An intelligent vehicle-mounted combined control system and method for electric vehicles

Abstract

The application discloses an intelligent vehicle-mounted combined control system and method for an electric vehicle, and relates to the technical field of electric vehicle control. The system comprises a chassis domain, an energy domain, a thermal management domain and a central computing platform. The platform is provided with a dynamic flexible boundary negotiation arbitration module. The chassis domain initiates an extreme value power negotiation request under a high-risk working condition. The energy domain outputs a lossless transient power envelope value based on a neural network model. The arbitration module compares the two to obtain an allocated power and decouples the allocated power into a motor torque instruction and a braking force instruction. The system is also provided with a slow and fast time scale layered cooperative control module and a lightweight digital twin transient overrun evaluation module, which are used for evaluating and stimulating the hardware transient overrun capability. The application realizes dynamic two-way negotiation between the chassis domain and the energy domain, eliminates the time scale dislocation of multi-domain response, improves the hardware transient power utilization rate, and improves the vehicle driving stability under extreme working conditions.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle control technology, specifically to an intelligent on-board joint control system and method for electric vehicles. Background Technology

[0002] In the electric vehicle field, the chassis dynamics domain, energy domain, and thermal management domain, which were previously relatively independent control units, are gradually being integrated into a unified on-board computing platform. In existing multi-domain joint control schemes, the battery management system calculates the allowable charge and discharge power boundaries in real time based on state parameters such as cell temperature, state of charge, voltage, and internal resistance, combined with pre-calibrated lookup table rules or fixed safety margin coefficients. This calculation is then sent unilaterally to the drive motor controller and regenerative braking control module via the vehicle controller. When performing vehicle stability control, the chassis domain controller needs to distribute torque and braking force within this fixed boundary range. If the required torque exceeds this boundary, it is forcibly limited according to the boundary, and any insufficient torque is discharged. The above scheme cannot compensate for the loss of power. In this scheme, the energy domain and chassis domain have a one-way and static information transmission relationship. The energy domain does not perceive the instantaneous safety needs of the chassis domain, and the chassis domain cannot provide feedback and negotiation on the boundary of the energy domain. There is no dynamic two-way arbitration mechanism between the two. At the same time, the thermal management system is usually an independent slow adjustment loop, which only makes passive adjustments based on the current temperature. It does not form a unified collaborative optimization framework with the rapid control of the chassis domain and the energy domain. This leads to the risk of instability and slippage of the vehicle due to insufficient power supply in extreme conditions such as high-speed emergency obstacle avoidance on low-temperature icy and snowy roads. Therefore, we propose an intelligent on-board joint control system and method for electric vehicles. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent vehicle-mounted integrated control system and method for electric vehicles.

[0004] The technical problem this invention aims to solve is that in existing multi-domain joint control schemes for electric vehicles, the energy domain issues a rigid power boundary unilaterally based on fixed rules, while the chassis domain lacks the ability to dynamically negotiate with the energy domain. Under high-risk conditions such as high-speed emergency obstacle avoidance on low-temperature icy and snowy roads, when the chassis power domain's demand for instantaneous extreme power exceeds this rigid boundary, it will be forcibly limited, causing the vehicle to become unstable and slip due to insufficient torque or braking force supply. At the same time, the response time scales of the chassis system, electric drive system, and thermal management system differ by orders of magnitude, making it difficult for the existing joint control framework to achieve cross-scale real-time collaborative optimization within a unified system, resulting in control lag. Furthermore, existing technologies lack accurate dynamic evaluation methods for the transient over-limit capabilities of hardware such as batteries and motors under extreme conditions, making it impossible to fully explore the extreme potential of hardware while ensuring safety.

[0005] To address the aforementioned technical problems, this invention provides an intelligent onboard joint control system for electric vehicles, comprising a chassis domain, an energy domain, a thermal management domain, and a central computing platform. The chassis domain, energy domain, and thermal management domain are connected to the central computing platform via an onboard deterministic communication network. The chassis domain collects vehicle speed, yaw rate, lateral acceleration, steering wheel angle, and road adhesion coefficient estimates to determine if the vehicle has entered a high-risk operating condition. Upon determining that the vehicle has entered a high-risk operating condition, it generates an extreme power negotiation request and sends it to the central computing platform. Upon receiving the extreme power negotiation request, the energy domain invokes a pre-set neural network model, using battery state of charge, battery temperature, estimated battery internal resistance, estimated battery open-circuit voltage, and historical current load sequences as inputs, to output a lossless transient power envelope value in real time. The central computing platform is equipped with dynamic flexibility... The boundary negotiation and arbitration module is used to compare and arbitrate the extreme power negotiation request with the lossless transient power envelope value to obtain the allocated power value, and solve the power decoupling optimization problem based on the allocated power value to obtain the drive motor torque command and the line control power command, which are then issued and executed respectively. On this basis, the central computing platform can also set up a multi-timescale joint control module, which predicts the thermal load and generates dynamic boundary constraints for thermal management through a slow timescale submodule, and completes the optimal allocation of torque and braking force within a millisecond period through a fast timescale submodule with the constraints as conditions. The central computing platform can also set up a lightweight digital twin transient over-limit assessment module, which accurately calculates the transient over-limit capability of the system through a first-order thermal capacity-thermal resistance network model, and feeds back the assessment results to the dynamic flexible boundary negotiation and arbitration module to update the lossless transient power envelope value in real time. The present invention provides a corresponding intelligent on-board joint control method for electric vehicles.

[0006] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows:

[0007] 1. This invention replaces the one-way hard boundary restriction with a two-way dynamic negotiation mechanism, which can provide more sufficient instantaneous power support for the chassis domain under high-risk working conditions, effectively avoid vehicle instability caused by power deadlock, and improve driving safety under extreme working conditions;

[0008] 2. This invention achieves an organic combination of early prediction of thermal management boundaries and real-time allocation of rapid control through a hierarchical collaborative control method that decouples slow and fast time scales, thereby eliminating control lag problems and improving the response speed and control accuracy of multi-domain collaborative control.

[0009] 3. This invention uses a lightweight digital twin model to accurately calculate the dynamic heat capacity and heat dissipation flow of the system. Combined with the calculation of the nominal rated power of the battery (115 kW) and the maximum allowable transient over-limit output power (145 kW) in the specific embodiment, it can safely release transient traction or braking force (approximately 26% higher than the nominal value) within a short time window of hundreds of milliseconds, maximizing the utilization of hardware performance limits while ensuring system safety. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the overall closed-loop process of intelligent vehicle-mounted joint control for electric vehicles in an embodiment of the present invention;

[0011] Figure 2 This is a schematic diagram of the dynamic flexible boundary negotiation and arbitration sub-process in an embodiment of the present invention;

[0012] Figure 3 This is a schematic diagram of the lightweight digital twin transient over-limit evaluation process in an embodiment of the present invention. Detailed Implementation

[0013] The following is in conjunction with the appendix Figure 1-3 The specific embodiments of the present invention will be further described below. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0014] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0015] The intelligent vehicle-mounted joint control system for electric vehicles provided in this invention is divided into four main parts in terms of overall architecture: chassis domain, energy domain, thermal management domain, and central computing platform that supports the collaborative operation of the three. The parts interact with each other through vehicle-mounted deterministic communication networks such as vehicle Ethernet and time-sensitive networking. The chassis domain includes drive motor controller, brake-by-wire controller, steering and suspension controller, as well as vehicle status sensors such as wheel speed sensor, yaw rate sensor, and steering wheel angle sensor. The energy domain includes battery management system, battery pack, and voltage, current and temperature acquisition units. The thermal management domain includes battery cooling / heating actuator, motor cooling actuator, air conditioning compressor and its controller, and temperature sensor network. The central computing platform is divided into three logical functional modules: dynamic flexible boundary negotiation and arbitration module, multi-timescale joint control module (divided into slow timescale sub-module and fast timescale sub-module), and lightweight digital twin transient over-limit assessment module. The three functional modules cooperate with each other to form a closed-loop control system.

[0016] I. Dynamic Flexible Boundary Negotiation Mechanism:

[0017] This invention abandons the traditional battery management system's one-way fixed hard boundary arbitration method and introduces a dynamic flexible boundary negotiation arbitration module in the central computing platform to realize bidirectional real-time negotiation between the chassis domain and the energy domain.

[0018] The chassis domain controller periodically collects vehicle dynamic state parameters such as vehicle speed, yaw rate, lateral acceleration, steering wheel angle, and road adhesion coefficient estimate. Based on the preset stability margin threshold, it determines whether the vehicle has entered a high-risk condition. If it is determined to have entered a high-risk condition, the chassis domain controller calculates the extreme total control force and corresponding extreme power requirement required to maintain vehicle stability based on the vehicle stability control target (e.g., maintaining the yaw rate tracking target value), generates an extreme power negotiation request message, and sends it to the dynamic flexible boundary negotiation arbitration module of the central computing platform via the vehicle deterministic communication network.

[0019] Upon receiving the negotiation request, the energy domain control unit invokes a pre-set neural network model. Using the battery's current state of charge, battery temperature, estimated dynamic internal resistance, open-circuit voltage, and recent historical current load sequence as input, it outputs in real-time the lossless transient power envelope value for a short future time window. Its mathematical expression is as follows:

[0020] ;

[0021] In the formula, Indicates at time The output lossless transient power envelope value, This represents the mapping function of a pre-trained neural network. Indicates time The state of charge of the battery. Indicates time Battery temperature, Indicates time The estimated value of the battery dynamic internal resistance. Indicates time The estimated open-circuit voltage of the battery. Indicates from arrive Historical current load sequence within the time window, The duration of this historical window;

[0022] The dynamic flexible boundary negotiation and arbitration module of the central computing platform will handle extreme power negotiation requests issued by the chassis domain. The lossless transient power envelope value of the energy domain output By comparing and arbitrating, the actual allocable power value is obtained:

[0023] ;

[0024] In the formula, This indicates the allocated power value that can be assigned to the chassis domain after arbitration. This represents the extreme power negotiation request value issued by the chassis domain. The meaning is consistent with the above;

[0025] The arbitration module will further allocate power values. Decoupling involves combining the drive motor torque command and the line control power command, and its solution is achieved through the following optimization problem:

[0026] ;

[0027] In the formula, This represents the drive motor torque command obtained from the optimized solution. This represents the linear control power command obtained from the optimized solution. Indicates the torque variable of the drive motor. Indicates the variable of line control dynamics. Indicates the rolling radius of the wheel. This represents the target value of total longitudinal control force required for the chassis domain. Indicates motor torque The corresponding motor power function, The meaning is consistent with the above. , These are the weighting coefficients for the tracking error term and the power constraint term, respectively.

[0028] Finally, the obtained drive motor torque command is calculated. With line control power command The data is sent to the drive motor controller and the brake-by-wire actuator respectively, while the actual dynamic response of the vehicle after execution is continuously collected and fed back to the vehicle dynamic state monitoring link to form a closed-loop correction.

[0029] II. Cloud-edge collaborative multi-timescale model predictive control method:

[0030] To address the time-scale misalignment problem between the chassis, electric drive, and thermal management system, this invention proposes a hierarchical model predictive control method with long and short cycle decoupling, which is divided into a slow time-scale part and a fast time-scale part.

[0031] In the slow timescale portion, cloud servers or high-performance computing cores in vehicles collect heat-related state data such as battery temperature, motor temperature, ambient temperature, current and historical driving condition data (vehicle speed, acceleration, road gradient), and air conditioning load at intervals ranging from seconds to minutes. Based on the current state observations, the reinforcement learning agent uses a pre-trained policy network to predict the heat load sequence for several future slow timescale periods. According to the predicted heat load sequence and the thermal safety thresholds of the battery and motor (such as the upper limit of allowable temperature and the upper limit of allowable temperature rise rate), the dynamic boundary constraints of thermal management for each future period are calculated. The constraint sequence is sent to the edge control node through the vehicle deterministic communication network as a constraint input for fast timescale control, and at the same time drives the battery cooling / heating actuator and the air conditioning system to perform thermal management actions in advance.

[0032] In the fast timescale portion, edge control nodes collect current vehicle state parameters at millisecond intervals, including vehicle speed, yaw rate, longitudinal acceleration, and driver-demanded torque, while applying dynamic boundary constraints for thermal management from the slow timescale layer. As constraints, a model predictive control optimization problem is constructed, which includes the torque of the drive motor and the mechanical friction braking force. The objective function is:

[0033] ;

[0034] In the formula, This represents the objective cost function of model predictive control. Indicates the length of the prediction time domain. Indicates the prediction of the first time domain The system output vector of the step (including vehicle longitudinal acceleration, yaw rate, etc.) Indicates the first The reference target vector corresponding to the step, The weighting matrix representing the output error. Indicates the first The control input vector for each step (including the motor torque increment and braking force increment). The weighted matrix represents the control input. The optimization problem described above is simultaneously constrained by upper and lower limits of motor torque, upper and lower limits of braking force, and other constraints. The constraints of the equivalent power constraint of the conversion;

[0035] The optimization problem is solved offline using polyhedral partitioning technology to obtain a piecewise affine mapping relationship from the state space to the control law and store it as a lookup table. During online operation, the edge control node directly obtains the control increment based on the currently collected state variables through table lookup and affine operation, avoiding the computational delay caused by online iterative solution. The motor torque command and braking force command obtained from the solution are sent to the drive motor controller and the brake-by-wire actuator for execution, and the actual response after execution is fed back to the state acquisition stage of the next control cycle, forming a fast time scale closed-loop control.

[0036] III. Transient Limit Exceedance Capability Assessment Model Based on Lightweight Digital Twin:

[0037] To fully explore the safety potential of hardware under extreme conditions, this invention constructs a lightweight digital twin of battery-motor-thermal management on the vehicle side to accurately calculate the system's transient over-limit capability and feeds the results back to the aforementioned dynamic flexible boundary negotiation module to update the lossless transient power envelope value.

[0038] A lightweight digital twin model of the battery-motor-thermal management system is constructed within the vehicle-side computing platform. This model uses a first-order thermal capacity-thermal resistance network to describe the thermal dynamic characteristics of the battery and motor, and its thermal balance relationship is expressed as follows:

[0039] ;

[0040] In the formula, Indicates the battery's equivalent heat capacity. Indicates time Battery temperature, Indicates time The current flowing through the battery, Indicates time The internal resistance of the battery, This represents the equivalent thermal resistance from the battery to the environment. Indicates time The ambient temperature is given by the equation. The left side of the equation represents the product of the rate of change of battery temperature over time and the heat capacity. The first term on the right side represents the Joule heat power generated by the current passing through the internal resistance, and the second term represents the heat flow lost by the battery to the environment.

[0041] Vehicle-mounted sensors collect real-time data such as battery current, voltage, and surface temperature, synchronously updating the state variables in the digital twin model to ensure consistency between the digital twin and the real physical system. When a transient over-limit request is triggered under high-risk conditions, the digital twin model updates based on the current temperature... With respect to the preset critical temperature for irreversible battery damage Calculate the remaining heat capacity that the system can safely absorb from the current moment:

[0042] ;

[0043] In the formula, This represents the remaining heat margin that the system can safely absorb without exceeding the critical temperature. , , The meaning is consistent with the above. Indicates the moment when the out-of-limit request was triggered;

[0044] Combined with the preset allowed over-limit duration window (e.g., on the order of hundreds of milliseconds), calculate the upper limit of the transient over-limit allowable output power that can be safely released within this time window:

[0045] ;

[0046] In the formula, This represents the calculated upper limit of the transient over-limit allowable output power. This indicates the battery's nominal rated power. , The meaning is consistent with the above;

[0047] The calculated The updated lossless transient power envelope value is fed back to the dynamic flexible boundary negotiation and arbitration module, replacing the original neural network prediction output value or fusing and correcting it. This enables the negotiation module to allocate higher power values ​​under high-risk operating conditions. During the transient over-limit execution phase, the actual temperature change of the battery is continuously monitored in real time. If the temperature rise rate exceeds the preset rate threshold, or if the battery temperature approaches the critical temperature for irreversible damage, the module will take action. If the preset adjacent range is reached, the over-limit output will be immediately terminated and the power will be forcibly restored to the nominal power boundary to avoid permanent hardware damage.

[0048] IV. Closed-loop process of overall control method:

[0049] The three core modules mentioned above work together to form the closed-loop process of the overall control method of this invention: when the vehicle is driving normally, multi-domain conventional collaborative control is performed. The chassis domain continuously detects whether it has entered a high-risk working condition. If no high-risk working condition is detected, the conventional cycle is maintained. At the same time, slow time-scale thermal load prediction and boundary update are performed in parallel. If a high-risk working condition is detected, a dynamic flexible boundary negotiation mechanism is triggered. At the same time, a digital twin transient over-limit capability assessment is triggered in parallel. The outputs of the two are merged into the update link of the lossless transient power envelope value. Then, the torque and braking force distribution link of the fast time-scale model predictive control is entered. Then, the command issuance and vehicle dynamic response feedback link is entered. Finally, the system returns to the vehicle state detection node, forming a complete closed-loop control system covering the three major modules of negotiation, predictive control, and over-limit assessment. Specific Implementation

[0050] Taking the example of a vehicle encountering a sudden obstacle and needing to avoid it at high speed on a cold, icy road, the implementation process of this invention is illustrated. The chassis domain controller detects that the yaw rate deviation exceeds the preset stability margin threshold, determines that the vehicle has entered a high-risk condition, calculates that the extreme power requirement to maintain vehicle stability is approximately 150 kilowatts, generates an extreme power negotiation request and sends it to the central computing platform. The energy domain, based on the current battery state of charge of approximately 45%, battery temperature of approximately 5 degrees Celsius, dynamic internal resistance and open-circuit voltage estimates, outputs a lossless transient power envelope value of approximately 130 kilowatts through a neural network model. The dynamic flexible boundary negotiation arbitration module compares the two and obtains an initial power allocation value of 130 kilowatts. At the same time, the lightweight digital twin module detects the current battery temperature. The preset critical temperature is 32 degrees Celsius. The system calculates the remaining heat margin based on the battery's equivalent heat capacity at 55 degrees Celsius and the preset allowable over-limit duration window (approximately several hundred milliseconds). The calculated maximum allowable output power for transient over-limit is approximately 145 kilowatts. This value is fed back to the negotiation and arbitration module, which updates the allocated power value to 145 kilowatts. The arbitration module then solves the power decoupling optimization problem, decoupling the allocated power value into corresponding drive motor torque commands and line control power commands, and issues them for execution. Simultaneously, the fast-timescale model prediction control module, based on the pre-issued thermal management dynamic boundary constraints from the slow-timescale layer, further optimizes the allocation of torque and braking force within millisecond cycles. During over-limit execution, the system continuously monitors battery temperature changes, confirming that the temperature rise rate does not exceed the preset threshold and is not close to the critical temperature. The over-limit output is maintained until the preset time window ends, automatically returning to the nominal power boundary. Through this process, the vehicle obtains more sufficient instantaneous power support under extreme conditions compared to the original hard-boundary scheme, effectively avoiding instability and slippage caused by insufficient power.

[0051] VI. Alternative Implementation Methods:

[0052] In other embodiments, the lossless transient power envelope value in the dynamic flexible boundary negotiation mechanism may not be output by a neural network model, but instead obtained by solving the battery safety boundary in real time based on an electrochemical mechanism model (such as an equivalent circuit model combined with a lithium plating reaction rate model). This method is closer to the actual physical mechanism of the battery in terms of computational accuracy and can be used as a substitute for or fusion verification of the neural network model. In other embodiments, the slow time-scale thermal load prediction in the cloud-edge collaborative multi-time-scale joint control method may not be based on reinforcement learning, but instead on a rolling time series prediction method based on historical operating condition statistical laws (such as Kalman filtering or autoregressive moving average model), and combined with implicit model predictive control to perform online rolling optimization solution at the edge control node. This method reduces the dependence on the amount of training data and can be used as an alternative implementation method of reinforcement learning scheme in scenarios with limited computing power.

[0053] The above embodiments are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0054] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. An intelligent on-board joint control system for electric vehicles, characterized in that, It includes a chassis domain, an energy domain, a thermal management domain, and a central computing platform. The chassis domain, the energy domain, and the thermal management domain are respectively connected to the central computing platform through an on-board deterministic communication network. The thermal management domain includes battery cooling, heating actuators, and motor cooling actuators for regulating the temperature of the battery and drive motor. The chassis domain includes a drive motor controller, a brake-by-wire controller, and vehicle status sensors, used to collect vehicle speed, yaw rate, lateral acceleration, steering wheel angle, and road adhesion coefficient estimates, to determine whether the vehicle has entered a high-risk operating condition, and to generate an extreme power negotiation request when the vehicle is determined to be in a high-risk operating condition, which is then sent to the central computing platform through the vehicle-mounted deterministic communication network. The energy domain includes a battery management system and a battery pack. Upon receiving the extreme power negotiation request, it calls a preset neural network model, takes the battery state of charge, battery temperature, estimated battery internal resistance, estimated battery open-circuit voltage, and historical current load sequence as inputs, outputs a lossless transient power envelope value in real time, and sends the lossless transient power envelope value to the central computing platform. The central computing platform is equipped with a dynamic flexible boundary negotiation and arbitration module. This module compares and arbitrates the extreme power negotiation request with the lossless transient power envelope value to obtain an allocated power value. The allocated power value is the smaller of the extreme power negotiation request and the lossless transient power envelope value. The dynamic flexible boundary negotiation and arbitration module is also used to obtain drive motor torque commands and line control power commands based on the allocated power value by solving a power decoupling optimization problem with tracking error terms and power constraint terms as optimization objectives. The drive motor torque commands and line control power commands are then sent to the drive motor controller and the line control brake controller for execution, respectively.

2. The intelligent vehicle-mounted joint control system for electric vehicles according to claim 1, characterized in that, The central computing platform is also equipped with a multi-timescale joint control module, which includes a slow timescale sub-module and a fast timescale sub-module. The slow timescale submodule is used to periodically collect battery temperature, motor temperature, ambient temperature and driving condition data. It uses a pre-trained strategy network to predict the heat load sequence in multiple future slow timescale periods. Based on the heat load sequence and the preset thermal safety threshold, it calculates the thermal management dynamic boundary constraints corresponding to each future period. The thermal management dynamic boundary constraints are then sent to the fast timescale submodule through the vehicle deterministic communication network. The fast timescale submodule is used to collect the current state of the vehicle at millisecond intervals, and construct and solve a model predictive control optimization problem that includes the drive motor torque variable and the mechanical friction braking force variable, using the thermal management dynamic boundary constraints as constraints, to obtain torque commands and braking force commands. The torque commands and braking force commands are then sent to the drive motor controller and the brake-by-wire controller for execution, respectively.

3. The intelligent vehicle-mounted joint control system for electric vehicles according to claim 2, characterized in that, The fast timescale submodule solves the model predictive control optimization problem offline using polyhedral partitioning technology to obtain a piecewise affine mapping relationship from the state space to the control law, and stores the piecewise affine mapping relationship as a lookup table. When the fast timescale submodule is running online, it obtains the torque command and the braking force command by looking up the table and performing affine operations based on the collected current state of the vehicle.

4. The intelligent vehicle-mounted joint control system for electric vehicles according to claim 1, characterized in that, The central computing platform is also equipped with a lightweight digital twin transient limit assessment module. The lightweight digital twin transient limit assessment module is used to construct a lightweight digital twin model of the battery-motor-thermal management system. The lightweight digital twin model describes the thermal dynamic characteristics of the battery using a first-order thermal capacity-thermal resistance network. The lightweight digital twin transient over-limit assessment module is also used to synchronously update the state variables of the lightweight digital twin model based on the current, voltage, and temperature data collected in real time by the vehicle-side sensors. When it is determined that a high-risk over-limit request is triggered, it calculates the remaining heat margin that the system can safely absorb based on the current battery temperature and the preset irreversible damage critical temperature of the battery, and calculates the upper limit of the transient over-limit allowable output power in combination with the preset allowable over-limit duration window. The lightweight digital twin transient over-limit assessment module is also used to feed back the upper limit of the transient over-limit allowable output power to the dynamic flexible boundary negotiation and arbitration module as the updated lossless transient power envelope value.

5. The intelligent vehicle-mounted joint control system for electric vehicles according to claim 4, characterized in that, The lightweight digital twin transient over-limit assessment module is also used to monitor the actual temperature change of the battery in real time during the transient over-limit execution phase. When the temperature rise rate exceeds the preset rate threshold, the over-limit output is terminated. When the battery temperature enters the preset range near the critical temperature of irreversible damage to the battery, the over-limit output is also terminated, and the system is restored to the nominal power boundary.

6. A method for intelligent onboard joint control of electric vehicles, applied to an intelligent onboard joint control system for electric vehicles including a chassis domain, an energy domain, a thermal management domain, and a central computing platform, characterized in that, include: The chassis domain collects vehicle dynamic state parameters to determine whether the vehicle has entered a high-risk operating condition. When a high-risk operating condition is detected, the chassis domain generates an extreme power negotiation request and sends it to the central computing platform through the vehicle-mounted deterministic communication network. After receiving the extreme power negotiation request, the energy domain calls a preset neural network model, taking the battery state of charge, battery temperature, estimated battery internal resistance, estimated battery open circuit voltage, and historical current load sequence as inputs, and outputs the lossless transient power envelope value in real time. The central computing platform compares and arbitrates the extreme power negotiation request with the lossless transient power envelope value to obtain the allocated power value, which is the smaller value between the extreme power negotiation request and the lossless transient power envelope value. Based on the allocated power value, the central computing platform obtains the drive motor torque command and the line control power command by solving a power decoupling optimization problem with tracking error term and power constraint term as optimization objectives. The drive motor torque command and the brake-by-wire power command are respectively sent to the drive motor controller and the brake-by-wire controller for execution.

7. The intelligent on-board joint control method for electric vehicles according to claim 6, characterized in that, Also includes: The slow timescale submodule periodically collects battery temperature, motor temperature, ambient temperature and driving condition data, and uses a pre-trained strategy network to predict the heat load sequence in multiple future slow timescale periods. Based on the heat load sequence and the preset thermal safety threshold, it calculates the thermal management dynamic boundary constraints for each future period and sends the thermal management dynamic boundary constraints to the fast timescale submodule. The fast timescale submodule collects the current state of the vehicle at millisecond intervals, and constructs and solves a model predictive control optimization problem that includes drive motor torque variables and mechanical friction braking force variables, using the thermal management dynamic boundary constraints as constraints, to obtain torque commands and braking force commands. The torque commands and braking force commands are then sent to the drive motor controller and the brake-by-wire controller for execution, respectively.

8. The intelligent on-board joint control method for electric vehicles according to claim 6, characterized in that, The model predictive control optimization problem is solved offline using polyhedral partitioning technology to obtain a piecewise affine mapping relationship from the state space to the control law, which is stored as a lookup table. During online operation, the torque command and the braking force command are obtained by looking up the table and performing affine operations based on the collected current state of the vehicle.

9. The intelligent on-board joint control method for electric vehicles according to claim 6, characterized in that, Also includes: A lightweight digital twin model of a battery-motor-thermal management system is constructed, wherein the lightweight digital twin model describes the thermal dynamic characteristics of the battery using a first-order thermal capacity-thermal resistance network. The state variables of the lightweight digital twin model are updated synchronously based on the current, voltage, and temperature data collected in real time by the vehicle-side sensors. When a high-risk over-limit request is triggered, the remaining heat margin that the system can safely absorb is calculated based on the current battery temperature and the preset critical temperature for irreversible battery damage. The upper limit of the transient over-limit allowable output power is calculated in combination with the preset allowable over-limit duration window. The upper limit of the transient over-limit allowed output power is fed back as the updated lossless transient power envelope value.

10. The intelligent on-board joint control method for electric vehicles according to claim 6, characterized in that, Also includes: During the transient over-limit execution phase, the actual temperature change of the battery is monitored in real time. If the rate of temperature rise exceeds the preset rate threshold, the over-limit output is terminated. If the battery temperature enters the preset range near the critical temperature of irreversible damage to the battery, the over-limit output is also terminated, and the system is restored to the nominal power boundary.

Images