Thermoelectricity cooperative motor control method and device and vehicle

Abstract

This application relates to a thermoelectric co-control method, device, and vehicle for motors, addressing the problem that thermal models cannot be thermoelectrically coordinated with motor control. The method includes: determining motor copper losses, motor iron losses, controller conduction losses, and controller switching losses based on motor state parameters, motor controller state parameters, and coolant state parameters; constructing a lumped-parameter thermal network model including motor stator nodes, motor controller nodes, and coolant nodes; and solving an optimization problem based on the lumped-parameter thermal network model and losses, using the motor's d-axis and q-axis currents as optimization variables and preset constraints, to obtain the optimal current command for the current moment, which the motor then executes. By actively adjusting the ratio of d-axis and q-axis currents, the distribution of motor copper losses and controller losses is preemptively adjusted before the temperature exceeds limits, achieving coordinated optimization of system thermal state and electrical control while satisfying temperature, current, and voltage constraints.

Description

Technical Field

[0001] This application relates to the field of motor control, specifically to a thermoelectric motor control method, device, and vehicle. Background Technology

[0002] Electric vehicle drive systems are evolving towards higher power density, higher efficiency, and higher frequency. While the application of silicon carbide power devices improves efficiency, their high switching frequency and steep voltage change rate also make the system's thermal management issues more prominent and complex.

[0003] In existing technologies, thermal management of electric drive systems typically employs a passive protection strategy based on temperature thresholds. This involves monitoring the temperature at key points using temperature sensors and implementing power reduction or current limiting measures when the temperature exceeds a safe threshold. The drawback of this strategy is that existing thermal models cannot feed temperature prediction information back into the optimization process of current control commands, resulting in motor control and thermal management operating independently. When the system thermal load is high, the controller can only passively limit power output and cannot actively adjust the ratio of the motor's d-axis and q-axis currents to regulate the distribution of copper losses in the motor and controller, thereby failing to achieve active transfer of heat load between the motor and controller. This makes it difficult for the electric drive system to proactively optimize its thermal state while ensuring power performance. Summary of the Invention

[0004] This application provides a thermoelectric coordinated motor control method, device, and vehicle to solve the problem that thermal models cannot be thermoelectric coordinated with motor control in the prior art.

[0005] The technical solution of this application is as follows:

[0006] This application provides a thermoelectric co-controlled motor control method, including:

[0007] Collect motor status parameters, motor controller status parameters, and coolant status parameters;

[0008] Based on the motor status parameters, the motor controller status parameters, and the coolant status parameters, determine the motor copper loss, motor iron loss, controller conduction loss, and controller switching loss.

[0009] Based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters, and the coolant state parameters, a lumped parameter thermal network model is constructed, which includes motor stator nodes, motor controller nodes, and coolant nodes.

[0010] Based on the lumped parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss, the optimization problem is solved with the motor's d-axis current and q-axis current as optimization variables and preset constraints to obtain the optimal current command at the current moment.

[0011] The motor is controlled to execute the optimal current command.

[0012] By collecting motor status parameters, motor controller status parameters, and coolant status parameters, and based on these parameters determining motor copper losses, motor iron losses, controller conduction losses, and controller switching losses, the heat sources of each component of the electric drive system can be accurately identified, providing an accurate data foundation for subsequent thermal management.

[0013] Using the motor's d-axis and q-axis currents as optimization variables, and solving the optimization problem under preset constraints, the optimal current command is obtained, deeply integrating thermal management with motor control. Through a lumped-parameter thermal network model predictive control framework, the control command can be proactively adjusted while meeting temperature, current, and voltage constraints, avoiding the power performance loss caused by traditional passive power reduction strategies and achieving a coordinated balance between thermal safety and power output.

[0014] This method directly feeds back the temperature prediction information from the thermal model to the optimization process of the current control command, transforming the thermal management strategy from traditional passive power reduction protection to active current regulation based on real-time status monitoring. By actively adjusting the ratio of d-axis and q-axis currents, the distribution of motor copper losses and controller losses can be adjusted in advance before the temperature exceeds the limit. Under the premise of meeting temperature, current, and voltage constraints, the system's thermal state and electrical control are synergistically optimized, effectively avoiding power limiting protection caused by unilateral overheating, improving the continuous high-power output capability and overall reliability of the electric drive system, and solving the problem that the thermal model cannot achieve thermo-electric coordination with motor control in existing technologies.

[0015] In some possible embodiments, after the step of constructing a lumped-parameter thermal network model including motor stator nodes, motor controller nodes, and coolant nodes, and before the step of solving an optimization problem based on the lumped-parameter thermal network model, the motor copper losses, the motor iron losses, the controller conduction losses, and the controller switching losses, using the motor's d-axis current and q-axis current as optimization variables, and under preset constraints to obtain the optimal current command at the current moment, the method further includes:

[0016] Identify the thermal resistance parameters in the lumped parameter thermal network model online.

[0017] By identifying thermal resistance parameters online, the thermal model can adaptively track long-term changes such as cooling system performance degradation, coolant aging, and power module fatigue. Compared to thermal models with fixed parameters, the thermal model in this application can maintain the accuracy of temperature prediction throughout its entire life cycle, avoiding control deviations caused by parameter drift.

[0018] In some possible embodiments, after the motor executes the optimal current command, the method further includes:

[0019] When the temperature difference between the motor controller and the motor exceeds a threshold, the objective function of the optimization problem is modified to guide the heat load to transfer between the motor and the motor controller.

[0020] By monitoring the temperature difference between the motor controller and the motor, the objective function of the optimization problem is adjusted when the temperature difference exceeds a threshold, guiding the heat load to transfer between the motor and the controller. When the controller temperature is too high, the current command is adjusted to reduce controller losses and increase motor losses, causing heat to transfer to the motor side; conversely, when the motor temperature is too high, the opposite is true. This active heat load transfer mechanism can automatically balance the temperature distribution when the system's heat load is unbalanced, avoiding unilateral overheating that triggers power limiting protection, thereby improving the continuous high-power output capability and overall reliability of the electric drive system.

[0021] In some possible embodiments, the motor status parameters include three-phase current, motor speed, and motor winding temperature; the motor controller status parameters include DC bus voltage and controller heatsink temperature; the steps for determining motor copper losses, motor iron losses, controller conduction losses, and controller switching losses include:

[0022] Calculate the motor copper loss based on the three-phase current and the motor winding temperature;

[0023] Calculate the motor iron loss based on the motor speed;

[0024] Calculate the controller conduction loss based on the three-phase current and the controller heat sink temperature;

[0025] Calculate the controller switching losses based on the three-phase currents and the DC bus voltage;

[0026] The sum of the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss is taken as the total loss of the electric drive system.

[0027] By defining specific motor state parameters, including three-phase current, motor speed, and motor winding temperature, and specific motor controller state parameters, including DC bus voltage and controller heatsink temperature, this application clarifies the key input parameters required for thermal management, providing a complete data foundation for subsequent loss calculation and thermal model construction.

[0028] In the calculation of copper loss in motors, the effective value of stator phase current is calculated based on the three-phase current, and the stator resistance is corrected for temperature based on the motor winding temperature. This allows the copper loss calculation to reflect the impact of winding temperature changes on resistance in real time, thus improving the accuracy of loss calculation.

[0029] In the calculation of iron loss in motors, the stator magnetic field frequency is calculated based on the motor speed, and the iron loss is estimated by combining the magnetic flux density amplitude, so that the iron loss calculation can accurately reflect the impact of changes in motor operating conditions on iron core loss.

[0030] In the calculation of the controller conduction loss, the effective value of the phase current is calculated based on the three-phase current, and the conduction voltage drop and conduction resistance of the power device are corrected for temperature based on the controller heat sink temperature, so that the conduction loss calculation can reflect the influence of junction temperature change on the conduction characteristics of the device in real time.

[0031] In the calculation of controller switching losses, the peak phase current is calculated based on the three-phase current, and the switching energy is corrected for voltage and current based on the DC bus voltage, so that the switching loss calculation can accurately reflect the impact of bus voltage fluctuations and load current changes on switching losses.

[0032] By summing the motor copper loss, motor iron loss, controller conduction loss, and controller switching loss as the total loss of the electric drive system, this application establishes a complete system heat source model, providing accurate heat input for the subsequent construction of the thermal network model and online identification of thermal resistance parameters, thereby improving the adaptive accuracy of the thermal model and the reliability of temperature prediction.

[0033] In some possible embodiments, the coolant state parameters include coolant inlet temperature, coolant outlet temperature, and coolant flow rate;

[0034] Based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters, and the coolant state parameters, the steps for constructing a lumped parameter thermal network model including motor stator nodes, motor controller nodes, and coolant nodes include:

[0035] Based on the heat source, heat capacity, and thermal resistance relationship with the coolant node of the motor stator node, the first thermal balance equation of the motor stator node is constructed; wherein, the heat source of the motor stator node consists of the motor copper loss and the motor iron loss;

[0036] Based on the heat source, heat capacity, and thermal resistance relationship between the controller node and the coolant node, a second thermal balance equation for the controller node is constructed; wherein, the heat source of the controller node consists of the controller conduction loss and the controller switching loss;

[0037] Based on the thermal resistance relationship between the motor stator node, controller node and coolant node, as well as the heat capacity of the coolant node, the coolant flow rate, the coolant inlet temperature and the coolant outlet temperature, a third heat balance equation for the coolant node is constructed.

[0038] The first heat balance equation, the second heat balance equation, and the third heat balance equation together constitute the heat balance equation of the lumped parameter heat network model.

[0039] A complete three-node lumped-parameter thermal network model was established by constructing thermal balance equations for the motor stator node, controller node, and coolant node respectively. The first thermal balance equation describes the heat generation and transfer patterns of the motor stator node, based on the motor's copper and iron losses constituting the heat source, and considering the heat capacity of the stator node and its thermal resistance to the coolant node. The second thermal balance equation describes the heat generation and transfer patterns of the controller node, based on the controller's conduction and switching losses constituting the heat source, and considering the heat capacity of the controller node and its thermal resistance to the coolant node. The third thermal balance equation describes the heat accumulation and dissipation patterns of the coolant node, based on the thermal resistance relationships between the motor stator node, controller node, and coolant node, and considering the coolant node's heat capacity, coolant flow rate, coolant inlet temperature, and coolant outlet temperature. The three thermal balance equations are coupled to each other through the temperature difference and thermal resistance between nodes. The motor stator node and the controller node are directly coupled to the coolant node, while the motor stator node and the controller node are indirectly coupled through the coolant node. This fully reflects the thermal dynamic characteristics of the electric drive system and provides an accurate mathematical model basis for the subsequent online identification of thermal resistance parameters and temperature prediction.

[0040] In some possible embodiments, the step of identifying the thermal resistance parameters in the lumped-parameter thermal network model online includes:

[0041] Discretize the heat balance equation of the lumped parameter heat network model to construct a linear regression model;

[0042] The linear regression model is identified online using a recursive least squares algorithm with a forgetting factor to obtain the convective heat transfer resistance between the motor stator node and the coolant node, as well as the convective heat transfer resistance between the controller node and the coolant node.

[0043] By defining coolant state parameters including inlet temperature, outlet temperature, and flow rate, and establishing a lumped-parameter thermal network model with heat balance equations based on motor copper losses, iron losses, controller conduction losses, switching losses, motor winding temperatures, controller radiator temperatures, and coolant parameters, a complete description of the heat transfer relationships between the motor stator node, controller node, and coolant node is achieved. A linear regression model is constructed by discretizing the heat balance equations, and a recursive least squares algorithm with a forgetting factor is used to identify the convective heat transfer resistance between the motor stator node and the coolant node, as well as between the controller node and the coolant node, online. This enables the thermal model to adaptively track changes in thermal characteristics caused by cooling system performance degradation, coolant aging, and power module fatigue, solving the problem of accumulated temperature prediction errors after long-term use of fixed-parameter thermal models and improving the adaptive capability and long-term reliability of thermal management.

[0044] Preferably, the step of using a recursive least squares algorithm with a forgetting factor to perform online identification of the linear regression model to obtain the convective heat transfer resistance between the motor stator node and the coolant node, and the convective heat transfer resistance between the controller node and the coolant node, includes:

[0045] The initial values ​​of the parameters and the initial values ​​of the covariance matrix are set. The initial values ​​of the parameters include the initial values ​​of the convective heat transfer resistance between the motor stator node and the coolant node and the initial values ​​of the convective heat transfer resistance between the controller node and the coolant node. The initial value of the covariance matrix is ​​determined by the inverse of the cumulative information of the regression matrix and is used to characterize the uncertainty of the initial values ​​of the parameters.

[0046] In each control cycle, a regression matrix is ​​constructed based on the current total loss of the electric drive system, coolant inlet temperature, coolant flow rate, and the previous motor winding temperature, controller radiator temperature, and coolant outlet temperature; an observation vector is constructed based on the current motor winding temperature, controller radiator temperature, and coolant outlet temperature.

[0047] The gain matrix for the current time step is calculated based on the covariance matrix of the previous time step, the regression matrix of the current time step, and the forgetting factor. The covariance matrix and the regression matrix together determine the size of the gain matrix, which in turn determines the degree of influence of the measurement data at the current time step on the parameter update.

[0048] Update the parameter estimates at the current time step based on the parameter estimates from the previous time step, the gain matrix at the current time step, the observation vector at the current time step, and the regression matrix at the current time step.

[0049] Update the covariance matrix at the current time step based on the covariance matrix at the previous time step, the gain matrix at the current time step, the regression matrix at the current time step, and the forgetting factor.

[0050] Based on the updated parameter estimates, the convective heat transfer resistance between the motor stator node and the coolant node, as well as the convective heat transfer resistance between the controller node and the coolant node, are calculated under the current operating conditions.

[0051] By employing a recursive least squares algorithm with a forgetting factor to identify the linear regression model online, adaptive updates of convective heat transfer resistance are achieved.

[0052] First, by setting initial values ​​for the parameters and the covariance matrix, where the initial values ​​for the parameters are obtained through offline calibration under rated operating conditions and the initial value for the covariance matrix is ​​determined by the inverse of the cumulative information content of the regression matrix, a reasonable starting point is provided for parameter identification. At the same time, the magnitude of the initial value of the covariance matrix characterizes the uncertainty of the initial values ​​for the parameters, enabling the algorithm to make full use of new data for parameter correction in the initial stage.

[0053] In each control cycle, a regression matrix is ​​constructed based on the total loss of the electric drive system, the coolant inlet temperature, the coolant flow rate at the current moment, and the temperature data at the previous moment. An observation vector is constructed based on the temperature data at the current moment. This ensures that the regression matrix contains the heat source input and historical state information at the current moment, and the observation vector contains the measured temperature information at the current moment, providing a complete data foundation for parameter identification.

[0054] The gain matrix is ​​calculated based on the covariance matrix of the previous time step, the regression matrix of the current time step, and the forgetting factor. The size of the gain matrix is ​​determined by both parameter uncertainty and the amount of information in the input data. When parameter uncertainty is high or the input data is rich in information, the gain matrix increases accordingly, allowing new data to contribute more to parameter correction; conversely, it decreases, making parameter updates more stable. The forgetting factor controls the weight of historical data, enabling the algorithm to balance the speed of parameter tracking with stability against noise interference.

[0055] The parameter estimates for the current time step are updated based on the parameter estimates, gain matrix, observation vector, and regression matrix from the previous time step. The error between the observation vector and the predicted value is used to gradually correct the parameters, making the model's predicted values ​​continuously approach the measured values. At the same time, the covariance matrix is ​​updated to characterize the dynamic changes in the uncertainty of the parameter estimates. As data accumulates, the covariance matrix gradually decreases, indicating that the parameter estimates are becoming more and more accurate. The introduction of the forgetting factor reduces the weight of historical data, preventing the covariance matrix from being too small and thus causing a slow response to changes in operating conditions.

[0056] Finally, the convective heat transfer thermal resistance under the current operating conditions is calculated based on the updated parameter estimates, enabling the thermal model to reflect the changes in thermal characteristics caused by factors such as cooling system performance degradation, coolant aging, and power module fatigue in real time. This solves the problem of the cumulative increase in temperature prediction error after long-term use of fixed parameter thermal models, and improves the adaptive capability and long-term reliability of thermal management.

[0057] In some possible embodiments, the steps of solving the optimization problem based on the lumped parameter thermal network model and the total loss, using the d-axis current and q-axis current of the motor as optimization variables and with preset constraints, to obtain the optimal current command at the current moment include:

[0058] Set the prediction time domain, and use the d-axis current sequence and q-axis current sequence in the future prediction time domain as optimization variables;

[0059] Set an optimization objective function, wherein the optimization objective function is at least one of minimizing total system losses, minimizing temperature deviation from the reference value, or maximizing output torque;

[0060] Set constraints, which include temperature constraints, current hardware limit constraints, and voltage modulation limit constraints predicted based on the lumped parameter thermal network model.

[0061] An optimization problem is constructed using the optimization variables, the optimization objective function, and the constraints. The optimization problem is solved in each control cycle to obtain the optimal d-axis current command and the optimal q-axis current command at the current moment.

[0062] By applying model predictive control to the thermo-electric co-optimization of the electric drive system, forward-looking optimal control of the motor current command is achieved.

[0063] By setting a prediction time domain and using the d-axis current sequence and q-axis current sequence in the future prediction time domain as optimization variables, the controller can plan the current trajectory in multiple future control cycles, predict the system temperature change trend in advance, and avoid the short-sighted behavior of traditional control strategies that only make decisions at the current moment.

[0064] By setting the optimization objective function to minimize total system losses, minimize temperature deviation from the reference value, or maximize output torque, the control strategy can flexibly adjust the optimization direction according to different operating conditions. Under normal operating conditions, minimizing total system losses can be selected to improve energy efficiency; when the heat load is high, minimizing temperature deviation from the reference value can be selected to ensure system safety; and when power demand is urgent, maximizing output torque can be selected to fully utilize performance, thus achieving multi-objective coordinated optimization of energy efficiency, thermal safety, and power performance.

[0065] By setting constraints, including temperature constraints predicted by a lumped-parameter thermal network model, current hardware limits, and voltage modulation limits, the optimization problem can fully consider the physical limitations of the system during the solution process. Temperature constraints utilize the thermal network model to predict future temperatures, ensuring that the junction temperatures of power devices and permanent magnets remain within safe ranges. Current constraints guarantee that the effective values ​​of d-axis current, q-axis current, and phase current do not exceed the hardware's carrying capacity limits. Voltage constraints ensure that the d-axis voltage and q-axis voltage do not exceed the modulation limits of the DC bus voltage, making the optimal current command obtained from the solution engineering feasible.

[0066] An optimization problem is constructed using optimization variables, an optimization objective function, and constraints. This problem is solved in each control cycle to obtain the optimal d-axis current command and the optimal q-axis current command at the current moment, thus achieving rolling optimization control. Compared with traditional control strategies, this application, through look-ahead prediction, multi-objective optimization, and constraint processing, can proactively adjust the current command to balance system losses and temperature distribution while meeting system safety constraints. This avoids the loss of dynamic performance caused by passive power reduction and achieves optimal control for thermo-electric coordination.

[0067] In some possible embodiments, the optimization problem is as follows:

[0068] The optimization variables are the d-axis current and q-axis current at each time point in the future prediction time domain;

[0069] The optimization objective function is a weighted sum of at least one of minimizing total system losses, minimizing temperature deviation from the reference value, or maximizing output torque;

[0070] The constraints include: temperature constraints: the motor stator node temperature, controller node temperature, and power device junction temperature calculated from the controller node temperature, as predicted by the lumped parameter thermal network model, shall not exceed their respective safety thresholds.

[0071] Current constraints: The effective values ​​of the d-axis current, q-axis current, and the phase current calculated from the d-axis current and q-axis current shall not exceed their respective hardware limits;

[0072] Voltage constraint: The d-axis voltage and q-axis voltage calculated from the d-axis current and q-axis current shall not exceed the modulation limit of the DC bus voltage.

[0073] By using the predicted d-axis and q-axis current sequences in the future time domain as optimization variables, and employing a weighted sum of at least one of minimizing total system losses, minimizing temperature deviation from the reference value, or maximizing output torque as the objective function, a complete model predictive control optimization framework is constructed by setting multiple constraints, including temperature, current, and voltage constraints. Specifically, the temperature constraint ensures that the predicted stator node temperature of the motor, the node temperature of the controller, and the junction temperature of the power devices, based on the lumped-parameter thermal network model, do not exceed their respective safety thresholds, enabling the controller to proactively prevent temperature over-limits. The current constraint ensures that the effective values ​​of the d-axis current, q-axis current, and phase current do not exceed hardware limits, guaranteeing electrical safety. The voltage constraint ensures that the d-axis voltage and q-axis voltage do not exceed the modulation limit of the DC bus voltage, ensuring the feasibility of inverter modulation. Through online solving of this multi-constraint optimization problem, this application can proactively adjust the current command to achieve the optimal balance between system losses, temperature distribution, and output torque while satisfying system safety constraints, avoiding the power performance loss caused by traditional passive power reduction strategies.

[0074] In some possible embodiments, when the temperature difference between the motor controller and the motor exceeds a threshold, the step of modifying the objective function of the optimization problem to guide the transfer of heat load between the motor and the motor controller includes:

[0075] Obtain the temperature difference between the motor stator node and the controller node;

[0076] When the temperature difference exceeds a preset temperature difference threshold, and the current motor speed reaches a preset medium-high speed threshold, and the current motor output torque is lower than a preset medium-low load threshold, a temperature difference penalty term is added to the optimization objective function; the temperature difference penalty term is used to guide the optimization variables to adjust in the direction of reducing the temperature difference.

[0077] Solve the corrected optimization problem to obtain the optimal current command at the current moment.

[0078] By acquiring the temperature difference between the motor stator node and the controller node, when the temperature difference exceeds a preset threshold and simultaneously meets the conditions of medium-high speed and medium-low load, a temperature difference penalty term is added to the optimization objective function. This guides the optimization variables to adjust in the direction of reducing the temperature difference, and the optimal current command is obtained by solving the corrected optimization problem. This mechanism is triggered only when the system has adjustment space and does not affect power output, avoiding sacrificing driving performance due to heat load transfer in scenarios with urgent power demands such as rapid acceleration. When the controller temperature is higher than the motor temperature, the penalty term guides the optimization variables to reduce controller losses and increase motor losses, causing heat to transfer from the controller side to the motor side; conversely, it guides heat to transfer from the motor side to the controller side. By actively adjusting the heat load distribution, this application can automatically balance the system temperature field when the heat load is unbalanced, avoiding power limiting protection triggered by unilateral overheating, improving the continuous high power output capability and long-term operational reliability of the electric drive system, and automatically exiting the transfer mode after the temperature difference recovers, achieving dynamic balance control of the heat load.

[0079] This application also provides a thermoelectric co-operated motor control device, including:

[0080] The data acquisition module is used to acquire motor status parameters, motor controller status parameters, and coolant status parameters.

[0081] The loss determination module is used to determine the motor copper loss, motor iron loss, controller conduction loss and controller switching loss based on the motor state parameters, the motor controller state parameters and the coolant state parameters.

[0082] The thermal resistance parameter identification module is used to construct a lumped parameter thermal network model containing motor stator nodes, motor controller nodes, and coolant nodes based on the motor copper loss, motor iron loss, controller conduction loss, controller switching loss, motor state parameters, motor controller state parameters, and coolant state parameters.

[0083] The current control module is used to solve the optimization problem based on the lumped parameter thermal network model and the total loss, with the d-axis current and q-axis current of the motor as optimization variables and preset constraints, to obtain the optimal current command at the current moment.

[0084] The execution module is used to control the motor to execute according to the optimal current command.

[0085] This application also provides a vehicle including the aforementioned thermoelectric co-operated motor control device. Attached Figure Description

[0086] Figure 1 This is a structural block diagram of the vehicle according to the first embodiment of this application;

[0087] Figure 2This is a flowchart illustrating the thermoelectric coordinated motor control method in the second embodiment of this application;

[0088] Figure 3 This is a flowchart illustrating the thermoelectric coordinated motor control method in the third embodiment of this application;

[0089] Figure 4 This is a flowchart illustrating the thermoelectric coordinated motor control method in the fourth embodiment of this application;

[0090] Figure 5 This is a structural block diagram of the thermoelectric coordinated motor control device in the fifth embodiment of this application. Detailed Implementation

[0091] The first embodiment of this application provides a vehicle 100, which can be, but is not limited to, a pure electric vehicle (PEV / BEV), a hybrid electric vehicle (HEV), a range-extended electric vehicle (REEV), a plug-in hybrid electric vehicle (PHEV), a new energy vehicle, or a fuel vehicle. (See also...) Figure 1 The vehicle 100 includes: a motor 101, a motor controller 102, a cooling system 103, and a vehicle controller 104.

[0092] Motor 101 is a permanent magnet synchronous motor used to provide driving torque.

[0093] The motor controller 102 includes a three-phase inverter, which uses silicon carbide power devices to convert the DC bus voltage into a three-phase AC voltage to drive the motor.

[0094] The cooling system 103 includes, for example, a cooling pump, a radiator, a cooling fan, and a coolant circulation pipeline, for dissipating heat from the motor and the controller.

[0095] The vehicle controller 104 is communicatively connected to the motor controller 102, each sensor, and the cooling pump control unit of the cooling system 103, and is used to execute the thermoelectric coordinated motor control method in the second to fourth embodiments of this application.

[0096] After the vehicle 100 is started, the vehicle controller 104 collects the motor status parameters, motor controller status parameters and coolant status parameters in real time through various sensors.

[0097] The motor status parameters include three-phase current, motor speed, and motor winding temperature. The three-phase current is obtained through a Hall current sensor installed at the inverter output, the motor speed is obtained through a rotary transformer installed at the motor rotor shaft, and the motor winding temperature is obtained through a platinum resistance temperature sensor embedded at the stator winding end. The motor controller status parameters include DC bus voltage and controller heatsink temperature. The DC bus voltage is obtained through a resistor divider connected in parallel across the DC bus capacitor and an isolation amplifier, and the controller heatsink temperature is obtained through a thermocouple temperature sensor attached to the power module heatsink substrate. The coolant status parameters include coolant inlet temperature, coolant outlet temperature, and coolant flow rate. The coolant inlet temperature and coolant outlet temperature are obtained through temperature sensors installed at the inlet and outlet of the coolant circulation pipe, respectively, and the coolant flow rate is calculated from the coolant pump speed.

[0098] Based on the collected parameters, the vehicle controller 104 performs steps such as loss calculation, thermal model construction and identification, optimization problem solving, and current command output. It then sends the calculated optimal d-axis and q-axis current commands to the motor controller 102, where the current closed-loop control unit converts them into three-phase voltage commands. These commands are then used by the space vector pulse width modulation module to generate switching signals to drive the inverter, achieving precise control of the motor 101. Simultaneously, the vehicle controller 104 can send speed commands to the cooling pump control unit of the cooling system 103 as needed to adjust the coolant flow rate, collaboratively achieving thermal-electric optimization management of the electric drive system.

[0099] Reference Figure 2 The second embodiment of this application, based on the aforementioned vehicle, further provides a thermoelectric co-operational motor control method, including:

[0100] S101, collects motor status parameters, motor controller status parameters and coolant status parameters;

[0101] S102, based on the motor state parameters, the motor controller state parameters and the coolant state parameters, determine the motor copper loss, motor iron loss, controller conduction loss and controller switching loss;

[0102] S103, based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters and the coolant state parameters, construct a lumped parameter thermal network model including motor stator nodes, motor controller nodes and coolant nodes;

[0103] S104, based on the lumped parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss and the controller switching loss, the d-axis current and q-axis current of motor 101 are used as optimization variables, and the optimization problem is solved under preset constraints to obtain the optimal current command at the current moment.

[0104] S105, control motor 101 to execute according to the optimal current command.

[0105] In step S101, in a specific implementation, the motor state parameters include three-phase current, motor speed, and motor winding temperature. The unit of three-phase current is A, denoted as I. A I B I C The instantaneous current values ​​are collected in real time by current sensors installed at the inverter output, typically using Hall effect current sensors or shunt resistors, corresponding to the instantaneous current values ​​of phases A, B, and C, respectively. Motor speed, measured in rad / s and denoted as ω, is obtained through a rotary transformer or photoelectric encoder installed at the motor rotor shaft end, outputting the rotor's mechanical angular velocity. Motor winding temperature, measured in °C and denoted as T, is measured in °C. w The temperature is obtained by temperature sensors embedded at the ends of the stator windings, typically using platinum resistance temperature sensors or thermocouples, to directly measure the real-time temperature of the windings.

[0106] The motor controller status parameters include the DC bus voltage and the controller heatsink temperature. The unit of DC bus voltage is V, denoted as U. dc The temperature is obtained through a voltage sensor connected in parallel across the DC bus capacitor, typically using a resistor divider with an isolation amplifier. The controller heatsink temperature is measured in °C, denoted as T. s The temperature is obtained by a temperature sensor attached to the heat sink substrate of the power module, and this temperature reflects the heat dissipation boundary conditions of the power device.

[0107] Coolant condition parameters include coolant inlet temperature, coolant outlet temperature, and coolant flow rate. The unit for coolant inlet temperature is °C, denoted as T. cin The temperature is obtained through a temperature sensor installed at the inlet of the coolant circulation line. The coolant outlet temperature is measured in °C and denoted as T. cout The temperature is obtained through a temperature sensor installed at the outlet of the coolant circulation line. The unit of coolant flow rate is m³ / s. 3 / s, denoted as q c The flow rate is calculated by the speed of the cooling pump. Specifically, the coolant flow rate is equal to the coolant flow rate at the rated pump speed multiplied by the ratio of the current pump speed to the rated pump speed. This value reflects the heat dissipation capacity of the cooling system.

[0108] In step S102, the motor copper loss is calculated based on the three-phase current and the motor winding temperature;

[0109] Calculate the motor iron loss based on the motor speed;

[0110] Calculate the controller conduction loss based on the three-phase current and the controller heat sink temperature;

[0111] Calculate the controller switching losses based on the three-phase currents and the DC bus voltage;

[0112] The sum of the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss is taken as the total loss of the electric drive system.

[0113] The steps for calculating motor copper losses include:

[0114] First, based on the three-phase current I A I B I C Calculate the effective value of stator phase current I rms The calculation formula is as follows:

[0115]

[0116] Among them, I A I B I C This represents the instantaneous value of the three-phase current, in amperes (A).

[0117] Secondly, based on the motor winding temperature T w Correcting stator phase resistance R (T) w The calculation formula is as follows:

[0118]

[0119] Among them, R 25 The stator phase resistance is measured in Ω at a reference temperature of 25°C and is specified by the motor manufacturer. α cu T is the temperature coefficient of copper, with a value of 0.00393 / °C; w R(T) represents the motor winding temperature, measured in °C, directly by a temperature sensor. w () represents the temperature-corrected stator phase resistance, in Ω.

[0120] Finally, calculate the motor copper loss P. cu The calculation formula is as follows:

[0121]

[0122] Among them, P cu Copper loss of the motor, in watts (W); Irms This represents the effective value of the stator phase current.

[0123] The steps for calculating the iron loss of the motor based on the motor speed include:

[0124] First, calculate the stator magnetic field frequency f based on the motor speed ω and the number of pole pairs p. The calculation formula is as follows:

[0125]

[0126] Where p is the number of pole pairs of the motor, dimensionless; ω is the mechanical angular velocity of the motor, in rad / s; and f is the stator magnetic field frequency, in Hz.

[0127] Secondly, calculate the motor iron loss P. fe The calculation formula is as follows:

[0128]

[0129] Among them, P fe,rated This is the rated iron loss value under rated operating conditions, in W, calibrated by the motor factory test; f rated B is the rated frequency, in Hz, calculated from the rated speed; B is the stator core magnetic flux density amplitude under current operating conditions, in tons, estimated from the motor voltage equation; B rated The rated magnetic flux density amplitude, in tons (T), is determined by the motor design parameters; P fe This represents the iron loss of the motor, measured in watts (W).

[0130] The steps for calculating the controller conduction loss based on the three-phase current and the controller heat sink temperature include:

[0131] First, based on the controller heatsink temperature T s Correcting the on-state voltage drop V of power devices ce and on-resistance R on The calculation formula is as follows:

[0132]

[0133] Among them, V ce,25 The on-state voltage drop of the power device at a reference temperature of 25°C, in volts (V), is provided in the device datasheet; α v The pressure drop temperature coefficient is taken as -0.001 to -0.003 / °C; T s The controller heatsink temperature, in °C, is measured directly by a temperature sensor; V ce (T s () represents the temperature-corrected on-state voltage drop, in V.

[0134] Secondly, adjust the on-resistance of the power devices based on the controller heatsink temperature:

[0135]

[0136] Among them, R on,25 The on-resistance of the power device at a reference temperature of 25°C, in Ω, is provided in the device datasheet; α r T is the temperature coefficient of resistance, with a value ranging from 0.004 to 0.006 °C; s The controller heatsink temperature, in °C, is measured directly by a temperature sensor; R on (T s () represents the temperature-corrected on-resistance, in Ω.

[0137] Finally, calculate the controller conduction loss:

[0138]

[0139] Among them, I rms P represents the effective value of the stator phase current, in A. cond The controller conduction loss is expressed in watts (W).

[0140] The steps for calculating the controller switching losses based on the three-phase currents and the DC bus voltage include:

[0141] First, calculate the peak value of the stator phase current based on the effective value of the stator phase current:

[0142]

[0143] Among them, I peak This represents the peak value of the stator phase current, expressed in amperes (A).

[0144] Secondly, the energy required for a single activation has been adjusted:

[0145]

[0146] Among them, E on,ref The single-turn-on energy under rated conditions, in J, is provided in the device datasheet; I ref Rated current, in amperes (A), provided in the device datasheet; U dc This is the DC bus voltage, measured in volts (V), and is directly measured by a voltage sensor; U ref The rated voltage is given in volts (V) and is provided in the device datasheet; k1 is the current index, ranging from 1.5 to 2.0; k2 is the voltage index, ranging from 1.0 to 1.2; E on The corrected single-cycle activation energy is expressed in J.

[0147] Secondly, adjust the energy required for a single shutdown:

[0148]

[0149] Among them, E off,ref E represents the single-turn-off energy under rated conditions, in J, provided in the device datasheet. off The corrected single-turn-off energy is expressed in J.

[0150] Finally, the controller switching losses are calculated:

[0151]

[0152] Among them, f sw P is the controller switching frequency in Hz, a preset parameter of the system. sw The switching loss of the controller is expressed in watts (W).

[0153] The sum of the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss is taken as the total loss of the electric drive system:

[0154]

[0155] Among them, P loss This represents the total power loss of the electric drive system, expressed in watts (W).

[0156] Step S103 includes:

[0157] S1031, based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor winding temperature, the controller radiator temperature, the coolant inlet temperature, the coolant outlet temperature, and the coolant flow rate, establish the heat balance equation of the lumped parameter heat network model.

[0158] In step S1031, a first thermal balance equation for the motor stator node is constructed based on the heat source, heat capacity, and thermal resistance relationship with the coolant node of the motor stator node; wherein the heat source of the motor stator node consists of the motor copper loss and the motor iron loss.

[0159] Based on the heat source, heat capacity, and thermal resistance relationship between the controller node and the coolant node, a second thermal balance equation for the controller node is constructed; wherein, the heat source of the controller node consists of the controller conduction loss and the controller switching loss;

[0160] Based on the thermal resistance relationship between the motor stator node, controller node and coolant node, as well as the heat capacity of the coolant node, the coolant flow rate, the coolant inlet temperature and the coolant outlet temperature, a third heat balance equation for the coolant node is constructed.

[0161] The first, second, and third heat balance equations together constitute the heat balance equations of the lumped parameter heat network model.

[0162] The first heat balance equation is expressed as:

[0163]

[0164] Among them, C w The heat capacity of the motor stator joint is expressed in J / °C; T w The temperature of the motor stator node is expressed in °C and is characterized by the temperature of the motor windings; t is time, expressed in seconds; P cu Copper loss of the motor, measured in W; P fe Iron loss of the motor, measured in W; T c R represents the temperature of the coolant junction, in °C. conv,w The convective heat transfer thermal resistance between the motor stator node and the coolant node is expressed in °C / W and is the parameter to be identified.

[0165] The second heat balance equation is expressed as:

[0166]

[0167] Among them, C s The heat capacity of the controller node is expressed in J / °C; T s The temperature of the controller node, in °C, is characterized by the temperature of the controller's heatsink; P cond The controller conduction loss is expressed in W; P sw The switching loss of the controller is expressed in watts (W); R conv,s The convective heat transfer thermal resistance between the controller node and the coolant node is expressed in °C / W, and is the parameter to be identified.

[0168] The third heat balance equation is expressed as:

[0169]

[0170] Where Cc is the heat capacity of the coolant node, in J / °C; ρ is the density of the coolant, in kg / m³. 3 c p q represents the specific heat capacity of the coolant, expressed in J / (kg·°C); c Coolant flow rate, unit: m³ / s 3 / S, determined by the coolant flow rate in the coolant state parameters; T cinThe coolant inlet temperature, in °C, is determined by the coolant inlet temperature in the coolant state parameters; the coolant node temperature Tc is determined by both the coolant inlet temperature and the coolant outlet temperature, specifically Tc = (T... cin +T cout ) / 2, where T cout The coolant outlet temperature is expressed in °C and is determined from the coolant outlet temperature in the coolant state parameters.

[0171] In the second embodiment of this application, step S104 includes:

[0172] S1041, Set the prediction time domain, and use the d-axis current sequence and q-axis current sequence in the future prediction time domain as optimization variables;

[0173] S1042, Set the optimization objective function, wherein the optimization objective function is at least one of minimizing the total system loss, minimizing the temperature deviation from the reference value, or maximizing the output torque;

[0174] S1043, Set constraints, the constraints including temperature constraints, current hardware limit constraints and voltage modulation limit constraints predicted according to the lumped parameter thermal network model;

[0175] S1044, an optimization problem is constructed using the optimization variables, the optimization objective function, and the constraints. The optimization problem is solved in each control cycle to obtain the optimal d-axis current command and the optimal q-axis current command at the current moment.

[0176] Specifically, if the prediction time domain consists of N control cycles, then the optimization variables can be expressed as:

[0177]

[0178] Where id(k+i|k) and iq(k+i|k) are the d-axis current and q-axis current predicted at the current time k for the i-th future time, respectively, in A; N is the prediction time domain length, ranging from 3 to 10 control cycles.

[0179] When the weighted sum of minimizing total system loss and minimizing temperature deviation from the reference value is used as the objective function, the optimization objective function J is expressed as:

[0180]

[0181] Among them, w1w1, w2w2, and w3w3 are weighting coefficients, all of which are non-negative numbers and are dynamically adjusted according to the control priority; P loss(k+i|k) represents the predicted total system loss at the i-th step, in W, calculated from the sum of motor copper loss, motor iron loss, controller conduction loss, and controller switching loss; T j (k+i|k) represents the predicted junction temperature of the power device at the i-th step, in °C, calculated by adding the controller node temperature to the controller loss and multiplying by the junction-to-substrate thermal resistance; Tj ∗ This is a reference value for the junction temperature of power devices, in °C, ranging from 80 to 100°C; T pm (k+i|k) represents the predicted permanent magnet temperature at the i-th step, in °C, calculated by subtracting motor losses from the motor stator node temperature and multiplying by the thermal resistance from the winding to the permanent magnet; Tpm ∗ This is a reference value for the temperature of the permanent magnet, in °C, and the range is 60 to 80°C.

[0182] When maximizing output torque is used as the objective function, the optimization objective function is expressed as:

[0183]

[0184] Among them, T e (k+i|k) represents the predicted electromagnetic torque of the motor at the i-th step, in N / m, calculated from the d-axis and q-axis currents using the following formula:

[0185]

[0186] Where p is the number of pole pairs of the motor, which is dimensionless; ψ pm The flux linkage of a permanent magnet is measured in Wb; L d L q These are the d-axis inductance and q-axis inductance, respectively, in H.

[0187] When a mixed objective function is used, the optimization objective function is expressed as:

[0188]

[0189] Among them, w t This is the torque weighting coefficient, which is dynamically adjusted according to the vehicle's driving conditions.

[0190] Next, execute S1043 to set constraints. The constraints include temperature constraints, current hardware limit constraints, and voltage modulation limit constraints.

[0191] Temperature constraints are expressed as:

[0192]

[0193] Among them, T j,min and T j,maxTpm,min and Tpm,max are the minimum and maximum safe thresholds for the junction temperature of the power device, respectively, in °C, and are determined by the device datasheet; Tpm,min and Tpm,max are the minimum and maximum safe thresholds for the permanent magnet temperature, respectively, in °C, and are determined by the motor design.

[0194] The current hardware limit constraint is expressed as:

[0195]

[0196] Among them, i d,max i represents the maximum current along the d-axis, in A. q,max I is the maximum q-axis current, in A. max The maximum effective value of the stator phase current is expressed in amperes (A). All three values ​​are determined by the power devices of the controller and the load-bearing limits of the motor windings.

[0197] The voltage modulation limit constraint is expressed as:

[0198]

[0199] Among them, u d and u q These are the d-axis voltage and q-axis voltage, respectively, in V, derived from the motor voltage equation; U dc is the DC bus voltage, in volts (V), measured directly by a voltage sensor; S is the modulation ratio margin, ranging from 0.05 to 0.1, used to avoid voltage distortion.

[0200] In execution S1044, the optimization problem is solved using a quadratic programming algorithm or a model predictive control-specific solver. The optimal d-axis current command obtained at the current moment is denoted as id. ∗ (k), the optimal q-axis current command is denoted as iq ∗ (k) is then output to the current controller for execution.

[0201] In optimization problems:

[0202] The optimization variables are the d-axis current and q-axis current at each time point in the future prediction time domain; assuming the prediction time domain is N control cycles, the optimization variables are expressed as:

[0203]

[0204] Among them, i d (k+i|k) and i q (k+i|k) represents the d-axis current and q-axis current predicted at the current time k for the i-th future time, in A; N is the prediction time domain length, ranging from 3 to 10 control cycles.

[0205] The optimization objective function is a weighted sum of at least one of minimizing total system losses, minimizing temperature deviation from the reference value, or maximizing output torque; when using a weighted sum of minimizing total system losses and minimizing temperature deviation from the reference value, the optimization objective function is expressed as:

[0206]

[0207] Where w1, w2, and w3 are weighting coefficients, all of which are non-negative; P loss (k+i|k) represents the predicted total system loss at the i-th step, in W; T j (k+i|k) represents the predicted junction temperature of the power device at the i-th step, in °C; Tj ∗ This is a reference value for the junction temperature of power devices, in °C; T pm (k+i|k) represents the predicted temperature of the permanent magnet at the i-th step, in °C; T pm ∗ This is a reference value for the temperature of the permanent magnet, in °C.

[0208] When maximizing output torque is adopted, the optimization objective function is expressed as:

[0209]

[0210] Among them, T e (k+i|k) represents the predicted electromagnetic torque of the motor at the i-th step in the future, in Newton-meters.

[0211] When a mixed objective function is used, the optimization objective function is expressed as:

[0212]

[0213] Among them, w t This is the torque weighting coefficient.

[0214] The constraints include temperature constraints, current constraints, and voltage constraints.

[0215] The temperature constraint, based on the lumped-parameter thermal network model, stipulates that the motor stator node temperature, controller node temperature, and power device junction temperature calculated from the controller node temperature must not exceed their respective safety thresholds. The temperature constraint is expressed as follows:

[0216]

[0217] Among them, T j,min and T j,maxThese are the minimum and maximum safe threshold temperatures for the junction temperature of the power device, respectively, in °C, and are determined by the device datasheet; T pm,min and T pm,max These are the minimum and maximum safe temperature thresholds for the permanent magnet, respectively, in °C, and are determined by the motor design. The junction temperature of the power device is calculated by adding the controller node temperature to the controller losses and multiplying by the junction-to-substrate thermal resistance, i.e.:

[0218]

[0219] Among them, T s R represents the controller node temperature, in °C. th,jc Thermal resistance from junction to substrate of power device, in °C / W, provided in device datasheet.

[0220] Current constraints: The effective values ​​of the d-axis current, q-axis current, and the phase current calculated from the d-axis current and q-axis current shall not exceed their respective hardware limits; the current constraints are expressed as:

[0221]

[0222] Among them, i d,max i represents the maximum current along the d-axis, in A. q,max I is the maximum q-axis current, in A. max These are the maximum effective values ​​of the stator phase currents, expressed in amperes (A). All three values ​​are determined by the power devices in the controller and the load-bearing limits of the motor windings. The relationship between the effective value of the phase current and the d-axis and q-axis currents is as follows:

[0223]

[0224] Voltage constraint: The d-axis voltage and q-axis voltage calculated from the d-axis current and q-axis current shall not exceed the modulation limit of the DC bus voltage. The voltage constraint is expressed as:

[0225]

[0226] in, u d and u q These are the d-axis voltage and q-axis voltage, respectively, in V, derived from the motor voltage equation:

[0227]

[0228] Among them, R s L is the stator resistance, in Ω; ω is the mechanical angular velocity of the motor, in rad / s; d L q These are the d-axis inductance and q-axis inductance, respectively, in H; ψ pm The flux linkage of a permanent magnet is measured in Wb; U dcis the DC bus voltage, in V, which is directly measured by a voltage sensor; S is the modulation ratio margin, ranging from 0.05 to 0.1.

[0229] An optimization problem is constructed using the aforementioned optimization variables, objective function, and constraints. This problem is solved in each control cycle using a quadratic programming algorithm or a model predictive control solver to obtain the optimal d-axis current command and optimal q-axis current command at the current moment. The optimal d-axis current command obtained is denoted as... The optimal q-axis current command is denoted as And output it to the current controller for execution.

[0230] In the first embodiment described above, this method can directly feed back the temperature prediction information from the thermal model to the optimization process of the current control command, transforming the thermal management strategy from traditional passive power reduction protection to active current regulation based on real-time status monitoring. By actively adjusting the ratio of the d-axis current and the q-axis current, the distribution of motor copper losses and controller losses can be adjusted in advance before the temperature exceeds the limit. Under the premise of meeting temperature constraints, current constraints, and voltage constraints, the system's thermal state and electrical control are synergistically optimized, effectively avoiding power limiting protection caused by unilateral overheating, improving the continuous high-power output capability and overall reliability of the electric drive system, and solving the problem that the thermal model cannot perform thermoelectric coordination with motor control in the prior art.

[0231] Reference Figure 3 In the third embodiment of this application, in order to improve the prediction results of the lumped parameter thermal network model, a step S106 is added between steps S103 and S104, based on the second embodiment described above. Specifically, step S106 involves: identifying the thermal resistance parameters in the lumped parameter thermal network model online.

[0232] In the third embodiment of this application, step S106 specifically includes:

[0233] S1061, Discretize the heat balance equation and construct a linear regression model;

[0234] S1062, the recursive least squares algorithm with forgetting factor is used to identify the linear regression model online, and the convective heat transfer resistance between the motor stator node and the coolant node and the convective heat transfer resistance between the controller node and the coolant node are obtained.

[0235] Step S1062 includes: setting initial values ​​for parameters and initial values ​​for the covariance matrix. The initial values ​​for parameters include the initial values ​​for convective heat transfer resistance between the motor stator node and the coolant node, and the initial values ​​for convective heat transfer resistance between the controller node and the coolant node. The initial values ​​for the covariance matrix are determined by the inverse of the cumulative information of the regression matrix and are used to characterize the uncertainty of the initial values ​​for parameters.

[0236] In each control cycle, a regression matrix is ​​constructed based on the current total loss of the electric drive system, coolant inlet temperature, coolant flow rate, and the previous motor winding temperature, controller radiator temperature, and coolant outlet temperature; an observation vector is constructed based on the current motor winding temperature, controller radiator temperature, and coolant outlet temperature.

[0237] The gain matrix for the current time step is calculated based on the covariance matrix of the previous time step, the regression matrix of the current time step, and the forgetting factor. The covariance matrix and the regression matrix together determine the size of the gain matrix, which in turn determines the degree of influence of the measurement data at the current time step on the parameter update.

[0238] Update the parameter estimates at the current time step based on the parameter estimates from the previous time step, the gain matrix at the current time step, the observation vector at the current time step, and the regression matrix at the current time step.

[0239] Update the covariance matrix at the current time step based on the covariance matrix at the previous time step, the gain matrix at the current time step, the regression matrix at the current time step, and the forgetting factor.

[0240] Based on the updated parameter estimates, the convective heat transfer resistance between the motor stator node and the coolant node, as well as the convective heat transfer resistance between the controller node and the coolant node, are calculated under the current operating conditions.

[0241] The initial values ​​of the convective heat transfer resistance between the motor stator node and the coolant node, and the initial values ​​of the convective heat transfer resistance between the controller node and the coolant node, were obtained through offline experimental calibration under rated operating conditions. The initial parameter values ​​were set as follows:

[0242]

[0243] Among them, R conv,w(0) R represents the initial value of the convective heat transfer thermal resistance between the motor stator node and the coolant node, in °C / W. conv,s(0) The initial value of the convective heat transfer thermal resistance between the controller node and the coolant node is expressed in °C / W. Both values ​​were obtained through offline experimental calibration under rated operating conditions.

[0244] The initial value of the covariance matrix is ​​determined by the inverse of the cumulative information content of the regression matrix. It characterizes the uncertainty of the initial parameter values ​​and is typically set to a large positive number multiplied by the identity matrix. The initial value of the covariance matrix is ​​set as follows:

[0245]

[0246] Where α is a large positive number, ranging from 10³ to 10⁻⁶. 5 ;I is a 2×2 identity matrix; P(0) is used to characterize the uncertainty of the initial values ​​of the parameters.

[0247] Each row of the regression matrix corresponds to a prediction equation for a temperature node, and its elements consist of the current loss, flow rate, inlet temperature, and the temperature value from the previous time step. The regression matrix is ​​represented as follows:

[0248]

[0249] The elements of the regression matrix are derived from the total loss P of the electric drive system at the current moment. loss (k) Coolant inlet temperature T cin (k), Coolant flow rate q c (k), and the motor winding temperature T at the previous moment. w (k−1), controller heatsink temperature T s (k−1), and coolant outlet temperature T cout (k−1) constitutes the matrix. Specifically, according to the discretized heat balance equation, the elements of the regression matrix can be expressed as:

[0250]

[0251]

[0252]

[0253]

[0254]

[0255]

[0256] Where Δt is the control period in seconds; Cw, Cs, and Cc are the heat capacities of the motor stator node, controller node, and coolant node, respectively, in J / °C; Tc(k−1) is the temperature of the coolant node at the previous moment, calculated from the coolant inlet and outlet temperatures. c (k−1)=(T cin (k−1)+T cout (k−1)) / 2.

[0257] Meanwhile, based on the current motor winding temperature, controller radiator temperature, and coolant outlet temperature, an observation vector is constructed, which contains the three measured temperature values ​​at the current moment.

[0258] The observation vector y(k) is represented as:

[0259]

[0260] Wherein, the observation vector y(k) is determined by the current motor winding temperature T. w (k) Controller heatsink temperature T s (k) and coolant outlet temperature T cout (k) constitutes.

[0261] The size of the gain matrix is ​​determined by both the covariance matrix and the regression matrix, thus determining the degree of influence of the current measurement data on parameter updates. When parameter uncertainty is high or the input data contains a large amount of information, the gain matrix increases accordingly, making the contribution of new data to parameter correction greater. The gain matrix K(k) at the current time is expressed as:

[0262]

[0263] Where K(k) is the gain matrix, which is dimensionless; P(k−1) is the covariance matrix of the previous time step; ϕ(k) is the regression matrix of the current time step, which is a 3×2 matrix whose elements consist of the total loss of the electric drive system at the current time step, the coolant inlet temperature, the coolant flow rate, and the node temperatures of the previous time step, used to describe the linear relationship between the input data and the parameters to be identified; ϕ ⊤ (k) is the transpose of the regression matrix at the current time, which is a 2×3 matrix; λ is the forgetting factor, which ranges from 0.95 to 0.99 and is used to control the weight of historical data in parameter updates.

[0264] The parameter estimates include the reciprocal of the convective heat transfer resistance between the motor stator node and the coolant node, and the reciprocal of the convective heat transfer resistance between the controller node and the coolant node. The update formula is: the current parameter estimate equals the previous parameter estimate plus the difference between the gain matrix multiplied by the observation vector and the transpose of the regression matrix multiplied by the previous parameter estimate. This difference is the model prediction error. This error is used to progressively correct the parameters, making the model predictions approximate the measured values. The steps for obtaining the current parameter estimate θ(k) include:

[0265]

[0266] Where θ(k) is the parameter estimate at the current time. θ1(k)=1 / Rconv,w(k), θ2(k)=1 / Rconv,s(k); θ(k−1) is the parameter estimate of the previous time step.

[0267] The covariance matrix is ​​updated to characterize changes in the uncertainty of parameter estimates. As data accumulates, the covariance matrix gradually decreases, indicating that the parameter estimates are becoming more accurate. Simultaneously, the forgetting factor reduces the weight of historical data, preventing the covariance matrix from becoming too small and thus causing a sluggish response to changes in operating conditions. The covariance matrix at the current moment is expressed as:

[0268]

[0269] Where P(k) is the covariance matrix at the current time; I is the identity matrix.

[0270] Through the formula:

[0271]

[0272] The convective heat transfer resistances between the motor stator node and the coolant node, and between the controller node and the coolant node, under the current operating conditions are calculated. Here, Rconv,w(k) represents the convective heat transfer resistance between the motor stator node and the coolant node under the current operating conditions, in °C / W; Rconv,s(k) represents the convective heat transfer resistance between the controller node and the coolant node under the current operating conditions, in °C / W.

[0273] In the third embodiment described above, the addition of step S106 enables the lumped parameter thermal network model to adaptively track changes in thermal characteristics caused by cooling system performance degradation and power module fatigue throughout its entire life cycle, thereby improving the long-term accuracy of temperature prediction and enhancing the reliability of thermoelectric synergistic control.

[0274] In the fourth embodiment of this application, reference is made to Figure 4 After step S105, step S107 is also executed: when the temperature difference between the motor controller and the motor exceeds a threshold, the objective function of the optimization problem is modified to guide the heat load to transfer between the motor and the motor controller.

[0275] In step S107, the temperature difference between the motor stator node and the controller node is obtained. The motor stator node temperature is determined by the motor winding temperature T. w Characterized by the controller node temperature T, the controller heatsink temperature is determined. s Characterized by this, the temperature difference is expressed as:

[0276]

[0277] Where ΔT is the temperature difference, in °C; when ΔT>0, it means the controller temperature is higher than the motor temperature, and when ΔT<0, it means the motor temperature is higher than the controller temperature.

[0278] Determine if the temperature difference exceeds a preset temperature difference threshold. When |ΔT| > ΔT th When the temperature difference exceeds the limit, the condition is met, where ΔT th The temperature difference threshold is 10 to 20°C and is determined based on the system's thermal characteristics calibration.

[0279] Simultaneously, it determines whether the current motor speed has reached the preset medium-high speed threshold and whether the current motor output torque is lower than the preset medium-low load threshold. The medium-high speed threshold ωth is taken as 50% to 70% of the motor's rated speed, i.e., ω th =η ω ⋅ω rated , where η ω The value ranges from 0.5 to 0.7, ω rated Rated motor speed, in rad / s. Low / medium load threshold T. low Take 30% to 50% of the motor's rated torque, i.e., T low =η T ⋅T rated , where η T The value ranges from 0.3 to 0.5, T rated This is the rated torque of the motor, measured in Newton-meters.

[0280] When the conditions of excessive temperature difference, medium-high speed, and medium-low load are simultaneously met, a heat load transfer mode is triggered, and a temperature difference penalty term is added to the optimization objective function. The modified optimization objective function is expressed as follows:

[0281]

[0282] Where J is the original optimization objective function; w ΔT ΔT is the temperature difference penalty weighting coefficient, ranging from 0.5 to 2 times the original objective function weight, with the specific value determined based on system characteristics; ΔT(k+i|k) is the predicted temperature difference between the motor stator node and the controller node in the i-th step, in °C; ΔT th The temperature difference threshold is expressed in °C; the max(0,⋅) function means that when the absolute value of the predicted temperature difference exceeds the threshold, the excess part is taken, otherwise zero is taken; the summation range is all times in the future prediction time domain.

[0283] When ΔT>0, meaning the controller temperature is higher than the motor temperature, the temperature difference penalty term guides the optimization variables to adjust in the direction of reducing controller losses and increasing motor losses. Specifically, by fine-tuning the modulation ratio or current vector angle, the controller switching losses and conduction losses are reduced. At the same time, by adjusting the ratio of d-axis current and q-axis current, the motor copper losses are increased while keeping the output torque basically unchanged, thereby reducing the controller temperature and increasing the motor temperature, thus narrowing the temperature difference.

[0284] When ΔT < 0, meaning the motor temperature is higher than the controller temperature, the temperature difference penalty term guides the optimization variables to adjust in the direction of increasing controller losses and reducing motor losses, thus achieving the purpose of reducing the temperature difference.

[0285] Solve the corrected optimization problem to obtain the optimal d-axis current command and the optimal q-axis current command at the current moment, and output them to the current controller for execution.

[0286] When the steady-state temperature satisfies |ΔT|≤ΔT th When the temperature difference exceeds the limit, the system automatically exits the heat load transfer mode, reverts to the original optimized objective function, and no longer adds a temperature difference penalty term.

[0287] In the fourth embodiment described above, when the temperature difference between the motor controller and the motor exceeds the threshold, the objective function of the optimization problem is corrected. This can actively guide heat to transfer from the high-temperature side to the low-temperature side when the thermal load is unbalanced, reduce the temperature difference between the motor and the controller, avoid power limiting protection triggered by unilateral overheating, achieve dynamic balance of the system temperature field, and further improve the continuous high-power output capability and long-term operational reliability of the electric drive system.

[0288] By constructing a lumped-parameter thermal network model integrating the motor, controller, and coolant, and using a recursive least squares algorithm with a forgetting factor to identify thermal resistance parameters online, the thermal model can adaptively track long-term drift such as cooling system performance degradation, coolant aging, and power module solder fatigue. This solves the problem of the cumulative increase in temperature prediction error after long-term use of fixed-parameter thermal models in existing technologies, and improves the reliability and adaptability of thermal management.

[0289] Using d-axis and q-axis currents as optimization variables, temperature constraints, current constraints, and voltage constraints are incorporated into a unified model predictive control framework to achieve integrated thermal synergistic optimization of the motor and controller. Compared to traditional passive power reduction protection strategies, the methods in the second to fourth embodiments of this invention can proactively adjust control commands while ensuring the safety of the electric drive system, avoiding power performance loss caused by simple power limiting, and achieving an optimal balance between thermal safety and power output.

[0290] By monitoring the temperature difference between the motor stator node and the controller node, a heat load transfer mode is triggered under medium-to-high speed and medium-to-low load conditions. A temperature difference penalty term is added to the optimization objective function to guide the active transfer of heat between the motor and the controller. When the controller temperature is too high, the controller loss is reduced and the motor loss is increased; conversely, when the motor temperature is too high, the loss is reduced. This allows the system to automatically balance the temperature distribution when the heat load is unbalanced, avoiding power limiting protection caused by unilateral overheating and extending the continuous high-power output capability of the electric drive system.

[0291] Reference Figure 5 The fifth embodiment of this application provides a thermoelectric co-operated motor control device, comprising:

[0292] The acquisition module 201 is used to acquire motor status parameters, motor controller status parameters, and coolant status parameters;

[0293] The total loss determination module 202 is used to determine the motor copper loss, motor iron loss, controller conduction loss and controller switching loss based on the motor state parameters, the motor controller state parameters and the coolant state parameters;

[0294] The thermal resistance parameter identification module 203 is used to construct a lumped parameter thermal network model containing motor stator nodes, motor controller nodes and coolant nodes based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters and the coolant state parameters.

[0295] The current control module 204 is used to solve the optimization problem based on the lumped parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss and the controller switching loss, with the motor d-axis current and q-axis current as optimization variables and preset constraints, to obtain the optimal current command at the current moment.

[0296] The execution module 205 is used to control the motor to execute according to the optimal current command.

[0297] By using an online identification mechanism, the thermal model can adapt to parameter changes throughout its entire life cycle. Model predictive control enables thermal-electric synergistic optimization, and temperature difference penalty terms enable proactive heat load transfer. The organic combination of these three elements allows the electric drive system to maintain an efficient, safe, and balanced operating state across all operating conditions, significantly improving the overall performance of the system.

[0298] The apparatus in the fifth embodiment of this application is an apparatus corresponding to the methods in the second to fourth embodiments described above, and can achieve the same technical effect as the aforementioned methods.

[0299] This application also provides a vehicle including the aforementioned thermoelectric co-operated motor control device.

[0300] It should be understood that the application of this application is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims. Those skilled in the art can understand that implementing all or part of the processes of the above embodiments and making equivalent changes according to the claims of this application still fall within the scope of this application.

Claims

1. A thermoelectrically cooperative motor control method, characterized by, include: Collect motor status parameters, motor controller status parameters, and coolant status parameters; Based on the motor status parameters, the motor controller status parameters, and the coolant status parameters, determine the motor copper loss, motor iron loss, controller conduction loss, and controller switching loss. Based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters, and the coolant state parameters, a lumped parameter thermal network model is constructed, which includes motor stator nodes, motor controller nodes, and coolant nodes. Based on the lumped parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss, the optimization problem is solved with the motor's d-axis current and q-axis current as optimization variables and preset constraints to obtain the optimal current command at the current moment. The motor is controlled to execute the optimal current command.

2. The thermoelectric synergistic motor control method according to claim 1, characterized in that, After the step of constructing a lumped-parameter thermal network model including motor stator nodes, motor controller nodes, and coolant nodes, and before the step of solving the optimization problem based on the lumped-parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss, and the controller switching loss, using the motor's d-axis current and q-axis current as optimization variables, and under preset constraints to obtain the optimal current command at the current moment, the method further includes: Identify the thermal resistance parameters in the lumped parameter thermal network model online.

3. The thermoelectric synergistic motor control method according to claim 2, characterized in that, After the motor executes the optimal current command, the method further includes: When the temperature difference between the motor controller and the motor exceeds a threshold, the objective function of the optimization problem is modified to guide the heat load to transfer between the motor and the motor controller.

4. The thermoelectric synergistic motor control method according to claim 2 or 3, characterized in that, The coolant status parameters include coolant inlet temperature, coolant outlet temperature, and coolant flow rate; Based on the motor copper loss, the motor iron loss, the controller conduction loss, the controller switching loss, the motor state parameters, the motor controller state parameters, and the coolant state parameters, the steps for constructing a lumped parameter thermal network model including motor stator nodes, motor controller nodes, and coolant nodes include: Based on the heat source, heat capacity, and thermal resistance relationship with the coolant node of the motor stator node, the first thermal balance equation of the motor stator node is constructed; wherein, the heat source of the motor stator node consists of the motor copper loss and the motor iron loss; Based on the heat source, heat capacity, and thermal resistance relationship between the controller node and the coolant node, a second thermal balance equation for the controller node is constructed; wherein, the heat source of the controller node consists of the controller conduction loss and the controller switching loss; Based on the thermal resistance relationship between the motor stator node, controller node and coolant node, as well as the heat capacity of the coolant node, the coolant flow rate, the coolant inlet temperature and the coolant outlet temperature, a third heat balance equation for the coolant node is constructed. The first heat balance equation, the second heat balance equation, and the third heat balance equation together constitute the heat balance equation of the lumped parameter heat network model.

5. The thermoelectric synergistic motor control method according to claim 4, characterized in that, The steps for online identification of thermal resistance parameters in the lumped parameter thermal network model include: Discretize the heat balance equation of the lumped parameter heat network model to construct a linear regression model; The linear regression model is identified online using a recursive least squares algorithm with a forgetting factor to obtain the convective heat transfer resistance between the motor stator node and the coolant node, as well as the convective heat transfer resistance between the controller node and the coolant node.

6. The thermoelectric synergistic motor control method according to claim 5, characterized in that, The steps of using a recursive least squares algorithm with a forgetting factor to perform online identification of the linear regression model, and obtaining the convective heat transfer resistance between the motor stator node and the coolant node, and the convective heat transfer resistance between the controller node and the coolant node, include: The initial values ​​of the parameters and the initial values ​​of the covariance matrix are set. The initial values ​​of the parameters include the initial values ​​of the convective heat transfer resistance between the motor stator node and the coolant node and the initial values ​​of the convective heat transfer resistance between the controller node and the coolant node. The initial value of the covariance matrix is ​​determined by the inverse of the cumulative information of the regression matrix and is used to characterize the uncertainty of the initial values ​​of the parameters. In each control cycle, a regression matrix is ​​constructed based on the current total loss of the electric drive system, coolant inlet temperature, coolant flow rate, and the previous motor winding temperature, controller radiator temperature, and coolant outlet temperature; an observation vector is constructed based on the current motor winding temperature, controller radiator temperature, and coolant outlet temperature. The gain matrix for the current time step is calculated based on the covariance matrix of the previous time step, the regression matrix of the current time step, and the forgetting factor. The covariance matrix and the regression matrix together determine the size of the gain matrix, which in turn determines the degree of influence of the measurement data at the current time step on the parameter update. Update the parameter estimates at the current time step based on the parameter estimates from the previous time step, the gain matrix at the current time step, the observation vector at the current time step, and the regression matrix at the current time step. Update the covariance matrix at the current time step based on the covariance matrix at the previous time step, the gain matrix at the current time step, the regression matrix at the current time step, and the forgetting factor. Based on the updated parameter estimates, the convective heat transfer resistance between the motor stator node and the coolant node, as well as the convective heat transfer resistance between the controller node and the coolant node, are calculated under the current operating conditions.

7. The thermoelectric co-operational motor control method according to claim 1, characterized in that, Based on the lumped parameter thermal network model and the total loss, the steps to solve the optimization problem using the motor's d-axis current and q-axis current as optimization variables and with preset constraints to obtain the optimal current command at the current moment include: Set the prediction time domain, and use the d-axis current sequence and q-axis current sequence in the future prediction time domain as optimization variables; Set an optimization objective function, wherein the optimization objective function is at least one of minimizing total system losses, minimizing temperature deviation from the reference value, or maximizing output torque; Set constraints, which include temperature constraints, current hardware limit constraints, and voltage modulation limit constraints predicted based on the lumped parameter thermal network model. An optimization problem is constructed using the optimization variables, the optimization objective function, and the constraints. The optimization problem is solved in each control cycle to obtain the optimal d-axis current command and the optimal q-axis current command at the current moment.

8. The thermoelectric synergistic motor control method according to claim 3, characterized in that, When the temperature difference between the motor controller and the motor exceeds a threshold, the step of correcting the objective function of the optimization problem to guide the transfer of heat load between the motor and the motor controller includes: Obtain the temperature difference between the motor stator node and the controller node; When the temperature difference exceeds a preset temperature difference threshold, and the current motor speed reaches a preset medium-high speed threshold, and the current motor output torque is lower than a preset medium-low load threshold, a temperature difference penalty term is added to the optimization objective function; the temperature difference penalty term is used to guide the optimization variables to adjust in the direction of reducing the temperature difference. Solve the corrected optimization problem to obtain the optimal current command at the current moment.

9. A thermoelectric co-operated motor control device, characterized in that, include: The data acquisition module is used to acquire motor status parameters, motor controller status parameters, and coolant status parameters. The loss determination module is used to determine the motor copper loss, motor iron loss, controller conduction loss and controller switching loss based on the motor state parameters, the motor controller state parameters and the coolant state parameters. The thermal resistance parameter identification module is used to construct a lumped parameter thermal network model containing motor stator nodes, motor controller nodes, and coolant nodes based on the motor copper loss, motor iron loss, controller conduction loss, controller switching loss, motor state parameters, motor controller state parameters, and coolant state parameters. The current control module is used to solve the optimization problem based on the lumped parameter thermal network model, the motor copper loss, the motor iron loss, the controller conduction loss and the controller switching loss, with the motor d-axis current and q-axis current as optimization variables and preset constraints, to obtain the optimal current command at the current moment. The execution module is used to control the motor to execute according to the optimal current command.

10. A vehicle, characterized in that, Includes the thermoelectric co-operated motor control device as described in claim 9.

Images