Method and system for controlling three-in-one electric drive system of electric vehicle

Abstract

The invention relates to the technical field of electric vehicle control, and discloses an electric vehicle three-in-one electric drive system control method and system, and the method comprises the steps: obtaining vehicle data, and recognizing a driving condition based on the vehicle data; triggering a control prediction model according to a driving condition, constructing a multi-target cost function based on vehicle data, and calculating a wheel end target torque and a required motor output power; and the wheel end target torque and the needed motor output power are issued to a controller, a control instruction is generated, and three-in-one electric drive cooperative execution is achieved. The method is complete in structure, the control link is closed-loop, the method is suitable for dynamic scheduling control of a high-integration-level electric drive system, and the method has the comprehensive technical advantages of being high in real-time performance, high in energy efficiency, self-adaptive in thermal management and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle control, and in particular to a control method and system for a three-in-one electric drive system of an electric vehicle. Background Art

[0002] As the power density of electric vehicles continues to rise, three-in-one electric drives—highly integrated motors, electronic controls, and reducers—are gradually replacing separate solutions and becoming the mainstream. However, existing production models often use fixed Eco / Sport modes or simple table-lookup PI control. These systems are unable to recognize complex driving scenarios in real time and dynamically adjust the torque-power mapping. Furthermore, they passively limit current only when device temperatures reach a threshold, leading to high energy consumption, sudden drops in power output, and localized overheating.

[0003] In addition, the electro-thermal-mechanical coupling of the motor, inverter and gearbox exhibits strong nonlinear characteristics across time scales; traditional single-time domain or hierarchical control is difficult to meet the driver's instantaneous power needs in milliseconds, while proactively constraining temperature rise safety within seconds, and lacks the ability to continuously adapt as the vehicle ages and the environment changes. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: a method for controlling a three-in-one electric drive system of an electric vehicle, comprising: acquiring vehicle data, and identifying driving conditions based on the vehicle data; The control prediction model is triggered according to the driving conditions, and a multi-objective cost function is constructed based on vehicle data to calculate the target wheel-end torque and required motor output power. The wheel-end target torque and required motor output power are sent to the controller to generate control instructions, realizing the coordinated execution of the three-in-one electric drive.

[0006] As a preferred solution of the electric vehicle three-in-one electric drive system control method described in the present invention, the vehicle data includes power pedal, brake master cylinder, steering wheel, vehicle body inertia, wheel speed, road longitudinal slope, battery state of charge and ambient temperature data.

[0007] As a preferred solution of the control method of the three-in-one electric drive system of an electric vehicle described in the present invention, wherein: the identification of driving conditions includes stacking vehicle data in chronological order to form a three-dimensional tensor ,Will Input three layers of one-dimensional convolution for feature extraction, and input the convolution output into a single-layer gated recurrent unit network. After softmax transformation, the driving condition probability distribution is obtained, and the driving condition and confidence level corresponding to the maximum component are output. The driving condition probability distribution corresponds to six types of driving conditions, including urban constant speed, start-stop following, strong acceleration, high-speed cruising, uphill and downhill.

[0008] As a preferred embodiment of the control method for a three-in-one electric drive system of an electric vehicle according to the present invention, wherein: triggering the control prediction model according to the driving condition includes taking the driving condition, confidence level and vehicle data as inputs of the control prediction model; The control prediction model is a dual-time domain prediction model, divided into a fast time domain and a slow time domain. The slow time domain takes precedence. At the beginning of the current control cycle, the slow time domain uses a four-node RC thermal network to perform a rolling prediction of the temperature evolution in the next 1 second. The highest predicted temperature and the limit temperature are extracted to calculate the temperature rise margin factor. The energy consumption weight and temperature control weight are dynamically adjusted according to the temperature rise margin factor. After obtaining the temperature rise margin factor, the fast time domain maps the temperature rise margin factor into wheel-end torque, current, motor speed, and PWM frequency, while simultaneously solving for the optimal pedal and brake transients. The wheel-end torque, current, motor speed, and PWM frequency actually issued in the fast time domain are immediately fed back to the next prediction in the slow time domain, achieving information synchronization between the two domains.

[0009] As a preferred solution of the control method of the three-in-one electric drive system of an electric vehicle described in the present invention, the slow time domain includes abstracting the main heat source and heat dissipation path of the three-in-one electric drive into four equivalent thermal nodes, namely the motor stator, the inverter module, the reducer gearbox and the radiator coupled with the environment, reading the motor stator temperature, the inverter module temperature, the reducer gear oil temperature and the radiator outlet temperature based on the vehicle data, and forming a four-node temperature vector ; Calculate the heating power of the four nodes based on the control instructions issued in the previous fast domain cycle: ; in, represents the column vector of the heating power of the four nodes at time k; Indicates the motor stator copper loss; represents the inverter switching loss; Indicates the mechanical friction and oil shear loss of the reducer; Represents matrix transpose; Considering each node as an isothermal block with heat capacity, heat transfer between nodes is only through heat conduction or forced convection, and establishing a resistance and capacitance equivalent network for the four nodes: ; in, represents the heat capacity of node i; represents the instantaneous temperature of node i; express About time The derivative of Represents the node index; represents the instantaneous temperature of node j; represents the equivalent thermal resistance between node i and node j; represents the self-heating power of node i; Using forward Euler discretization with a fixed step size of Δt=20 ms, at time Discretization yields: ; in, Represents the temperature vector of the four nodes predicted to the next moment; The column vector representing the four-node temperature at time k; represents the discrete sampling step; represents the heat capacity diagonal matrix; represents the four-node overall admittance matrix; represents the environmental coupling column vector; Indicates the ambient temperature; represents the kth discrete moment; Take the maximum predicted value among the four node temperatures , calculate the temperature rise margin factor: ; in, Indicates the temperature rise margin factor; Indicates the set limit temperature; represents the maximum predicted value among the four node temperatures; The temperature rise margin factor is used to adjust the weight of the cost function in the multi-objective cost function in real time, and the upper limits of the wheel end torque, current and PWM frequency are proportionally reduced to establish a soft constraint upper bound table; A multi-objective cost function is constructed using weighted quadratic performance indicators to optimize and output the temperature rise margin factor, cost function weight, and soft constraint upper bound table.

[0010] As a preferred solution of the control method of the three-in-one electric drive system of an electric vehicle described in the present invention, wherein: the fast time domain includes, according to the vehicle data, obtaining the time The vehicle linear speed, motor mechanical speed and wheel end output torque executed in the previous control cycle are used to establish a small signal linear electromechanical coupling model; Based on the temperature rise margin factor output in the slow time domain and the soft constraint upper bound table, a quadratic cost function is constructed. The weight in the quadratic cost function is the weight of the cost function output in the slow time domain, and the first-step control quantity is obtained through the solver. The first-step control quantity is added to the current working point to form the target current and target gear. Then, the d-axis current command is generated in combination with the field weakening principle. The voltage command is calculated through the current loop. The target PWM frequency is: ; in, Indicates the target PWM frequency; Indicates taking the minimum value; Indicates the upper limit of PWM frequency of slow time domain output; Indicates the reference switching frequency of the vehicle calibration; Calculate the target wheel-end torque and required motor output power and send them for execution.

[0011] A control system for a three-in-one electric drive system of an electric vehicle using any of the methods described in the present invention, wherein: an identification module acquires vehicle data and identifies driving conditions based on the vehicle data; The calculation module triggers the control prediction model according to the driving conditions, builds a multi-objective cost function based on vehicle data, and calculates the target wheel-end torque and required motor output power; The control module sends the wheel-end target torque and required motor output power to the controller, generates control instructions, and realizes the coordinated execution of the three-in-one electric drive.

[0012] Beneficial effects of the present invention: The present invention realizes the coordinated optimization of torque output and electric-thermal constraints under different driving conditions by constructing a dual-time domain model predictive control framework that integrates multi-modal real-time perception, rapid identification of working conditions and temperature rise margin regulation. The slow time domain prediction module is combined with the thermal network model to evaluate the temperature rise trend of key nodes in advance and dynamically generate soft constraints and weight scheduling factors; the fast time domain optimization module accurately tracks driving needs at a high frequency, generates the optimal current and shifting strategy within feasibility, and realizes millisecond-level response and thermal safety parallel control. The method of the present invention has a complete structure and a closed-loop control link. It is suitable for dynamic scheduling control of highly integrated electric drive systems and has comprehensive technical advantages such as strong real-time performance, high energy efficiency, and adaptive thermal management. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0014] Figure 1 An overall flow chart of a control method for a three-in-one electric drive system of an electric vehicle provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0015] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0016] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a control method for a three-in-one electric drive system of an electric vehicle, comprising: S1: Acquire vehicle data and identify driving conditions based on the vehicle data.

[0017] Furthermore, multimodal real-time data acquisition is performed within the vehicle's central controller, synchronously acquiring 22 raw signals at a fixed 10ms cycle, including those from the power pedal, brake master cylinder, steering wheel, vehicle inertia, wheel speed, road grade, battery state of charge, and ambient temperature. These signals are provided by Hall effect sensors, piezoelectric pressure transducers, magnetic encoders, six-axis IMUs, wheel speed sensors, GNSS air pressure fusion modules, battery management systems, and external temperature sensors. These signals are uniformly triggered via onboard buses such as TTCAN, LIN, and SENT, and timestamped using the IEEE-1588 protocol.

[0018] Specifically, the 22 original signals are power pedal opening, power pedal opening rate of change, brake master cylinder pressure, brake master cylinder pressure rate of change, steering wheel angle, steering wheel angular velocity, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle body yaw angular velocity, vehicle linear velocity, front left wheel speed, front right wheel speed, rear left wheel speed, rear right wheel speed, road longitudinal slope angle, battery state of charge SOC, battery pack total current, battery pack total voltage, maximum cell temperature, ambient temperature, body roll angular velocity and body pitch angular velocity.

[0019] The collected signal first undergoes local second-order Butterworth denoising and five-point differential derivative calculations. It is then normalized based on the range constant stored in the EEPROM to form a standardized vector. To ensure data integrity, the present invention employs dual-channel redundancy and implements CRC-16 checksums for critical signals, providing immediate compensation and fault flagging for missing samples or out-of-bounds anomalies. Finally, the processed vector is written to a shared memory area for subsequent condition identification steps. The raw data is retained in a ring buffer for 2 seconds to support algorithm self-learning and fault tracing.

[0020] Furthermore, within every 10ms control cycle, the vehicle central control unit (VCU) automatically reads the most recent 1s (50 frames) of normalized data vectors from a circular buffer, stacks them chronologically into a three-dimensional tensor, and feeds this into a pre-configured lightweight convolutional-recurrent hybrid neural network model. This model first extracts local feature variations of each signal on short timescales using a three-layer one-dimensional convolutional network. It then aggregates and models global temporal dependencies across signals within a 1s period using a gated recurrent unit (GRU). The model ultimately outputs a log-probability vector of length six, which is then softmax-transformed to produce probability distributions representing six driving conditions (urban constant speed, stop-start following, strong acceleration, high-speed cruising, uphill and downhill). The GRU model utilizes a publicly available gated recurrent unit model. After parameter quantization and pruning, it enables real-time inference within an automotive-grade MCU using less than 3kB of memory, without requiring modifications to the GRU architecture itself.

[0021] It should be noted that the six driving conditions can be mapped to core segments of current mainstream driving cycles (WLTP, NEDC, US FTP-75, etc.), representing the vast majority of the vehicle's statistical distribution. Discretely classifying complex, continuous driving conditions into six typical scenarios significantly reduces the complexity and computational effort of the driving condition identification model, meeting the real-time inference requirements of automotive-grade MCUs. Through this classification, the present invention dynamically adjusts the weights and constraints in the model predictive control cost function based on differences in four dimensions: power demand curve, energy flow, thermal load, and mechanical transmission characteristics. This allows for optimally efficient, responsive, and thermally safe coordinated electric drive control in a variety of complex scenarios.

[0022] To ensure the reliability of the recognition results, the present invention prefers to use the maximum probability value as the operating condition confidence level. When the confidence level is not less than 0.50, the corresponding operating condition label is directly output with the confidence level attached. If the confidence level is less than 0.50, the system temporarily reverts to the default "standard operating condition" label and sets the confidence level to 0.50 to maintain the continuity and stability of the control strategy. The inference latency of the entire recognition process is strictly limited to less than 3ms. The recognition results are encapsulated into a structure along with the timestamp and written to the VCU shared memory area for asynchronous reading by the next model predictive control module within 1ms. At the same time, the original tensor and recognition results are retained for 2s to support subsequent incremental learning and fault tracing.

[0023] S2: Trigger the control prediction model according to the driving conditions, build a multi-objective cost function based on vehicle data, and calculate the wheel-end target torque and required motor output power.

[0024] Furthermore, the vehicle's central controller aggregates all key state information within a single control cycle to form a complete input vector. This vector not only includes the driving condition label and its confidence level from the previous step, but also incorporates vehicle speed, motor speed, reference torque derived from pedal mapping, road grade, battery state of charge and charge / discharge current, as well as ambient temperature and the real-time temperatures of the three major heat sources in the electric drive: the motor stator, inverter power module, and reducer gearbox.

[0025] Based on the complete input vector, this paper constructs a dual-time domain model predictive control framework to meet the dual requirements of instantaneous dynamic response and second-level thermal and energy foresight. The framework consists of two levels of predictors, fast domain and slow domain, which operate sequentially within the same control loop, feeding back data to ultimately output a coherent set of target control variables.

[0026] The slow-domain predictor is activated first. This predictor uses a 1-second (50-step) rolling window with a 20ms step size. It embeds a four-node RC linear thermal network, simplified for automotive specifications, to predict the temperature evolution of the motor, inverter, gearbox, and heat sink over the next second. During the calculation process, the slow-domain predictor maps the gear position, PWM frequency, and current-torque command from the fast domain into node heat generation power. After the prediction is complete, the controller extracts the peak node temperature and calculates the temperature rise margin factor based on the device's extreme temperature and ambient temperature. This factor, ranging from 0 to 1, intuitively describes the thermal margin for the next second: a margin approaching 0 indicates that the temperature is about to peak, requiring immediate power reduction; a margin approaching 1 indicates minimal thermal risk, allowing for appropriate easing of power output. The thermal margin refers to the difference between the maximum heat output that can be removed and the actual heat output under the current environment and cooling conditions. The slow domain then injects the temperature-rise margin factor into its own multi-objective cost function, continuously scheduling energy consumption weights, temperature control weights, and soft current-limit thresholds, while also providing the next cycle's gear switching and PWM frequency-limiting strategies. This completes a proactive balance of the energy-temperature-rise coupling conflict and pushes the calculated temperature-rise margin factor, updated weights, and upper limit parameters to the fast domain in real time.

[0027] Specifically, the main heat sources and heat dissipation paths of the three-in-one electric drive are abstracted into four equivalent thermal nodes, namely the motor stator, inverter module, reducer gearbox and radiator coupled with the environment. Based on the vehicle data, the motor stator temperature, inverter module temperature, reducer gear oil temperature and radiator outlet temperature are read to form a four-node temperature vector .

[0028] Calculate the heating power of the four nodes based on the control instructions issued in the previous fast domain cycle: ; in, represents the column vector of the heating power of the four nodes at time k; Indicates the motor stator copper loss; represents the inverter switching loss; Indicates the mechanical friction and oil shear loss of the reducer; Denotes the matrix transpose. There are established methods for calculating motor copper losses, inverter switching losses, and gearbox losses, so we will not elaborate on them in this article. They can be calculated using Joule's law and the approximate linear relationship between IGBT / SiC device switching energy and frequency.

[0029] Considering each node as an isothermal block with heat capacity, heat transfer between nodes is only through heat conduction or forced convection, and establishing a resistance and capacitance equivalent network for the four nodes: ; in, represents the heat capacity of node i; represents the instantaneous temperature of node i; express About time The derivative of Represents the node index; represents the instantaneous temperature of node j; represents the equivalent thermal resistance between node i and node j; represents the self-heating power of node i. The self-heating of the heat sink node is zero, and the convection heat dissipation can be equivalent to the thermal resistance between the heat sink and the external environment temperature.

[0030] In order to achieve millisecond-level rolling prediction on automotive-grade MCU, forward Euler discretization with a fixed step size of Δt=20 ms is adopted. Discretization yields: ; in, Represents the temperature vector of the four nodes predicted to the next moment; The column vector representing the four-node temperature at time k; represents the discrete sampling step; represents the heat capacity diagonal matrix; represents the four-node overall admittance matrix; represents the environmental coupling column vector; Indicates the ambient temperature; represents the kth discrete moment. By pre-merging the constant term, it can be written in the standard linear discrete state space form: ; in, represents the state transition matrix; represents the heating power input matrix; Represents the ambient temperature input matrix. Since A, B, and E only depend on 、 and Δt, which can be calculated offline during the calibration phase and then solidified into lookup table constants. During runtime, each step requires only a single 4×4 matrix-vector multiplication (approximately 40 MAC instructions), completing a single-step recursion in less than 0.05ms. After multiple iterations, the temperature trajectory for the next 1 second is obtained, from which the predicted peak temperature is extracted to calculate the temperature rise margin factor λ.

[0031] In order to express the heat dissipation margin from the extreme temperature with a single scalar, the present invention uses the device extreme safety temperature (including the factory safety margin) as the upper bound and the current ambient temperature as the lower bound, normalizes the predicted peak value, and takes the maximum predicted value among the four node temperatures. , calculate the temperature rise margin factor: ; in, Indicates the temperature rise margin factor; Indicates the set limit temperature; Represents the maximum predicted value among the four node temperatures. =1, it means there is almost no temperature rise pressure in the next 1s; When it approaches 0, it means the temperature is about to reach its peak and the power output needs to be reduced immediately. Thermal safety information can be transmitted using a single fixed-point number, greatly reducing the amount of data exchange and model dimensions between the fast and slow domains.

[0032] The temperature rise margin factor is injected into the slow time domain cost function and the constraint upper limit in real time, so that energy consumption, temperature control and power balance can be continuously adjusted with the thermal margin, which refers to the remaining space between the current or predicted temperature and the allowable limit temperature. The controller adopts a linear scheduling law: ; in, represents the energy consumption weight; represents the initial energy consumption weight; Indicates the temperature control weight; Indicates the initial temperature control weight. The bigger the car, the more important the energy consumption is. The larger the value, the more important the temperature control is. When the thermal margin is reduced, the energy consumption weight is weakened and the temperature control weight is amplified. At the same time, the physical upper bounds of the wheel end torque, current and PWM frequency are also soft-contracted by the same factor, and a soft constraint upper bound table is established. As the temperature decreases, the corresponding upper limit is reduced proportionally to ensure that the power peak is automatically reduced when the temperature approaches the limit without any abrupt hard current limiting.

[0033] After completing the reconstruction of weights and soft constraints, the slow domain uses weighted quadratic performance indicators to construct a multi-objective cost function: ; in, represents the total cost function of the slow-horizon rolling predictive control; Represents the prediction step index; 、 、 、 It represents the weight coefficient of the cost function, which is scheduled online with the temperature rise margin factor; Indicates the Stepper motor outputs mechanical power; Indicates the Step 1 prediction of wheel end torque; The reference torque representing the synchronization step; Indicates the Step 1: Predict the motor stator temperature; Indicates the motor safety temperature threshold; Represents the difference between two consecutive steps of predicted torque. The constraint set consists of the upper bound of the soft-contracted constraint and the temperature hard constraint. The temperature hard constraint is derived from component specifications and vehicle reliability calibration and is fixed in the controller in the form of a calibration table.

[0034] Since the thermal network has been explicitly recursively derived in the previous step, the optimization problem consists solely of linearized dynamics and a convex quadratic objective. The controller employs a gradient-projection-multiple-processor (GP-MPC) algorithm: Using the optimal solution from the previous cycle as the initial value, the controller converges to the current local optimum within 0.3ms after 6-8 iterations, outputting the total control variable. It's important to note that this total control variable does not directly drive the actuators; rather, it serves as a slow, global scheduling decision—informing the fast domain in advance of the available thermal margin, torque upper bound, gear preference, and weight coefficients for the current cycle. This ensures that the fast domain remains within the safe feasible region during millisecond-level refinement of control, maintaining consistency with thermal management objectives.

[0035] Optimization results are immediately transmitted via a real-time bus: target wheel-end torque, target motor power, gear shift instructions, and target PWM switching frequency. Within the same control cycle, the temperature rise margin factor, updated weight set, and soft constraint upper bound table are written to shared memory for direct access by the fast domain in the next 10ms loop. The current, gear, and frequency actually executed in the fast domain are then transmitted back to the slow domain for the next round of thermal power estimation, achieving closed-loop consistency of thermal, electrical, and mechanical information. This repetitive cycle allows the vehicle to respond to driver power demands within milliseconds while maintaining proactive control of temperature rise risks and energy consumption within the next 1s. This fundamentally addresses the technical challenge of traditional fixed-strategy strategies that struggle to balance power, energy consumption, and temperature control.

[0036] Furthermore, the fast-domain predictor immediately takes over the control law solution upon obtaining the latest temperature-rise margin factor. The fast domain uses a 200ms (10-step) window, still rolling with a 20ms step size. However, the internal dynamics retain only the linearized motor-vehicle transmission model, ignoring the slow thermal network, significantly reducing the state dimensionality. The controller first soft-scaling the upper bounds of the allowable torque, current, and PWM frequency according to the temperature-rise margin factor: the smaller the margin, the more the upper bounds are shrunk, ensuring that instantaneous commands are always subject to thermal safety constraints. The fast domain then constructs a quadratic cost function centered on precise reference torque tracking and torque rate of change suppression. It then invokes a fast quadratic programming solver based on hard real-time multiplier iteration. In less than 0.6ms, the target wheel-end torque, target motor power, gear shift command, and target PWM switching frequency for the current control cycle are determined. These results are immediately distributed to actuators such as the motor controller, inverter, and two-speed reducer, driving the vehicle to deliver power as planned.

[0037] Specifically, according to the vehicle data, obtain the time The vehicle linear speed, motor mechanical speed and wheel-end output torque executed in the previous control cycle are used to establish a small-signal linearized electro-mechanical coupling model.

[0038] According to the temperature rise margin factor and soft constraint upper bound table output in the slow time domain, a quadratic cost function is constructed, which is expressed as: ; in, represents the total cost of the fast time domain objective function; Represents the instantaneous torque error weight coefficient; Represents the torque change rate weight; Indicates the Step 1 predicts the wheel end torque. It should be noted that The torque error weight is expressed in both the fast and slow time domain cost functions, but in the slow domain, it is the torque error weight within the control cycle, while in the fast domain, it is the instantaneous torque error weight. β controls the torque error weight, while η limits the torque ramp rate. Both are automatically scaled by the weight vector provided by the slow domain, ensuring smoother output when thermal margin is insufficient.

[0039] Because both the system model and the objective function are linear quadratic, the controller uses a precompiled real-time fast quadratic programming (RT-FQP) solver. RT-FQP uses the previous optimal solution as a warm start in each cycle, requiring only 3-5 iterations to satisfy the KKT condition. For 10×2 control variables, the fixed-point arithmetic operation count is less than 400 MAC operations, resulting in a measured solution time of 0.55ms, fully integrated into the 10ms control cycle.

[0040] The first-step control quantity is obtained through the solver and superimposed on the current working point to form the target current and target gear. Then, the d-axis current command is generated in combination with the weak magnetic principle, and the voltage command is calculated through the current loop. The target PWM frequency is: ; in, Indicates the target PWM frequency; Indicates taking the minimum value; Indicates the upper limit of PWM frequency of slow time domain output; Indicates the reference switching frequency for vehicle calibration.

[0041] In the fast time-domain control loop, after receiving the d-axis and q-axis current commands, the driver first regulates the inverter through a closed-loop current loop (traditional motor field-oriented control (FOC)) to ensure that the three-phase stator current follows the target value in real time. Simultaneously, the controller uses the known number of motor pole pairs, the excitation flux constant, and the current q-axis current feedback to instantly calculate the motor-side electromagnetic torque, an intermediate quantity, in software according to the commonly used torque conversion rules for permanent magnet synchronous motors. This electromagnetic torque is then converted into wheel-end output torque based on the gear ratio and efficiency of the reducer and recorded as the target wheel-end torque for the current cycle.

[0042] Estimating motor power is more straightforward. The inverter's DC bus port measures the bus voltage and bus current within the same control cycle. Multiplying these two values ​​and applying a short-term moving average yields the electric drive input power. Because the fast time window covers only 200 milliseconds, the motor's mechanical power and DC-side electrical power are approximately conserved over this short timescale. Therefore, the controller uses this power as the target motor power for the cycle.

[0043] After obtaining the target wheel-end torque and target motor power, the controller writes them into the execution buffer together with the current gear position instruction and the PWM target switching frequency. At the end of the control cycle, the hardware layer directly sends them to the inverter and motor drive unit.

[0044] S3: Send the wheel-end target torque and required motor output power to the controller to generate control instructions to achieve coordinated execution of the three-in-one electric drive.

[0045] When the fast time-domain control loop completes its calculations, the central controller generates the target wheel-end torque, target motor power, target current command, target inverter PWM switching frequency, and next gear position command. The three-in-one electric drive system then coordinates execution in the following sequence at the end of the same cycle.

[0046] To ensure the internal thermal balance of the three-in-one system, temperature samples of the motor, inverter, and reducer are sent back to the central controller at a 20ms cycle during the execution phase. If the temperature of any node shows an over-threshold trend, the controller immediately calls the soft constraint upper limit table to further reduce the current upper limit, torque upper limit, and PWM frequency of the next cycle, and increase the pump speed or fan speed on the radiator side to suppress thermal shock before it occurs.

[0047] All execution results—including actual current, effective gear position, measured inverter switching frequency, and the latest temperatures of the three nodes—are written to a shared memory area in the form of event frames, providing realistic initial values ​​for the next round of fast time-domain prediction. They are also fed into the slow time-domain thermal power estimation module, ensuring full consistency between the thermal model and physical execution. This repetitive cycle allows the motor, electronic control, and reducer to maintain synchronous coordination within millisecond timescales, meeting the driver's immediate power needs while maintaining the system operating point within safe and efficient temperature zones, achieving the overall real-time, adaptive, and integrated electro-thermal-mechanical control objectives of the present invention.

[0048] Example 2: In an exemplary embodiment, a control system for a three-in-one electric drive system of an electric vehicle is also provided, including an identification module for acquiring vehicle data and identifying driving conditions based on the vehicle data.

[0049] The calculation module triggers the control prediction model according to the driving conditions, builds a multi-objective cost function based on vehicle data, and calculates the wheel-end target torque and required motor output power.

[0050] The control module sends the wheel-end target torque and required motor output power to the controller, generates control instructions, and realizes the coordinated execution of the three-in-one electric drive.

[0051] If the above functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0052] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0053] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, and then editing, interpreting, or processing in another suitable manner as necessary, and then storing it in a computer memory.

[0054] It should be understood that various components of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof may be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0055] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A control method for a three-in-one electric drive system of an electric vehicle, characterized in that: include: Acquire vehicle data and identify driving conditions based on the vehicle data; The control prediction model is triggered according to the driving conditions, and a multi-objective cost function is constructed based on vehicle data to calculate the target wheel-end torque and required motor output power. The wheel-end target torque and required motor output power are sent to the controller to generate control instructions, realizing the coordinated execution of the three-in-one electric drive.

2. The method for controlling a three-in-one electric drive system of an electric vehicle according to claim 1, wherein: The vehicle data includes power pedal, brake master cylinder, steering wheel, vehicle body inertia, wheel speed, road longitudinal slope, battery state of charge and ambient temperature data.

3. The control method of a three-in-one electric drive system for an electric vehicle according to claim 2, characterized in that: The identification of driving conditions includes stacking vehicle data in time sequence to form a three-dimensional tensor ,Will Input three layers of one-dimensional convolution for feature extraction, and input the convolution output into a single-layer gated recurrent unit network. After softmax transformation, the driving condition probability distribution is obtained, and the driving condition and confidence level corresponding to the maximum component are output. The driving condition probability distribution corresponds to six types of driving conditions, including urban constant speed, start-stop following, strong acceleration, high-speed cruising, uphill and downhill.

4. The control method of a three-in-one electric drive system for an electric vehicle according to claim 3, characterized in that: Triggering the control prediction model according to the driving condition includes taking the driving condition, the confidence level, and the vehicle data as inputs to the control prediction model; The control prediction model is a dual-time domain prediction model, divided into a fast time domain and a slow time domain. The slow time domain takes precedence. At the beginning of the current control cycle, the slow time domain uses a four-node RC thermal network to perform a rolling prediction of the temperature evolution in the next 1 second. The highest predicted temperature and the limit temperature are extracted to calculate the temperature rise margin factor. The energy consumption weight and temperature control weight are dynamically adjusted according to the temperature rise margin factor. After obtaining the temperature rise margin factor, the fast time domain maps the temperature rise margin factor into wheel-end torque, current, motor speed, and PWM frequency. Simultaneously, the optimal solution is obtained for pedal and brake transients, and control instructions are issued. The wheel-end torque, current, motor speed, and PWM frequency actually issued in the fast time domain are immediately fed back to the next prediction in the slow time domain, achieving information synchronization between the two domains.

5. The control method of a three-in-one electric drive system for an electric vehicle according to claim 4, characterized in that: The slow time domain includes abstracting the main heat sources and heat dissipation paths of the three-in-one electric drive into four equivalent thermal nodes, namely the motor stator, inverter module, reducer gearbox and radiator coupled with the environment. Based on vehicle data, the motor stator temperature, inverter module temperature, reducer gear oil temperature and radiator outlet temperature are read to form a four-node temperature vector ; Calculate the heating power of the four nodes based on the control instructions issued in the previous fast domain cycle: ； in, represents the column vector of the heating power of the four nodes at time k; Indicates the motor stator copper loss; represents the inverter switching loss; Indicates the mechanical friction and oil shear loss of the reducer; Represents matrix transpose; Considering each node as an isothermal block with heat capacity, heat transfer between nodes is only through heat conduction or forced convection, and establishing a resistance and capacitance equivalent network for the four nodes: ； in, represents the heat capacity of node i; represents the instantaneous temperature of node i; express About time The derivative of Represents the node index; represents the instantaneous temperature of node j; represents the equivalent thermal resistance between node i and node j; represents the self-heating power of node i; Using forward Euler discretization with a fixed step size of Δt=20 ms, at time Discretization yields: ； in, Represents the temperature vector of the four nodes predicted to the next moment; The column vector representing the four-node temperature at time k; represents the discrete sampling step; represents the heat capacity diagonal matrix; represents the four-node overall admittance matrix; represents the environmental coupling column vector; Indicates the ambient temperature; represents the kth discrete moment; Take the maximum predicted value among the four node temperatures , calculate the temperature rise margin factor: ； in, Indicates the temperature rise margin factor; Indicates the set limit temperature; represents the maximum predicted value among the four node temperatures; The temperature rise margin factor is used to adjust the weight of the cost function in the multi-objective cost function in real time, and the upper limits of the wheel end torque, current and PWM frequency are proportionally reduced to establish a soft constraint upper bound table; A multi-objective cost function is constructed using weighted quadratic performance indicators to optimize and output the temperature rise margin factor, cost function weight, and soft constraint upper bound table.

6. The control method of a three-in-one electric drive system for an electric vehicle according to claim 5, characterized in that: The fast time domain includes obtaining the time according to the vehicle data The vehicle linear speed, motor mechanical speed and wheel end output torque executed in the previous control cycle are used to establish a small signal linear electromechanical coupling model; Based on the temperature rise margin factor output in the slow time domain and the soft constraint upper bound table, a quadratic cost function is constructed. The weight in the quadratic cost function is the weight of the cost function output in the slow time domain, and the first-step control quantity is obtained through the solver. The first-step control quantity is added to the current working point to form the target current and target gear. Then, the d-axis current command is generated in combination with the field weakening principle. The voltage command is calculated through the current loop. The target PWM frequency is: ； in, Indicates the target PWM frequency; Indicates taking the minimum value; Indicates the upper limit of PWM frequency of slow time domain output; Indicates the reference switching frequency of the vehicle calibration; Calculate the target wheel-end torque and required motor output power and send them for execution.

7. A control system for a three-in-one electric drive system of an electric vehicle, applied to a control method for a three-in-one electric drive system of an electric vehicle according to any one of claims 1 to 6, characterized in that: include, Identification module, which obtains vehicle data and identifies driving conditions based on the vehicle data; The calculation module triggers the control prediction model according to the driving conditions, builds a multi-objective cost function based on vehicle data, and calculates the target wheel-end torque and required motor output power; The control module sends the wheel-end target torque and required motor output power to the controller, generates control instructions, and realizes the coordinated execution of the three-in-one electric drive.

Images