify key fault scenarios and influencing factors, so as to carry out targeted training. This can not only reduce training time, but also improve training efficiency, so that trainees can master the action logic of protection devices and fault handling skills more quickly.

[0027] Strengthen fault handling ability: By simulating different fault scenarios and calculating fault impact coefficients, trainees can have a deeper understanding of the impact of faults on the power system and the operating principles of protection devices. This helps to enhance trainees' ability to handle faults in actual work and improve the reliability and safety of the power system.

[0028] Feedback and iteration: The training platform based on the failure impact coefficient can collect feedback data from trainees during the training process to evaluate the training effect and optimize the training content. Through continuous feedback and iteration, the functions and performance of the training platform can be gradually improved to make it more in line with actual needs.

[0029] Promoting technological innovation: The calculation method and application of the fault influence coefficient provide new ideas for the technological innovation of the protection device training platform. By deeply studying the calculation method and influencing factors of the fault influence coefficient, it is possible to promote technological innovation in algorithm optimization, data processing and other aspects of the training platform, and further improve its performance and practicality.

[0030] In summary, the fault influence coefficient has significant benefits for the protection device training platform method based on action logic optimization, including improving simulation accuracy, enhancing practicality, and promoting continuous optimization. These benefits will help improve the efficiency and effectiveness of the training platform and provide strong support for the reliable operation and fault handling of the power system.

[0031] The current state coefficient is a comprehensive parameter used to quantitatively describe the specific state or behavior characteristics of current in the power system, which may involve multiple aspects such as current intensity, stability, phase relationship, etc. In the training platform, this coefficient is used to simulate and analyze the changes in current under different fault conditions, so as to help trainees understand the action logic of the protection device and optimize its settings to ensure that when a fault occurs in the power system, the protection device can respond accurately and quickly, effectively remove the fault, and ensure the safe and stable operation of the power system.

[0032] The specific method for obtaining the current state coefficient is as follows: The real-time current data within the time t is collected in real time by the current sensor and the real-time current data is fitted into a current curve based on MATLAB; Perform data calibration on current data and remove abnormal current values; Integrate the calibrated current values ​​into a current set and extract the current mean value, the current standard value and the peak current from the current set as current characteristic values; The current state coefficient is calculated according to the current characteristic value based on a preset current state coefficient calculation formula.

[0033] The specific calculation formula of the current state coefficient is as follows:

[0034] in, is the current state coefficient, is the mean current, is the standard value of current, is the peak current, is the i-th current data in the set, and n is the number of elements in the current data set collected within time t.

[0035] The current state coefficient has the following benefits for the protection device training platform based on action logic optimization: Improve training effect: As a comprehensive indicator of current data, the current state coefficient can more intuitively reflect the operating status of the power system. During the training process, trainees can better understand the operating rules and fault characteristics of the power system by observing the changes in the current state coefficient. This helps to improve trainees' understanding and mastery of the action logic of the protection device, thereby improving the training effect.

[0036] Optimize the action logic of protection devices: The current state coefficient can reflect the real-time operating status of the power system and provide an important basis for optimizing the action logic of protection devices. By analyzing the relationship between the current state coefficient and the action logic of protection devices, the shortcomings of protection devices under specific working conditions can be found and targeted optimization and improvement can be carried out. This helps to improve the reliability and accuracy of protection devices and reduce the risk of false operation and refusal to operate.

[0037] Enhance fault simulation and diagnosis capabilities: On the training platform, the current state coefficient can be used to simulate various fault conditions of the power system. By observing the changes in the current state coefficient before and after the fault occurs, students can better understand the causes and propagation paths of the fault. This helps improve students' ability to diagnose and troubleshoot power system faults, and provides strong support for fault handling in actual work.

[0038] Promoting technological innovation and R&D: The application of current state coefficients provides new ideas and methods for technological innovation and R&D of protection devices. By deeply studying the relationship between the current state coefficient and the operating state of the power system, more intelligent, efficient and reliable protection devices can be developed. This will help promote technological progress and industrial upgrading in the power industry and provide strong guarantees for the safe and stable operation of the power system.

[0039] In summary, the current state coefficient has significant benefits for the protection device training platform based on action logic optimization. It can not only improve the training effect, optimize the action logic of the protection device, enhance the fault simulation and diagnosis capabilities, but also promote technological innovation and research and development. Therefore, in the construction and application of the training platform, full attention should be paid to the application and promotion of the current state coefficient.

[0040] The switch state coefficient is a comprehensive parameter used to describe the state or characteristics of a switch device under specific conditions. It may involve the current state of the switch (such as open or closed), the number of actions, the action time, the stability and other aspects, and is used to quantitatively analyze the performance and behavior of the switch device in the power system. By simulating and analyzing the switch state coefficient, the training platform can evaluate whether the action logic of the protection device is reasonable, thereby helping trainees optimize the settings of the protection device and improve the safety and stability of the power system.

[0041] The specific method for obtaining the switch state coefficient is as follows: Obtain historical status data and perform data cleaning and denoising; Observe the protection devices in the power grid system. When an abnormality is detected, measure and record the number of times the current switch is operated, record the time interval from the detection of the fault to the tripping of the protection device, and calculate the tripping action coefficient; Combining the fault current value with the tripping action coefficient, the switch state coefficient is further calculated.

[0042] The specific calculation formula of the tripping action coefficient is as follows:

[0043] The calculation formula of the switch state coefficient is as follows:

[0044] In the formula, is the tripping action coefficient, is the adjustment factor, and its value is generally between 0 and 1. t is the time interval from when the protection device detects a fault to when it trips. is the number of switch operations in time interval j, is the switching state coefficient, is the total number of measurements, is the current value at the time of fault, is the rated current of the switch.

[0045] The switch state coefficient has the following benefits for the protection device training platform based on action logic optimization: Reflecting the real switch status: The switch status coefficient can more accurately reflect the status of the switch in actual operation by comprehensively considering multiple key parameters of the switch (such as the number of operations, tripping time interval, fault current value, etc.). This enables the training platform to simulate switch actions that are closer to the real situation, thereby improving the fidelity and accuracy of the training.

[0046] Optimize action logic: By analyzing the switch state coefficient, it is possible to identify whether the action logic of the switch under specific conditions is reasonable. Based on these analyses, the action logic of the protection device can be optimized to make it more in line with actual operation requirements.

[0047] Enhance students' understanding of protective devices: By simulating the actions of protective devices under different switch state coefficients, students can more intuitively understand the working principle and action logic of protective devices. This helps students better master the use methods and precautions of protective devices.

[0048] Improve students' troubleshooting ability: On the training platform, students can simulate troubleshooting under different switch state coefficients. Through repeated practice, students can gradually improve the speed and accuracy of troubleshooting, so as to better deal with various emergencies in actual work.

[0049] Provide data support: As an important indicator reflecting the switch state, the switch state coefficient can provide data support for the technical innovation of the protection device. By analyzing a large amount of switch state coefficient data, problems in the design and use of the protection device can be discovered, and directions for its improvement can be provided.

[0050] Promote technological progress: The analysis and application of switch state coefficients can promote the continuous advancement of protection device technology. For example, by optimizing the action logic and algorithm of the protection device, the accuracy and reliability of its fault detection can be improved, thereby ensuring the safe and stable operation of the power system.

[0051] In summary, the switch state coefficient has many benefits for the protection device training platform based on action logic optimization. It not only improves the fidelity and accuracy of the training platform, but also improves the training effect and promotes the innovation and development of protection device technology. Therefore, when constructing and optimizing the protection device training platform, the role and influence of the switch state coefficient should be fully considered.

[0052] S2, constructing a protection action evaluation model based on the BP neural network according to the fault detection data and status data, and obtaining the protection action evaluation coefficient; The specific calculation formula of the protection action evaluation model is:

[0053] in, is the protection action evaluation coefficient, is the fault influence coefficient, is the current state coefficient, is the switching state coefficient, is the preset fault influence coefficient weight factor, is the preset current state coefficient weight factor, is the preset switch state coefficient weight factor.

[0054] S3, optimizing and adjusting the action logic of the protection device through a genetic algorithm according to the protection action evaluation coefficient; The optimization and adjustment of the action logic of the protection device by using a genetic algorithm according to the protection action evaluation coefficient is specifically as follows: Based on the protection action evaluation coefficient, the optimization target of the protection device action logic is obtained; Decompose the action logic of the protection device into several parameters, and encode these parameters into binary string and real number string encoding forms; Randomly generate a set of initial action logics as the initial population of the genetic algorithm; Use the protection action evaluation coefficient to evaluate each action logic in the population and obtain its fitness value; According to the fitness value, select the action logic with high fitness for iteration, generate a new population, and perform crossover and mutation operations on the action logic in the new population to obtain a new gene combination and action logic; Repeat the evaluation and iteration steps until the predetermined number of iterations is reached or the action logic that meets the optimization goal is found; In each iteration, the optimal action logic is selected according to the fitness value, and a new population is generated through crossover and mutation operations; The optimal action logic finally obtained is decoded into specific action logic and applied to the protection device.

[0055] S4, based on the protection action evaluation model after action logic optimization, build a virtual simulation platform and conduct training simulation through several typical fault scenarios.

[0056] Build a virtual simulation platform: Design a system architecture that takes into account both the requirements of virtual reality technology and the needs of the training program. This includes determining the hardware equipment (such as head-mounted displays, handheld controllers, etc.) and software systems (such as modeling and rendering of virtual environments, implementation of user interactions, etc.) required for the platform.

[0057] Data model building: Based on real scenarios or computer-generated data, a power system model is built in a virtual environment. This includes elements such as terrain, buildings, and power equipment, and the scalability of the model is considered to facilitate subsequent adjustments and optimizations based on training needs.

[0058] Interaction design: Design the interaction between users and the virtual environment, such as using handheld controllers to grab objects, operate buttons, etc. At the same time, use visual, sound and other elements to enhance the interaction effect, so that students can experience the operation and fault handling process of the power system in an immersive way.

[0059] Performance optimization: Optimize the performance of the VR training platform to increase frame rate, reduce latency, and reduce screen tearing. This includes reducing the number of polygons, using appropriate texture compression algorithms, and optimizing the rendering process.

[0060] Select typical fault scenarios: According to the actual situation of the power system and training requirements, select several typical fault scenarios for simulation. These scenarios can include different types of faults such as single-phase ground short circuit, two-phase short circuit, three-phase short circuit, etc.

[0061] Set fault parameters: Set fault parameters in the virtual simulation platform, such as fault type, fault location, fault current, etc. These parameters should be able to truly reflect the fault conditions in the actual power system.

[0062] Simulate fault occurrence: Start the virtual simulation platform to simulate the process of fault occurrence. During the simulation, the operating status of the power system and the fault change process can be observed in real time.

[0063] Analyze protection actions: Analyze the actions of the protection device after a fault occurs based on the protection action evaluation model. This includes the accuracy, timeliness, selectivity, etc. of the action.

[0064] Trainees: Guide trainees to perform fault handling operations on the virtual simulation platform. Through repeated practice, trainees can gradually master the methods and techniques of power system fault handling and improve their fault handling capabilities and levels.

[0065] Feedback and optimization: Collect feedback from trainees during the training process and optimize and improve the virtual simulation platform and training content based on these feedback. This helps improve training effectiveness and trainee satisfaction.

[0066] Embodiment 2, a protection device training platform system based on action logic optimization, comprises the following modules: a data acquisition module, a protection action evaluation model building module, an action logic optimization module and a training simulation module; Data acquisition module: used to obtain fault detection data and status data when the protection device is in protection; Protection action evaluation model building module: used to build a protection action evaluation model based on BP neural network according to fault detection data and status data to obtain protection action evaluation coefficient; Action logic optimization module: used to optimize and adjust the action logic of the protection device through genetic algorithm according to the protection action evaluation coefficient; Training simulation module: used to build a virtual simulation platform based on the model optimized by action logic, and conduct training simulation through several typical fault scenarios.

[0067] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0068] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0069] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0070] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0071] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0072] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A protection device training platform method based on action logic optimization, characterized in that: The following steps are involved: Obtain fault detection data and status data when the protection device is in protection; According to the fault detection data and status data, a protection action evaluation model is constructed based on the BP neural network to obtain the protection action evaluation coefficient; According to the protection action evaluation coefficient, the action logic of the protection device is optimized and adjusted through the genetic algorithm; Using the model with optimized action logic, a virtual simulation platform is constructed to conduct training simulations through several typical fault scenarios.

2. The protection device training platform method based on action logic optimization according to claim 1 is characterized in that: The fault detection data includes a fault influence coefficient; the state data includes a current state coefficient and a switch state coefficient; The specific method for obtaining the fault influence coefficient is as follows: Set several fault scenarios based on the actual operation of the power system and the action logic of the protection device; Simulate several fault scenarios through simulation software or experimental platform; Real-time recording of fault detection data of protection devices for several fault scenarios; The fault detection data is divided into equal parts according to the time series to obtain a number of fault detection data based on the time series; Construct a fault tree based on several time series-based fault detection data and use the fault influence coefficient as the top event of the fault tree; A probability value is assigned to each basic event of the fault tree and a weighted calculation is performed to obtain the fault impact coefficient of each fault scenario.

3. The protection device training platform method based on action logic optimization according to claim 2 is characterized in that: The specific method for obtaining the current state coefficient is as follows: The real-time current data within the time t is collected in real time by the current sensor and the real-time current data is fitted into a current curve based on MATLAB; Perform data calibration on current data and remove abnormal current values; Integrate the calibrated current values ​​into a current set and extract the current mean value, the current standard value and the peak current from the current set as current characteristic values; The current state coefficient is calculated according to the current characteristic value based on a preset current state coefficient calculation formula.

4. The protection device training platform method based on action logic optimization according to claim 3 is characterized in that: The specific method for obtaining the switch state coefficient is as follows: Obtain historical status data and perform data cleaning and denoising; Observe the protection devices in the power grid system. When an abnormality is detected, measure and record the number of times the current switch is operated, record the time interval from the detection of the fault to the tripping of the protection device, and calculate the tripping action coefficient; Combining the fault current value with the tripping action coefficient, the switch state coefficient is further calculated.

5. The protection device training platform method based on action logic optimization according to claim 4 is characterized in that: The optimization and adjustment of the action logic of the protection device by using a genetic algorithm according to the protection action evaluation coefficient is specifically as follows: Based on the protection action evaluation coefficient, the optimization target of the protection device action logic is obtained; Decompose the action logic of the protection device into several parameters, and encode these parameters into binary string and real number string encoding forms; Randomly generate a set of initial action logics as the initial population of the genetic algorithm; Use the protection action evaluation coefficient to evaluate each action logic in the population and obtain its fitness value; According to the fitness value, select the action logic with high fitness for iteration, generate a new population, and perform crossover and mutation operations on the action logic in the new population to obtain a new gene combination and action logic; Repeat the evaluation and iteration steps until the predetermined number of iterations is reached or the action logic that meets the optimization goal is found; In each iteration, the optimal action logic is selected according to the fitness value, and a new population is generated through crossover and mutation operations; The optimal action logic finally obtained is decoded into specific action logic and applied to the protection device.

6. The protection device training platform method based on action logic optimization according to claim 5 is characterized in that: The specific calculation formula of the fault influence coefficient is: in, is the fault influence coefficient, n is the total number of basic events in the fault tree, is the ith basic event in the fault tree, For basic events The probability of occurrence, For basic events The weight in the fault tree, is the fault duration of the basic event.

7. The protection device training platform method based on action logic optimization according to claim 6 is characterized in that: The specific calculation formula of the current state coefficient is as follows: in, is the current state coefficient, is the mean current, is the standard value of current, is the peak current, is the i-th current data in the set, and n is the number of elements in the current data set collected within time t.

8. The protection device training platform method based on action logic optimization according to claim 7 is characterized in that: The specific calculation formula of the tripping action coefficient is as follows: The calculation formula of the switch state coefficient is as follows: In the formula, is the tripping action coefficient, is the adjustment factor, and its value is generally between 0 and 1. t is the time interval from when the protection device detects a fault to when it trips. is the number of switch operations in time interval j, is the switching state coefficient, is the total number of measurements, is the current value at the time of fault, is the rated current of the switch.

9. The protection device training platform method based on action logic optimization according to claim 8 is characterized in that: The specific calculation formula of the protection action evaluation model is: in, is the protection action evaluation coefficient, is the fault influence coefficient, is the current state coefficient, is the switching state coefficient, is the preset fault influence coefficient weight factor, is the preset current state coefficient weight factor, is the preset switch state coefficient weight factor.

10. A protection device training platform system based on action logic optimization using any one of claims 1 to 9, characterized in that: It includes data acquisition module, protection action evaluation model building module, action logic optimization module and training simulation module; Data acquisition module: used to obtain fault detection data and status data when the protection device is in protection; Protection action evaluation model building module: used to build a protection action evaluation model based on BP neural network according to fault detection data and status data to obtain protection action evaluation coefficient; Action logic optimization module: used to optimize and adjust the action logic of the protection device through genetic algorithm according to the protection action evaluation coefficient; Training simulation module: used to build a virtual simulation platform based on the model optimized by action logic, and conduct training simulation through several typical fault scenarios.

Images