Charging pile main power circuit fault diagnosis system construction method based on digital-physical hybrid platform

Abstract

The invention relates to the technical field of fault diagnosis, in particular to a charging pile main power circuit fault diagnosis system construction method based on a digital-physical hybrid platform, and the method comprises the steps: constructing a fault database based on historical fault data; diversified fault features are injected into a charging pile main circuit discrete model in the core processing module through a random extraction strategy, and a fault reaching the standard is dynamically forbidden in combination with a closed-loop verification mechanism, so that a diagnosis model continuously adapts to an unknown fault type. Fault diagnosis models in a core processing module and an auxiliary processing module in an upper computer are respectively optimized by verifying diagnosis results, fault library dynamic injection is combined with double-level diagnosis verification, the limitation of center AI static modeling in the prior art is broken through, and a digital-physical hybrid verification closed loop is constructed, so that the fault diagnosis accuracy is improved. The two times of verification and optimization of the charging pile fault diagnosis model in the hardware-in-the-loop are realized, and the accuracy and the real-time performance of the fault diagnosis of the charging pile main power circuit under the complex working condition are comprehensively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault diagnosis, in particular to a charging pile main power circuit fault diagnosis system construction method based on a digital-physical hybrid platform. BACKGROUND

[0002] With the vigorous development of the new energy industry, electric vehicles have become the core development direction of the industry due to their advantages of being clean, efficient, energy-saving and environmentally friendly, and the scale of supporting charging infrastructure continues to expand. Among them, direct current charging piles have become the mainstream application scenario due to their high-power fast charging capability. However, the high-frequency switching characteristics and complex working conditions of the main power circuit of the charging pile have led to a significant increase in its failure rate, causing equipment downtime, a sharp increase in maintenance costs, and safety hazards, which seriously restricts the popularization process of electric vehicles.

[0003] At present, fault diagnosis methods based on artificial intelligence, such as deep neural networks and BP algorithms, highly depend on historical fault data and pre-defined patterns, and are difficult to dynamically respond to unknown fault types and real-time working condition changes, and lack generalization ability; pure software diagnosis models cannot achieve real-time closed-loop verification at the hardware layer, and the diagnosis results are deployed to physical devices without verification, leading to an increase in misdiagnosis rate and a decrease in robustness; diagnosis model optimization relies on manually labeled data, lacks automatic fault injection and dynamic database update mechanisms, and is difficult to adapt to rapidly evolving fault scenarios.

[0004] Therefore, it is urgent to build a collaborative diagnosis platform that integrates digital simulation and physical hardware, and to achieve self-optimization and hard deployment of fault diagnosis models through closed-loop verification, thereby fundamentally improving diagnosis accuracy, real-time performance, and system reliability. SUMMARY

[0005] The present application provides a charging pile main power circuit fault diagnosis system construction method based on a digital-physical hybrid platform, which can effectively solve the problems in the background art.

[0006] In order to achieve the above purpose, the technical solution adopted by the present application is: A charging pile main power circuit fault diagnosis system construction method based on a digital-physical hybrid platform, comprising the following steps: Constructing a digital simulation model based on the physical circuit of the charging pile, obtaining historical fault data of the charging pile and constructing a fault database; Extracting fault feature signals from the fault database through a controller and inputting them into the digital simulation model for fault simulation and generating a first diagnosis result; Based on the comparison accuracy rate of the first diagnosis result and the fault database, it is determined whether to optimize the digital simulation model; Reconstructing a fault diagnosis model based on the diagnosis module of the digital simulation model; The fault feature simulation signal generated by the digital simulation model is processed by the fault diagnosis model to obtain a second diagnosis result; Based on the comparison accuracy rate of the second diagnosis result and the fault database, it is judged whether to optimize the fault diagnosis model; After updating the fault database, the fault feature signal is continuously extracted, the accuracy rates of the diagnosis results of different fault types are continuously compared, and the charging pile main power circuit fault diagnosis system is output after cyclic optimization.

[0007] Further, the digital simulation model is deployed in an upper computer, and the controller and the fault database are deployed in an auxiliary processing module; The upper computer performs discretization processing on the digital simulation model to obtain a charging pile main circuit discrete model, and deploys the charging pile main circuit discrete model to a core processing module; The upper computer constructs the fault diagnosis model in the auxiliary processing module based on the diagnosis module.

[0008] Further, the controller uses a random extraction strategy to randomly extract the fault feature signals stored in the fault database and injects the control signals into the controller; The controller sends the control signals containing the fault feature signals to the core processing module and controls the charging pile main circuit discrete model to perform fault simulation.

[0009] Further, the core processing module converts the charging pile fault feature data after fault simulation into fault feature simulation signals and sends them to a demonstration module of the upper computer; The demonstration module analyzes the fault feature simulation signals and the charging and grid-connected basic signals of the charging pile, and displays the changes of the current fault feature simulation signals through a waveform diagram.

[0010] Further, the upper computer converts the fault feature simulation signals into digital signals through a data processing module and performs fault diagnosis through the diagnosis module to obtain the first diagnosis result; The upper computer compares the first diagnosis result with the fault types corresponding to the extracted fault feature signals in the fault database, and judges whether to optimize the diagnosis module in the digital simulation model by calculating the accuracy rate.

[0011] Further, the upper computer constructs the fault diagnosis model based on the current diagnosis module and deploys it to the auxiliary processing module through a hardware language; sending a diagnostic control signal to the core processing module through the controller, the core processing module converting a fault feature signal into a fault feature simulation signal and sending the fault feature simulation signal to the fault diagnosis model in the auxiliary processing module to obtain the second diagnostic result; The auxiliary processing module compares the second diagnostic result with the fault feature signal extracted from the fault database, and judges whether to optimize the fault diagnosis model by calculating the accuracy rate.

[0012] Further, when the accuracy rate of the first diagnostic result does not reach a preset standard, a fault feature simulation signal is generated by the charging pile main circuit discrete model to iteratively train the diagnostic module; When the accuracy rate of the second diagnostic result does not reach a preset standard, a fault feature simulation signal is generated by the charging pile main circuit discrete model to iteratively train the fault diagnosis model.

[0013] Further, the upper computer downloads the charging pile main circuit discrete model to the core processing module through a high-speed interface, and the core processing module uploads the fault feature simulation signal to the demonstration module of the upper computer through a high-speed interface.

[0014] Further, the high-speed interface is set as a PCIE interface, and the core processing module and the auxiliary processing module both adopt a programmable logic device FPGA.

[0015] Further, updating the fault database includes adding a disable identifier to a fault type that has been successfully diagnosed, and selecting unverified fault features during subsequent extraction.

[0016] The beneficial effects of the present application are: In the present application, a fault database is constructed based on historical fault data, diversified fault features are injected into the charging pile main circuit discrete model in the core processing module through a random extraction strategy, and a closed-loop verification mechanism is combined to dynamically disable fault types that have reached the standard, so that the diagnostic model continuously adapts to unknown fault types. The software layer diagnostic mechanism in the upper computer and the hardware layer diagnostic mechanism in the auxiliary processing module are optimized through verification of the diagnostic results, and the dynamic injection of the fault library is combined with double-level diagnostic verification, which breaks through the limitations of the existing AI static modeling, and through the construction of a digital-physical hybrid verification closed loop, the charging pile fault diagnosis model is verified and optimized twice in the hardware-in-the-loop, and the accuracy and real-time performance of the charging pile main power circuit fault diagnosis under complex working conditions are comprehensively improved. BRIEF DESCRIPTION OF DRAWINGS

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the construction method of the charging pile main power circuit fault diagnosis system based on the digital-physical hybrid platform in this invention; Figure 2 This is a schematic diagram illustrating the fault diagnosis performed by the host computer through the core processing module on the fault database extracted from the auxiliary processing module in an embodiment of the present invention. Figure 3 This is a schematic diagram of the auxiliary processing module directly diagnosing the core processing module in this invention; Figure 4 This is a flowchart of the fault diagnosis process for the main power circuit of the charging pile in the digital-physical hybrid platform of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly attached to the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementation.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] This invention discloses a method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform. The system comprises three parts: a host computer, a core processing module, and an auxiliary processing module, to perform main power circuit fault diagnosis. The host computer includes a digital simulation model, a demonstration module, and a data processing module; the digital simulation model includes a diagnostic module. The core processing module contains a discrete model of the charging pile main circuit constructed based on the digital simulation model. The auxiliary processing module includes a fault diagnosis model, a fault database, and a controller. The controller sends control signals to control the grid connection, charging, and fault diagnosis of the discrete model of the charging pile main circuit in the core processing module. The host computer performs fault diagnosis on faults extracted from the fault database in the auxiliary processing module through the core processing module.

[0023] like Figures 1 to 4 The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform, as shown, includes the following steps: A digital simulation model is constructed based on the physical circuit of the charging pile. Historical fault data of the charging pile is acquired and a fault database is built. Fault feature signals are extracted from the fault database through the controller and input into the digital simulation model to simulate faults and generate a first diagnostic result. Based on the accuracy of the comparison between the first diagnostic result and the fault database, it is determined whether the digital simulation model should be optimized. The fault diagnosis model is reconstructed based on the diagnostic module of the digital simulation model. The fault diagnosis model processes the fault feature simulation signals generated by the digital simulation model to obtain a second diagnostic result. Based on the accuracy of the comparison between the second diagnostic result and the fault database, it is determined whether the fault diagnosis model should be optimized. After updating the fault database, fault feature signals are extracted again, and the accuracy of the diagnostic results for different fault types is continuously compared. After iterative optimization, the fault diagnosis system for the main power circuit of the charging pile is output.

[0024] The system comprises a digital simulation model deployed in the host computer, and controllers and fault databases deployed in the auxiliary processing module. The host computer discretizes the digital simulation model to obtain a discrete model of the charging pile's main circuit, and then deploys this discrete model to the core processing module. Based on the diagnostic module, the host computer constructs a fault diagnosis model within the auxiliary processing module. The host computer sends the discrete model of the charging pile's main circuit to the core processing module via a high-speed interface, and the core processing module uploads simulated fault characteristic signals to the host computer's demonstration module via the same high-speed interface. The high-speed interface is a PCIe interface, and both the core processing module and the auxiliary processing module utilize FPGA (Field-Programmable Gate Array) logic devices.

[0025] In this invention, a fault database is constructed based on historical fault data. Diverse fault features are injected into the discrete model of the charging pile main circuit in the core processing module using a random sampling strategy. Combined with a closed-loop verification mechanism, faults that have already met the criteria are dynamically disabled, allowing the diagnostic model to continuously adapt to unknown fault types. The software-level diagnostic mechanism in the host computer and the hardware-level diagnostic mechanism in the auxiliary processing module are optimized by verifying the diagnostic results. By employing dynamic fault database injection combined with two-level diagnostic verification, the limitations of existing AI static modeling are overcome. Through the construction of a digital-physical hybrid verification closed loop, the charging pile fault diagnosis model undergoes two verifications and optimizations in the hardware-in-the-loop, comprehensively improving the accuracy and real-time performance of charging pile main power circuit fault diagnosis under complex operating conditions.

[0026] In this embodiment, the controller uses a random sampling strategy to randomly sample fault feature signals stored in the fault database and inject them into the controller's control signals. The controller sends the control signals containing the fault feature signals to the core processing module and controls the discrete model of the charging pile main circuit to perform fault simulation.

[0027] Specifically, such as Figure 2 As shown, the discrete model of the charging pile main circuit is sent to the core processing module via a high-speed interface. The controller in the auxiliary processing module sends a random sampling strategy to the fault database, randomly sampling fault characteristic signals present in the current fault database based on the auxiliary processing module, and injecting them directly as part of the control signal. The controller based on the auxiliary processing module generates control signals and sends them to the core processing module to realize grid connection, charging, and fault diagnosis control of the discrete model of the charging pile main circuit. After receiving the control signals uploaded by the controller based on the auxiliary processing module, the core processing module injects the current charging pile data information into the discrete model of the charging pile main circuit and performs fault simulation.

[0028] In this embodiment, the core processing module converts the fault characteristic data of the charging pile after fault simulation into a fault characteristic simulation signal and sends it to the demonstration module of the host computer. The demonstration module analyzes the fault characteristic simulation signal and the charging and grid connection basic signals of the charging pile, and displays the changes of the current fault characteristic simulation signal through waveform diagram.

[0029] Specifically, the core processing module uses a digital-to-analog converter to convert the charging pile fault characteristic data after fault simulation into fault characteristic simulation signals, and then uploads them to the host computer demonstration module via a high-speed interface. The host computer demonstration module analyzes the basic signals such as charging and grid connection of the current charging pile, as well as the fault characteristic signals, and displays the changes in the current fault characteristic simulation signals through waveform graphs. After completing the real-time demonstration, the real-time demonstration results are transmitted to the host computer data processing module and diagnostic module.

[0030] In this embodiment, the host computer converts the analog fault feature signal into a digital signal through the data processing module, and performs fault diagnosis through the diagnostic module to obtain a first diagnostic result. The host computer compares the first diagnostic result with the fault type corresponding to the fault feature signal extracted from the fault database, and determines whether to optimize the diagnostic module in the digital simulation model by calculating the accuracy.

[0031] Furthermore, the host computer, based on the current diagnostic module, deploys it to the auxiliary processing module through hardware language to form a fault diagnosis model; the controller sends diagnostic control signals to the core processing module, the core processing module converts the fault feature signals into fault feature simulation signals and sends them to the fault diagnosis model in the auxiliary processing module to obtain a second diagnostic result; the auxiliary processing module compares the second diagnostic result with the fault feature signals extracted from the fault database, and determines whether to optimize the fault diagnosis model by calculating the accuracy.

[0032] If the accuracy of the first diagnostic result does not reach the preset standard, the fault feature simulation signal is generated by the discrete model of the charging pile main circuit to iteratively train the diagnostic module; if the accuracy of the second diagnostic result does not reach the preset standard, the fault diagnosis model is generated by the discrete model of the charging pile main circuit to iteratively train the fault diagnosis model.

[0033] Specifically, the host computer data processing and diagnostic module converts the simulated fault characteristic signals after real-time demonstration into digital signals for real-time fault characteristic data processing. The host computer diagnostic module receives the fault characteristics collected by the data processing module and performs fault diagnosis on the current state of the charging pile, obtaining a first diagnostic result. The host computer then performs a cyclical comparison between the first diagnostic result and the fault characteristic signal types extracted from the fault database to confirm the diagnostic accuracy. If the accuracy does not meet the requirements, the diagnostic module in the host computer continues to be trained using the data generated by the core processing module until the accuracy reaches the expected level.

[0034] like Figure 3 As shown, after the host computer data processing and diagnosis module completes a cyclic comparison with the fault database, the diagnosis module in the host computer reconstructs the fault diagnosis model in the diagnosis module of the auxiliary processing module using hardware language. The controller in the auxiliary processing module sends direct diagnostic control signals to the core processing module. The core processing module converts the fault characteristic signals into fault characteristic analog signals through a digital-to-analog converter and sends them to the fault diagnosis model in the auxiliary processing module. The auxiliary processing module cyclically compares the fault diagnosis model with the fault characteristic signals extracted from the pre-established fault database. If the accuracy does not meet the requirements, it continues to train the fault diagnosis model in the auxiliary processing module using the data generated by the core processing module.

[0035] In this embodiment, updating the fault database includes adding a disable flag to fault types that have been successfully diagnosed, and selecting unverified fault features during subsequent extraction.

[0036] Specifically, after the accuracy of two rounds of comparison reaches the standard, the controller based on the auxiliary processing module updates the fault extraction strategy issued to the fault database. In the next round of fault extraction, the fault characteristic signal feature will no longer be extracted. After the fault extraction strategy is updated, the next round of fault extraction is carried out, so that the digital side and the physical side form a closed loop, forming a closed-loop experimental verification of the system. Finally, a safe and efficient construction of a charging pile hardware fault diagnosis system that can be practically applied is completed.

[0037] Those skilled in the art should understand that this invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to this invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a fault diagnosis system for the main power circuit of a charging pile based on a digital-physical hybrid platform, characterized in that, Includes the following steps: A digital simulation model is built based on the physical circuit of the charging pile to obtain historical fault data of the charging pile and construct a fault database. Fault feature signals are extracted from the fault database, input into the digital simulation model to simulate the fault, and a first diagnostic result is generated. Based on the accuracy obtained by comparing the first diagnostic result with the fault type corresponding to the fault feature signal extracted from the fault database, it is determined whether to optimize the digital simulation model. The fault diagnosis model is reconstructed based on the diagnostic module of the digital simulation model or the optimized digital simulation model. The second diagnostic result is obtained by processing the fault feature simulation signal generated by the digital simulation model through the fault diagnosis model or the reconstructed fault diagnosis model. Based on the accuracy obtained by comparing the second diagnostic result with the fault type corresponding to the fault feature signal extracted in this fault database, it is determined whether to optimize the fault diagnosis model. After updating the fault database, continue to extract fault feature signals, continuously compare the accuracy of diagnostic results for different fault types, and output the fault diagnosis system after iterative optimization.

2. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 1, characterized in that, The digital simulation model is deployed on the host computer, and the controller and the fault database are both deployed in the auxiliary processing module; The host computer discretizes the digital simulation model to obtain a discrete model of the charging pile main circuit, and then deploys the discrete model of the charging pile main circuit to the core processing module. The host computer constructs the fault diagnosis model in the auxiliary processing module based on the diagnostic module.

3. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 2, characterized in that, The controller extracts fault feature signals from the fault database. The controller uses a random sampling strategy to randomly extract fault feature signals stored in the fault database and injects them into the controller's control signals. The controller sends a control signal containing fault characteristic signals to the core processing module and controls the discrete model of the charging pile main circuit to perform fault simulation.

4. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 3, characterized in that, The core processing module converts the fault characteristic data of the charging pile after fault simulation into fault characteristic simulation signals and sends them to the demonstration module of the host computer. The demonstration module analyzes the fault characteristic simulation signals and the charging and grid connection basic signals of the charging pile, and displays the changes in the current fault characteristic simulation signals through waveform diagrams.

5. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 4, characterized in that, The host computer converts the analog signals of the fault characteristics into digital signals through the data processing module, and performs fault diagnosis through the diagnostic module to obtain the first diagnostic result. The host computer compares the first diagnostic result with the fault type corresponding to the fault feature signal extracted from the fault database, and determines whether to optimize the diagnostic module in the digital simulation model by calculating the accuracy.

6. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 5, characterized in that, The host computer, based on the current diagnostic module, deploys it to the auxiliary processing module using a hardware language to form the fault diagnosis model. The controller sends a diagnostic control signal to the core processing module, and the core processing module converts the fault characteristic signal into a fault characteristic simulation signal and sends it to the fault diagnosis model in the auxiliary processing module to obtain the second diagnostic result. The auxiliary processing module compares the second diagnostic result with the fault feature signals extracted from the fault database, and determines whether to optimize the fault diagnosis model by calculating the accuracy.

7. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 6, characterized in that, If the accuracy of the first diagnostic result does not reach the preset standard, the diagnostic module is iteratively trained by generating fault feature simulation signals through the discrete model of the charging pile main circuit. If the accuracy of the second diagnostic result does not reach the preset standard, the fault diagnosis model is iteratively trained by generating fault feature simulation signals through the discrete model of the charging pile main circuit.

8. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 4, characterized in that, The host computer sends the discrete model of the main circuit of the charging pile to the core processing module through a high-speed interface, and the core processing module uploads the fault characteristic simulation signal to the demonstration module of the host computer through a high-speed interface.

9. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 9, characterized in that, The high-speed interface is set to a PCIe interface, and both the core processing module and the auxiliary processing module use programmable logic devices (FPGAs).

10. The method for constructing a charging pile main power circuit fault diagnosis system based on a digital-physical hybrid platform according to claim 1, characterized in that, Updating the fault database includes adding a disable flag to fault types that have been successfully diagnosed, and selecting unverified fault features during subsequent extractions.

Images