A new energy automobile electric drive bench test device

Abstract

The application provides a new energy automobile electric drive bench testing device, which has the characteristics of soft and hard integrated combined testing, elimination of testing fragmentation defects, multi-degree-of-freedom flexible loading + wide temperature range simulation, testing working condition highly close to real vehicle, multi-source data intelligent mining + fault prediction, realization of testing digital closed loop and modular layered architecture, adaptation to multiple generations of electric drive iteration, strong universality, full-automatic timing cooperation, safety interlocking integration and the like, which can greatly improve testing authenticity and automation degree and reduce enterprise research and development and real vehicle testing cost.

Description

Technical Field

[0001] This invention relates to the field of testing equipment technology for new energy vehicle power systems, and more specifically to a new energy vehicle electric drive bench testing device that integrates hardware-in-the-loop (HIL) semi-physical simulation, electric drive power stage bench loading, and intelligent mining and analysis of multi-source test data. It is suitable for joint testing of the performance, reliability, extreme conditions, and fault diagnosis of three-in-one / multi-in-one electric drive assemblies, drive motors, and motor controllers (MCUs). Background Technology

[0002] With the rapid expansion of the new energy vehicle industry, electric drive systems, as the core power components of the vehicle, have seen a surge in demand from companies for electric drive assembly simulation verification, bench durability testing, and controller strategy calibration. Currently, mainstream electric drive testing equipment in the industry is divided into two independent types: pure hardware-in-the-loop simulation platforms and traditional electric drive dynamometer benches. Using these separately results in several inherent technical drawbacks. The devices are separate, and the hardware and software tests cannot be synchronized, resulting in severe test distortion. The existing HIL platform only realizes signal-level simulation of VCU / BMS / MCU, without real power load, and can only verify control logic. It cannot reproduce real power-level operating conditions such as motor temperature rise, overload, and torque loss. Traditional dynamometers can only complete single-motor loading tests, lacking a whole-vehicle virtual environment. They cannot simulate slope, cyclic operating conditions, battery voltage fluctuations, and whole-vehicle torque coordination strategies. The test data deviates greatly from the real vehicle, requiring a lot of real-vehicle calibration to make up for it. The development cost is high and the cycle is long. Incomplete coverage of operating conditions and low level of automation in extreme / failure condition testing: Traditional test benches only support steady-state speed and torque testing. Fault conditions and dynamic extreme conditions require manual parameter modification and cannot generate fault injection test cases in batches. Existing loading mechanisms are mostly fixed rigid connections, which cannot simulate the dynamic load disturbances caused by changes in the half-shaft angle and suspension vibration of real vehicles, resulting in distorted temperature rise, efficiency, and NVH test results. Test data is siloed, lacking a unified data-driven evaluation model: HIL simulation data, bench sensor data, and battery simulation data are collected by three separate software programs with incompatible data formats. Only scattered data can be manually exported for simple statistics, and it is impossible to automatically discover potential performance degradation and hidden faults of electric drives. A standardized electric drive performance database has not been established, and there is a lack of data support for component selection and iterative optimization. Poor modularity of the test bench and weak adaptability for model iteration: The existing bench simulation model is decoupled from the hardware loading unit. When adding a new motor type or a new control algorithm, the entire simulation framework needs to be rebuilt, resulting in a long model calibration cycle. Traditional benches do not support automatic calibration or online model updates, making it difficult to adapt to the testing requirements of the new generation of 800V high-voltage electric drives and silicon carbide controllers. The search revealed that CN219608423U and CN112114210A only optimized the dynamometer loading structure, half-shaft alignment mechanism, and environmental temperature chamber. They did not address the core pain points of deep integration between HIL hardware-in-the-loop simulation and power-stage electric drive test bench, data-driven intelligent diagnosis, and full-condition automated joint testing. They lacked an integrated and complete architecture of "simulation modeling - hardware-in-the-loop - power test bench loading - data mining closed loop". Summary of the Invention

[0003] To address the shortcomings of existing technologies, such as the independence of HIL simulation benches and electric drive dynamometers, low simulation accuracy of test conditions, inability to share data, insufficient automated testing capabilities for extreme / fault conditions, and lack of a closed-loop data optimization system, this invention provides a new energy vehicle electric drive bench testing device. This device achieves a fully integrated closed-loop testing process, encompassing virtual simulation modeling, hardware-in-the-loop semi-physical interaction, electric drive power stage loading testing, intelligent mining of multi-source data, and online iterative optimization of the model. This significantly improves the realism and automation of testing, while reducing enterprise R&D and real-vehicle testing costs.

[0004] The technical solution adopted by the present invention to achieve the technical objective is: a new energy vehicle electric drive bench test device, the structure of which includes: a whole vehicle virtual simulation modeling unit, a hardware-in-the-loop real-time interaction unit, an electric drive power loading test unit, a multi-dimensional sensing acquisition unit, a data intelligent mining and fault diagnosis unit, and a collaborative central control scheduling unit.

[0005] As a further improvement of the present invention, the whole vehicle virtual simulation modeling unit, hardware-in-the-loop real-time interaction unit, electric drive power loading test unit, multi-dimensional sensing acquisition unit, data intelligent mining and fault diagnosis unit communicate bidirectionally with the collaborative central control scheduling unit through the EtherCAT real-time bus, and are uniformly controlled by the collaborative central control scheduling unit to form a closed-loop test architecture of simulation-semi-physical interaction-power loading-data mining.

[0006] As a further improvement of the present invention, the whole vehicle virtual simulation modeling unit includes a Simulink multi-domain simulation modeling submodule, a model automatic calibration submodule, and a multi-condition test case automatic generation submodule.

[0007] As a further improvement of the present invention, the hardware-in-the-loop real-time interaction unit is equipped with a physical vehicle controller (VCU), a battery management system (BMS), and a motor controller (MCU), and is equipped with multiple programmable fault injection channels. It interacts with the vehicle virtual simulation modeling unit in real time and sends torque control commands to the electric drive power loading test unit.

[0008] As a further improvement of the present invention, the electric drive power loading test unit includes a bidirectional feedback dynamometer, a multi-degree-of-freedom flexible transmission fixture, a high and low temperature environment simulation chamber, and a driver simulation submodule; the multi-degree-of-freedom flexible transmission fixture integrates an angle fine-tuning mechanism and a vertical vibration simulation module, which is used to simulate the suspension vibration and dynamic changes of the half-shaft tilt angle of a real vehicle.

[0009] As a further improvement of the present invention, the multi-dimensional sensing and acquisition unit is equipped with distributed high-frequency synchronous acquisition sensors to collect multi-dimensional raw data such as electric drive torque, speed, temperature, high voltage signal, vibration and noise in real time and upload them to the data intelligent mining and fault diagnosis unit.

[0010] As a further improvement of the present invention, the data intelligent mining and fault diagnosis unit incorporates a time-series data cleaning submodule, a data-driven performance evaluation model, a latent fault identification model, and a standardized electric drive database submodule.

[0011] As a further improvement of the present invention, the collaborative central control and scheduling unit has a built-in real-time timing synchronization controller and a safety interlock protection module, which uniformly synchronizes the running sequence of each unit, monitors high voltage, temperature, and speed risks in real time, and executes safety protection actions.

[0012] As a further improvement of the present invention, the Simulink multi-domain simulation modeling submodule incorporates an equivalent model of the power battery, a 400V / 800V high-voltage electric drive model, a vehicle dynamics model, and a VCU / BMS collaborative control model, and supports FMI standard co-simulation; the model automatic calibration submodule accesses measured data from multi-dimensional sensor acquisition units and uses the least squares algorithm to automatically correct the simulation model parameters.

[0013] As a further improvement of the present invention, the fault injection channel supports real-time injection of fault signals such as bus overvoltage / undervoltage, current offset, CAN bus frame loss, temperature sensor failure, and torque command sudden change, and generates automated test cases for normal operating conditions, extreme overload conditions, and electrical fault conditions in batches.

[0014] As a further improvement of the present invention, the bidirectional feedback dynamometer loading unit adopts a symmetrical dual dynamometer arrangement, has four-quadrant dynamic loading capability, no torque loading blind zone at low speed, and feeds working energy back to the power grid; the high and low temperature environment simulation chamber has a temperature control range of -40℃ to 150℃, and integrates a water-cooled / oil-cooled cycle simulation system to reproduce the heat dissipation conditions of the electric drive vehicle.

[0015] As a further improvement of the present invention, the multi-degree-of-freedom flexible transmission tooling is equipped with a rapid centering and positioning mechanism, which is compatible with two-in-one, three-in-one, and multi-in-one integrated electric drive assemblies, without the need to replace the special tooling; the tooling servo angle adjustment range is ±15°, and the vibration excitation frequency is 0~50Hz.

[0016] As a further improvement of the present invention, the time-series data cleaning submodule automatically removes abnormal noise points generated by electromagnetic and vibration interference, and unifies the formats of HIL simulation data, sensor measured data, and fault injection log data; the latent fault identification model identifies latent faults such as electric drive bearing wear, power device thermal aging, and increased reducer clearance by comparing the deviation between the simulation theoretical output value and the measured data.

[0017] As a further improvement of the present invention, the standardized electric drive database submodule automatically stores test data of different models of electric drives under full operating conditions, supports the retrieval of electric drive models, test conditions, and performance indicators, and provides data support for vehicle matching and component iterative optimization.

[0018] As a further improvement of the present invention, the collaborative central control and scheduling unit realizes millisecond-level multi-device timing synchronization; the safety interlock protection module monitors bus overcurrent, motor overtemperature, and dynamometer overspeed in real time, and automatically cuts off the high-voltage power supply and unloads the dynamometer load when a risk is triggered.

[0019] As a further improvement of the present invention, the closed-loop test method for electric drive of new energy vehicles is characterized by comprising the following steps: S1 Modeling and Calibration Phase: The vehicle virtual simulation modeling unit builds a vehicle simulation model of the power battery, electric drive, VCU / BMS, imports bench test data to complete automatic model calibration, and generates multiple sets of automated test case examples; S2 Hardware-in-the-Loop Semi-Physical Simulation Stage: Connect the physical VCU, BMS, and MCU to the hardware-in-the-loop real-time interactive unit, and output the vehicle dynamic operating condition signal from the virtual simulation model to the physical controller, and inject various fault signals as needed; S3 power stage bench loading test phase: The physical controller outputs torque commands to the electric drive power loading test unit, the bidirectional dynamometer simulates the dynamic load of the road, the multi-degree-of-freedom flexible fixture replicates the mechanical environment of the vehicle, the high and low temperature chamber simulates the temperature and heat dissipation conditions of the whole vehicle, and the multi-dimensional sensor unit synchronously collects all-dimensional operating data. S4 Data Intelligent Mining and Diagnostic Phase: The data intelligent mining unit cleans multi-source heterogeneous data in a unified manner, quantifies various performance indicators of the electric drive through the performance evaluation model, outputs fault warnings through the hidden fault identification model 53, and archives all test data into the standardized electric drive database. S5 Model Iteration and Optimization Phase: The measured performance and fault data are back-imported into the vehicle virtual simulation modeling unit, the simulation model parameters and VCU / BMS collaborative control strategy are updated online, and S2-S4 are executed in a loop to complete the closed-loop iterative test.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention's HIL semi-physical simulation unit and power-stage electric drive dynamometer real-time linkage architecture breaks through the existing barriers of signal simulation and power loading separation; it simultaneously completes VCU / BMS control strategy calibration and motor temperature rise, overload, and efficiency physical tests. The simulation environment fully matches the dynamics of the actual vehicle, improving the accuracy of test data by more than 40%, reducing actual vehicle calibration tests by 70%, and significantly reducing the enterprise's R&D cycle and testing costs.

[0021] 2. This invention features a flexible transmission fixture that can dynamically adjust the half-shaft tilt angle and simulate suspension vibration, combined with a high and low temperature environment chamber ranging from -40℃ to 150℃. This solves the problem that traditional rigid test benches cannot simulate the dynamic mechanics and temperature changes of real vehicles. The bidirectional feedback dynamometer has no low-speed loading blind spot and can fully cover all extreme working conditions such as starting, climbing, high speed, overload, and short circuit. The temperature rise, efficiency, and NVH test results are true and reliable, representing a technological breakthrough compared to the distortion problem of existing fixed fixture test benches.

[0022] 3. This invention integrates an independent intelligent data mining unit, unifies and connects simulation data, bench sensor data, and fault log data, and constructs a machine learning-driven performance evaluation and hidden fault identification model to automatically discover potential degradation risks in electric drives; at the same time, a standardized electric drive performance database is established to realize test data reuse and rapid component selection; existing benches only have basic data storage functions and lack automated mining, fault prediction, and model iteration closed-loop functions.

[0023] 4. The six modular independent design of the present invention allows the model units to be updated online and tooling to be changed quickly. It is compatible with 400V / 800V silicon carbide electric drives, single / dual motors, and multi-in-one integrated electric drives. When new test conditions or new controllers are added, only the simulation model needs to be updated. There is no need to modify the test bench hardware, which greatly reduces the cost of equipment upgrades and modifications and adapts to long-term iterative testing needs.

[0024] 5. This invention uses a central control unit to uniformly synchronize the timing of all equipment, automatically executing hundreds of sets of working conditions and fault test cases in batches without manual supervision; multiple safety interlocks for high voltage, overspeed, and overtemperature make the test safety significantly better than traditional distributed control bench equipment. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall device architecture of the present invention.

[0026] Figure 2 This is a schematic diagram of the internal structure of the electric drive power loading test unit.

[0027] Figure 3 This is a timing block diagram of the closed-loop test process of the present invention.

[0028] Figure 4 This is a front-view three-dimensional schematic diagram of a multi-degree-of-freedom flexible transmission tooling structure.

[0029] Figure 5 This is a side-view three-dimensional schematic diagram of a multi-degree-of-freedom flexible transmission tooling structure.

[0030] In the diagram: 1-Vehicle virtual simulation modeling unit; 2-Hardware-in-the-loop real-time interaction unit; 3-Electric drive power loading test unit; 4-Multi-dimensional sensing acquisition unit; 5-Data intelligent mining and fault diagnosis unit; 6-Collaborative central control scheduling unit; 7-EtherCAT real-time bus; 31-Bidirectional feedback dynamometer loading unit; 32-Multi-degree-of-freedom flexible transmission fixture; 33-High and low temperature environment simulation chamber; 34-Driver simulation submodule; 51-Time-series data cleaning submodule; 52-Performance evaluation model; 53-Hidden fault identification model; 54-Electric drive database submodule. Detailed Implementation

[0031] The present invention will be further described below with reference to the accompanying drawings: Example: Figures 1 to 5 As shown: This invention provides a test bench for electric drive systems in new energy vehicles. Its structure includes: a whole vehicle virtual simulation modeling unit 1, a hardware-in-the-loop real-time interaction unit 2, an electric drive power loading test unit 3, a multi-dimensional sensing acquisition unit 4, a data intelligent mining and fault diagnosis unit 5, and a collaborative central control scheduling unit 6.

[0032] Among them, the whole vehicle virtual simulation modeling unit 1, hardware-in-the-loop real-time interaction unit 2, electric drive power loading test unit 3, multi-dimensional sensing acquisition unit 4, and data intelligent mining and fault diagnosis unit 5 communicate bidirectionally with the collaborative central control scheduling unit 6 through the EtherCAT real-time bus 7, and are uniformly controlled by the collaborative central control scheduling unit 6 to form a closed-loop test architecture of simulation-semi-physical interaction-power loading-data mining.

[0033] The vehicle virtual simulation modeling unit 1 includes a Simulink multi-domain simulation modeling submodule, a model automatic calibration submodule, and a multi-condition test case automatic generation submodule.

[0034] The hardware-in-the-loop real-time interaction unit 2 is equipped with a physical vehicle controller (VCU), a battery management system (BMS), and a motor controller (MCU). It has multiple programmable fault injection channels, interacts with the vehicle virtual simulation modeling unit 1 in real time, and sends torque control commands to the electric drive power loading test unit 3.

[0035] The electric drive power loading test unit 3 includes a bidirectional feedback dynamometer 31, a multi-degree-of-freedom flexible transmission fixture 32, a high and low temperature environment simulation chamber 33, and a driver simulation submodule 34. The multi-degree-of-freedom flexible transmission fixture 32 integrates an angle fine-tuning mechanism and a vertical vibration simulation module to simulate the suspension vibration and dynamic changes of the half-shaft tilt angle of a real vehicle.

[0036] The multi-dimensional sensing and acquisition unit 4 is equipped with distributed high-frequency synchronous acquisition sensors to collect multi-dimensional raw data such as electric drive torque, speed, temperature, high voltage signal, vibration and noise in real time and upload them to the data intelligent mining and fault diagnosis unit 5.

[0037] The data intelligent mining and fault diagnosis unit 5 includes a time-series data cleaning submodule 51, a data-driven performance evaluation model 52, a latent fault identification model 53, and a standardized electric drive database submodule 54.

[0038] The collaborative central control and scheduling unit 6 has a built-in real-time timing synchronization controller and a safety interlock protection module, which uniformly synchronizes the running sequence of each unit, monitors high voltage, temperature, and speed risks in real time, and executes safety protection actions.

[0039] The Simulink multi-domain simulation modeling submodule includes a built-in equivalent model of the power battery, a 400V / 800V high-voltage electric drive model, a vehicle dynamics model, and a VCU / BMS collaborative control model, supporting FMI standard co-simulation; the automatic model calibration submodule accesses the measured data from the multi-dimensional sensor acquisition unit 4 and uses the least squares algorithm to automatically correct the simulation model parameters.

[0040] The fault injection channel supports real-time injection of fault signals such as bus overvoltage / undervoltage, current offset, CAN bus frame loss, temperature sensor failure, and torque command sudden change, and can generate automated test cases for normal operating conditions, extreme overload conditions, and electrical fault conditions in batches.

[0041] The bidirectional feedback dynamometer loading unit 31 adopts a symmetrical dual dynamometer arrangement, has four-quadrant dynamic loading capability, no torque loading blind zone at low speed, and feeds working energy back to the power grid; the high and low temperature environment simulation chamber 33 has a temperature control range of -40℃ to 150℃, integrates a water-cooled / oil-cooled cycle simulation system, and reproduces the heat dissipation conditions of the electric drive vehicle.

[0042] The multi-degree-of-freedom flexible transmission tooling 32 is equipped with a quick centering and positioning mechanism, which is compatible with two-in-one, three-in-one, and multi-in-one integrated electric drive assemblies, without the need to replace the special tooling; the tooling servo angle adjustment range is ±15°, and the vibration excitation frequency is 0~50Hz.

[0043] The time-series data cleaning submodule 51 automatically removes abnormal noise points caused by electromagnetic and vibration interference, and unifies the formats of HIL simulation data, sensor measured data, and fault injection log data; the latent fault identification model 53 identifies latent faults such as electric drive bearing wear, power device thermal aging, and increased reducer clearance by comparing the deviation between the simulation theoretical output value and the measured data.

[0044] The standardized electric drive database submodule 54 automatically stores full-condition test data of different electric drive models, supports the retrieval of electric drive models, test conditions, and performance indicators, and provides data support for vehicle matching and component iterative optimization.

[0045] Among them, the collaborative central control and scheduling unit 6 achieves millisecond-level multi-device timing synchronization; the safety interlock protection module monitors bus overcurrent, motor overtemperature, and dynamometer overspeed in real time, and automatically cuts off the high-voltage power supply and unloads the dynamometer load when a risk is triggered.

[0046] The closed-loop test method for electric drive of new energy vehicles is characterized by including the following steps: S1 Modeling and Calibration Phase: Vehicle Virtual Simulation Modeling Unit 1 builds a vehicle simulation model of the power battery, electric drive, VCU / BMS, imports bench test data to complete automatic model calibration, and generates multiple sets of automated test case examples; S2 Hardware-in-the-Loop Semi-Physical Simulation Stage: Connect the physical VCU, BMS, and MCU to the hardware-in-the-loop real-time interactive unit 2. The virtual simulation model outputs the vehicle dynamic operating condition signal to the physical controller and injects various fault signals as needed. S3 power stage bench loading test phase: The physical controller outputs torque commands to the electric drive power loading test unit 3, the bidirectional dynamometer simulates the dynamic load of the road, the multi-degree-of-freedom flexible fixture replicates the mechanical environment of the vehicle, the high and low temperature chamber simulates the temperature and heat dissipation conditions of the whole vehicle, and the multi-dimensional sensor unit synchronously collects all-dimensional operating data. S4 Data Intelligent Mining and Diagnosis Phase: The data intelligent mining unit cleans multi-source heterogeneous data in a unified manner, quantifies various performance indicators of the electric drive through the performance evaluation model 52, outputs fault warnings through the hidden fault identification model 53, and archives all test data into the standardized electric drive database. S5 Model Iteration and Optimization Stage: The measured performance and fault data are back-imported into the whole vehicle virtual simulation modeling unit 1, the simulation model parameters and VCU / BMS collaborative control strategy are updated online, and S2-S4 are executed in a loop to complete the closed-loop iterative test.

[0047] The specific functions and operation procedures of this embodiment are as follows: In this invention, the electric drive bench test device for new energy vehicles consists of six units connected bidirectionally via EtherCAT real-time industrial bus 7, and coordinated with the central control and scheduling unit 6 as the main controller to uniformly allocate the working sequence of each unit and achieve millisecond-level synchronization. The whole vehicle virtual simulation modeling unit 1 is equipped with an industrial real-time simulator, pre-installed with Simulink simulation software, and has built-in 400V / 800V power battery, three-in-one electric drive assembly, vehicle dynamics, and VCU / BMS control models; the model automatic calibration submodule reads the measured torque and temperature data from the multi-dimensional sensor acquisition unit 4 via TCP / IP, and automatically iteratively corrects the model's internal resistance and loss parameters; Hardware-in-the-loop real-time interaction unit 2 is equipped with real vehicle controller (VCU), battery management system (BMS), and motor controller (MCU) hardware, and features 32 programmable fault injection channels to simulate various faults in high-voltage circuits, sensors, and CAN bus. Simulation unit 1 outputs virtual battery SOC, vehicle speed, and road resistance signals to the physical controller, which outputs the target torque command in real time and synchronously sends it to electric drive power loading test unit 3 via bus. The electric drive power loading test unit 3 mainly includes a symmetrically arranged bidirectional feedback dynamometer 31 with a rated torque of 0~3000 N·m and a speed of 0~18000 rpm, and four-quadrant dynamic loading; the three-in-one electric drive assembly under test is fixed on a multi-degree-of-freedom flexible transmission fixture 32, which has a built-in servo angle adjustment module and a vertical vibration exciter, which can simulate ±15° half-shaft tilt angle and 0~50Hz suspension vibration; the electric drive under test is placed inside a high and low temperature environment simulation chamber 33, which integrates a water-cooled / oil-cooled circulation simulation system to accurately reproduce the heat dissipation conditions of the electric drive vehicle; the driver simulation submodule 34 includes accelerator, brake, and steering pedals, and collects driver operation commands and transmits them back to the simulation unit; The multi-dimensional sensing and acquisition unit 4 has high-frequency sensors distributed on the output shaft of the dynamometer, the electric drive stator, the IGBT module, and the reducer housing. The sampling frequency is 10kHz. The unit collects raw data of torque, speed, multi-point temperature, high voltage AC and DC current and voltage, and vibration and noise in real time, and transmits them synchronously to the data intelligent mining and fault diagnosis unit 5. The data intelligent mining and fault diagnosis unit 5 is deployed on an industrial server. The time-series data cleaning submodule 51 removes electromagnetic and vibration interference noise through filtering algorithms; the performance evaluation model 52 automatically generates full-condition efficiency MAP and temperature rise decay curves based on the gradient boosting tree algorithm; the hidden fault identification model 53 compares the deviation between simulation theoretical output and measured data to identify potential faults such as bearing wear and power device aging; and the electric drive database submodule 54 automatically archives test data for all models of electric drives and supports retrieval by model, operating condition, and performance index. The collaborative central control and dispatch unit 6 has a built-in real-time FPGA timing controller to unify and synchronize simulation calculations, fault injection, dynamometer loading, and data acquisition timing; it sets up multi-level safety interlock logic to monitor high-voltage overcurrent, motor over-temperature, and dynamometer overspeed, and cut off the high-voltage power supply and unload the dynamometer load in milliseconds when a risk is triggered to avoid equipment damage; Complete closed-loop testing workflow: Step 1: Model Building and Calibration (corresponding to the 2025-2026 phase) Operators build a full-vehicle simulation model of the target vehicle's power battery, electric drive, and VCU / BMS in the whole vehicle virtual simulation modeling unit 1, and input a small amount of preliminary bench test data. The model automatic calibration submodule completes parameter correction and eliminates the initial deviation between simulation and physical models; import standard operating condition test cases such as WLTC, NEDC, hill climbing, and overload, and generate automated test tasks in batches. Step 2: Hardware-in-the-Loop Semi-Physical Co-simulation (2026-2027 Phase) Connect the VCU, BMS, and MCU of the entity under test to the hardware-in-the-loop real-time interaction unit 2, and coordinate with the central control scheduling unit 6 to start timing synchronization; the virtual vehicle model outputs dynamic driving conditions in real time, the fault injection channel injects various fault signals as needed, the physical controller outputs real-time torque control commands, and synchronously transmits them to the electric drive loading unit; Step 3: Power-stage bench loading test. The bidirectional feedback dynamometer loader 31 receives torque commands from the controller and dynamically simulates road load; the multi-degree-of-freedom flexible fixture 32 synchronously adjusts the half-shaft tilt angle and vibration excitation to replicate the stress on the vehicle suspension; the high and low temperature chamber 33 sets the target ambient temperature and cooling flow rate to simulate the high and low temperature working environment of the whole vehicle; the multi-dimensional sensing unit 4 synchronously collects all operating parameters throughout the process and uploads them to the data mining unit in real time. Step 4: Intelligent Analysis of Multi-Source Data and Fault Diagnosis Data Intelligent Mining Unit 5 Unifies the parsing of simulation logs, sensor measured data, and fault injection records, and automatically cleans abnormal data; the performance evaluation model quantitatively outputs electric drive peak efficiency, continuous temperature rise limit, overload tolerance time, and torque response delay indicators; the hidden fault identification model automatically marks data deviation anomalies and outputs fault warning reports; all test data are automatically stored in the standardized electric drive database. Step 5: Online Iterative Optimization of Simulation Model (2027-2028 Phase) The measured performance and fault data output by the data mining unit are imported back into the vehicle virtual simulation modeling unit 1 to update the electric drive loss, thermal model, and control strategy parameters online, and optimize the collaborative control logic of VCU and BMS; the iterated simulation model is then used to issue test tasks again, completing the complete closed loop of "simulation modeling - semi-physical testing - data mining - model optimization", continuously improving simulation accuracy and testing efficiency; Application of 800V silicon carbide electric drive extreme failure testing: 1. The simulation unit builds a dedicated model of an 800V high-voltage battery silicon carbide electric drive and completes model calibration; 2. The HIL unit is connected to an 800V dedicated VCU, BMS, and MCU, and is configured with test cases for bus overvoltage 1000V, IGBT overtemperature, and phase current short circuit fault injection. 3. The dynamometer unit is set with a peak overload of 3000 N·m, the flexible fixture simulates the change of half-shaft tilt angle during high-speed cornering, and the high and low temperature chamber is set with a high temperature durability condition of 120℃. 4. Multi-dimensional sensors collect data on high-voltage circuit temperature rise, component losses, and vibration noise; 5. The data mining model automatically evaluates the thermal decay characteristics of silicon carbide devices, identifies latent aging faults of IGBTs, and stores the complete set of performance data of the 800V electric drive in the database; 6. Update the simulation model of silicon carbide devices based on measured loss data and optimize the high-voltage energy recovery control strategy for the whole vehicle.

[0048] Any technical solution that achieves the above-mentioned technical effects by utilizing the technical solutions described in this invention, or by designing similar technical solutions by those skilled in the art under the inspiration of the technical solutions described in this invention, falls within the protection scope of this invention.

Claims

1. A test bench for electric drive systems of new energy vehicles, comprising: The system comprises a whole vehicle virtual simulation modeling unit (1), a hardware-in-the-loop real-time interaction unit (2), an electric drive power loading test unit (3), a multi-dimensional sensing acquisition unit (4), a data intelligent mining and fault diagnosis unit (5), and a collaborative central control scheduling unit (6), characterized in that: The whole vehicle virtual simulation modeling unit (1), hardware-in-the-loop real-time interaction unit (2), electric drive power loading test unit (3), multi-dimensional sensing acquisition unit (4), and data intelligent mining and fault diagnosis unit (5) communicate bidirectionally with the collaborative central control scheduling unit (6) through the EtherCAT real-time bus (7), and are uniformly controlled by the collaborative central control scheduling unit (6) to form a simulation-semi-physical interaction-power loading-data mining closed-loop test architecture; The whole vehicle virtual simulation modeling unit (1) includes a Simulink multi-domain simulation modeling submodule, a model automatic calibration submodule, and a multi-condition test case automatic generation submodule; The hardware-in-the-loop real-time interaction unit (2) is equipped with a physical vehicle controller (VCU), a battery management system (BMS), and a motor controller (MCU). It has multiple programmable fault injection channels, interacts with the vehicle virtual simulation modeling unit (1) in real time, and sends torque control commands to the electric drive power loading test unit (3). The electric drive power loading test unit (3) includes a bidirectional feedback dynamometer (31), a multi-degree-of-freedom flexible transmission fixture (32), a high and low temperature environment simulation chamber (33), and a driver simulation sub-module (34); the multi-degree-of-freedom flexible transmission fixture (32) integrates an angle fine-tuning mechanism and a vertical vibration simulation module, which is used to simulate the vibration of the real vehicle suspension and the dynamic change of the half-shaft tilt angle; The multi-dimensional sensing and acquisition unit (4) is arranged with high-frequency synchronous acquisition sensors to collect multi-dimensional raw data of electric drive torque, speed, temperature, high voltage signal, vibration and noise in real time and upload them to the data intelligent mining and fault diagnosis unit (5). The data intelligent mining and fault diagnosis unit (5) includes a time-series data cleaning submodule (51), a data-driven performance evaluation model (52), a hidden fault identification model (53), and a standardized electric drive database submodule (54). The collaborative central control and scheduling unit (6) has a built-in real-time timing synchronization controller and a safety interlock protection module, which uniformly synchronizes the running sequence of each unit, monitors high voltage, temperature, and speed risks in real time, and executes safety protection actions.

2. The new energy vehicle electric drive test bench device according to claim 1, characterized in that: The Simulink multi-domain simulation modeling submodule has a built-in equivalent model of the power battery, a 400V / 800V high-voltage electric drive model, a vehicle dynamics model, and a VCU / BMS collaborative control model, and supports FMI standard co-simulation; the model automatic calibration submodule accesses the measured data of the multi-dimensional sensing acquisition unit (4) and uses the least squares algorithm to automatically correct the simulation model parameters.

3. The new energy vehicle electric drive test bench device according to claim 1, characterized in that: The fault injection channel supports real-time injection of fault signals such as bus overvoltage / undervoltage, current offset, CAN bus frame loss, temperature sensor failure, and torque command sudden change, and can generate automated test cases for normal operating conditions, extreme overload conditions, and electrical fault conditions in batches.

4. The electric drive test bench for new energy vehicles according to claim 1, characterized in that: The bidirectional feedback dynamometer loading unit (31) adopts a symmetrical dual dynamometer arrangement, has four-quadrant dynamic loading capability, low-speed no torque loading blind zone, and working energy is fed back to the power grid; the high and low temperature environment simulation chamber (33) has a temperature control range of -40℃ to 150℃, integrates a water-cooled / oil-cooled cycle simulation system, and reproduces the heat dissipation conditions of the electric drive vehicle.

5. The electric drive test bench device for new energy vehicles according to claim 1, characterized in that: The multi-degree-of-freedom flexible transmission tooling (32) is equipped with a quick centering and positioning mechanism, which is compatible with two-in-one, three-in-one, and multi-in-one integrated electric drive assemblies, without the need to replace the special tooling; the tooling servo angle adjustment range is ±15°, and the vibration excitation frequency is 0~50Hz.

6. The electric drive test bench device for new energy vehicles according to claim 1, characterized in that: The time-series data cleaning submodule (51) automatically removes abnormal noise points caused by electromagnetic and vibration interference, and unifies the format of HIL simulation data, sensor measured data, and fault injection log data; the hidden fault identification model (53) identifies hidden faults such as electric drive bearing wear, power device thermal aging, and increased reducer clearance by comparing the deviation between the simulation theoretical output value and the measured data.

7. The electric drive test bench device for new energy vehicles according to claim 1, characterized in that: The standardized electric drive database submodule (54) automatically stores test data of different electric drive models under full operating conditions, supports the retrieval of electric drive models, test conditions, and performance indicators, and provides data support for vehicle matching and component iterative optimization.

8. The new energy vehicle electric drive test bench device according to claim 1, characterized in that: The collaborative central control and scheduling unit (6) achieves millisecond-level multi-device timing synchronization; the safety interlock protection module monitors bus overcurrent, motor overtemperature, and dynamometer overspeed in real time, and automatically cuts off the high-voltage power supply and unloads the dynamometer load when a risk is triggered.

9. The closed-loop test method for electric drive of new energy vehicles according to any one of claims 1-8, characterized in that, Includes the following steps: S1 Modeling and Calibration Stage: Vehicle Virtual Simulation Modeling Unit (1) Build a vehicle simulation model of power battery, electric drive, VCU / BMS, import bench test data to complete automatic model calibration, and generate multiple sets of automated test case examples; S2 Hardware-in-the-Loop Semi-Physical Simulation Stage: Connect the physical VCU, BMS, and MCU to the hardware-in-the-loop real-time interactive unit (2), and the virtual simulation model outputs the vehicle dynamic operating condition signal to the physical controller, injecting various fault signals as needed; S3 power stage bench loading test phase: The physical controller outputs torque command to the electric drive power loading test unit (3), the bidirectional dynamometer simulates the dynamic load of the road, the multi-degree-of-freedom flexible tooling replicates the mechanical environment of the vehicle, the high and low temperature chamber simulates the temperature and heat dissipation conditions of the whole vehicle, and the multi-dimensional sensor unit synchronously collects all-dimensional operating data. S4 Data Intelligent Mining and Diagnosis Stage: The data intelligent mining unit cleans multi-source heterogeneous data in a unified manner, quantifies various performance indicators of electric drive through the performance evaluation model (52), outputs fault warnings through the hidden fault identification model (53), and archives all test data into the standardized electric drive database. S5 Model Iteration Optimization Stage: The measured performance and fault data are imported back into the vehicle virtual simulation modeling unit (1), the simulation model parameters and VCU / BMS collaborative control strategy are updated online, and S2-S4 are executed in a loop to complete the closed-loop iterative test.

Images