Hybrid vehicle thermal energy cooperative control method based on dynamic physical envelope

Abstract

The application provides a hybrid vehicle thermal energy cooperative control method based on a dynamic physical envelope, comprising: collecting vehicle state, thermal management state and environmental demand parameters; constructing an augmented heterogeneous state space model containing a three-way catalyst substrate temperature; calculating the maximum heat exchange power upper limit that can be achieved by a front-end heat dissipation module, and taking the upper limit as a hard constraint of an optimization problem; constructing an adaptive cost function containing a weight matrix and a cold start penalty multiplier; first-order Taylor expansion of the augmented heterogeneous state space model forms a local linear time-varying model, combined with the hard constraint and input constraint, the rate constraint, the adaptive cost function is minimized, the optimal control increment sequence is obtained, and the current control quantity is obtained according to the model predictive control rolling execution principle; setting a functional safety monitoring and fault safety response module for three-level response. The method improves the actual effect and precision of control, enhances the stability of the system and helps to prolong the service life of hardware.

Description

Technical Field

[0001] This invention relates to the field of vehicle thermal energy control technology, and specifically to a method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope. Background Technology

[0002] Hybrid vehicles typically include an internal combustion engine, a high-voltage battery, a drive motor, power electronic components, a passenger compartment air conditioning system, and an exhaust after-treatment system. The target operating temperature ranges for these components differ: the engine usually requires a higher coolant temperature range to balance combustion stability and friction loss control; the battery typically requires a narrower suitable temperature window; the motor and electronic control system require temperatures not exceeding the permissible limits of electromagnetic materials and power devices; and the three-way catalytic converter needs to maintain catalytic activity above the ignition temperature. Therefore, thermal management of hybrid vehicles is no longer a simple heat dissipation problem, but rather a system control problem involving the coupling and coordination of multiple heat sources, heat sinks, and actuators.

[0003] In existing technologies, solutions related to thermal management and energy management of hybrid vehicles mainly include: rule-based threshold control schemes, model-based predictive control schemes, and additional thermal management schemes for exhaust aftertreatment. Their main shortcomings are as follows:

[0004] 1) Lack of dynamic physical boundary constraints.

[0005] Many controllers assume that the heat dissipation module has sufficient or stable heat exchange capacity, without introducing the upper limit of heat exchange, which is jointly determined by the heat capacity ratio of the air side and the coolant side, the number of heat transfer units, and the transient temperature difference, into the constraint conditions. When the controller outputs a command that exceeds the current heat exchange limit of the heat dissipation module, the actuator may enter a saturation state, resulting in a mismatch between the control target and the actual heat flow capacity.

[0006] 2) Insufficient smoothness in multi-target switching.

[0007] Rule-based control typically relies on discrete thresholds, while predictive control often employs fixed cost function weights. When the heat supply and demand relationship continuously changes, fixed control logic struggles to consistently reflect the differences between heat shortages and heat abundances, easily leading to frequent adjustments in pump speed and valve position within the temperature critical zone.

[0008] 3) Catalyst status was not included in the system-level closed-loop prediction.

[0009] Existing solutions often only apply additional heating to the catalyst during the engine cold start phase, but there is insufficient online monitoring and forward-looking control of the catalyst substrate temperature during the pure electric coasting and start-stop switching processes unique to hybrid vehicles.

[0010] 4) Insufficient control feasibility and security degradation design.

[0011] The computing power, execution cycle, and real-time performance of the vehicle controller are all limited. If the optimization solver has no feasible solution or times out under certain operating conditions, the control system needs to have a clear degradation and fail-safe mechanism, but existing solutions do not disclose this sufficiently. Summary of the Invention

[0012] To address the shortcomings of existing technologies, this invention provides a method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope.

[0013] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0014] A method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope includes the following steps:

[0015] Step 1: Collect vehicle status, thermal management status, and environmental requirements parameters;

[0016] Step 2: Construct an augmented isomeric state-space model that includes the temperature of the three-way catalytic converter substrate;

[0017] Step 3: Based on the cross-microflow heat transfer mechanism, calculate the maximum heat transfer power limit that the front-end heat dissipation module can achieve online according to the current air-side and coolant-side heat capacity rates, number of heat transfer units and temperature difference, and use this limit as a hard constraint for the optimization problem;

[0018] Step 4: First, continuously adjust the state error weight matrix and control increment weight matrix based on the thermal energy supply and demand tension index. Then, design an exponential cold start penalty multiplier related to coolant temperature. Finally, construct an adaptive cost function that includes the weight matrix and the cold start penalty multiplier.

[0019] Step 5: Perform a first-order Taylor expansion on the augmented heterogeneous state-space model from Step 2 to form a locally linear time-varying model. Combine the hard constraints, input constraints, and rate of change constraints from Step 3 to minimize the adaptive cost function constructed in Step 4, thereby obtaining the optimal control increment sequence. In accordance with the rolling execution principle of model predictive control, the control increment from the first step in the sequence is superimposed with the control quantity from the previous time step to obtain the current control quantity.

[0020] Step 6: Configure the Functional Safety Monitoring and Fault Safety Response module to monitor solution time, solution status, and number of consecutive anomalies, and trigger different levels of degradation actions based on the anomaly level.

[0021] Step 7: Execute steps 1 to 6 in a rolling time-domain loop to form a closed-loop control process.

[0022] Furthermore, the vehicle status includes instantaneous vehicle speed, drive motor output power, and battery state of charge; thermal management status includes engine coolant temperature, coolant pump speed, valve opening, and related measurements for online reconfiguration of the three-way catalytic converter substrate temperature; environmental requirement parameters include ambient temperature, passenger compartment heat load requirements, and vehicle frontal airflow.

[0023] Furthermore, the augmented heterogeneous state-space model is expressed as follows:

[0024]

[0025]

[0026]

[0027] Where k is the slow dynamic scrolling optimization step number, Let be the system state vector. To control the input vector, This represents the external disturbance vector. This is the discretized nonlinear state transition function; The instantaneous speed of the vehicle. This refers to the real-time state of charge of the power battery. This refers to the engine coolant temperature. This refers to the substrate temperature of the three-way catalytic converter. For the output torque of the internal combustion engine, To drive the motor output torque This refers to the speed of the cooling water pump. This refers to the opening degree of a multi-way proportional valve.

[0028] Furthermore, the temperature of the three-way catalytic converter substrate The estimation formula is expressed as:

[0029]

[0030] in, This is an estimated value for the substrate temperature of the three-way catalytic converter. This is the predicted catalyst temperature value for the (j+1)th fast dynamic update step. For the first The Kalman gain matrix for each fast dynamic update step; For measurement vectors; To map the catalyst temperature to a nonlinear observation function for each measurement; This is the control input corresponding to the current fast dynamic step. This represents the external disturbance corresponding to the current fast dynamic step.

[0031] Furthermore, the upper limit of the maximum heat exchange power The formula is expressed as:

[0032]

[0033]

[0034] in, This refers to the inlet temperature of the coolant in the heat dissipation module. Ambient temperature; For minimum heat capacity, The ratio of heat capacity factor, The number of heat transfer units, This refers to the heat exchange efficiency factor.

[0035] Furthermore, the hard constraint formula is expressed as:

[0036]

[0037] in, In order to achieve the current slow dynamic control step The future time predicted at each moment The target heat exchange requirement for a slow dynamic control step; In order to achieve the current slow dynamic control step The future time predicted at each moment The maximum heat exchange power that can be achieved in one slow dynamic control step; To predict the length of the time domain.

[0038] Furthermore, the state error weight matrix and control increment weight matrix The formula is expressed as:

[0039]

[0040]

[0041]

[0042] in, The basic state error weight matrix; This is the incremental weight matrix added to the state error term when the heat becomes tighter; Based on the control increment weight matrix; This is the incremental weight matrix added to the control smoothing term when heat is abundant; As an indicator of the tension between heat energy supply and demand; No. Total predicted heat demand within one slow dynamic control step; This refers to the upper limit of the achievable maximum heat exchange power obtained in step three. It is the minimum regularization constant; This serves as the lower bound for the indicator of the tension between heat energy supply and demand. This is the upper limit of the indicator for the tension between heat energy supply and demand.

[0043] Furthermore, the exponential cold start penalty multiplier The formula is expressed as:

[0044]

[0045] in, This is the penalty amplitude coefficient; The penalty attenuation coefficient; The target operating temperature for the engine; This is the current coolant temperature.

[0046] Furthermore, the adaptive cost function The formula is expressed as:

[0047]

[0048] in, To predict the length of the time domain; To control the length of the time domain; In order to achieve the current slow dynamic control step The next step towards the future The state prediction vector for each slow dynamic control step. The reference trajectory for the state prediction vector; For the future The control increment of a slow dynamic control step; The baseline fuel consumption rate for the predicted time; This refers to the weighting coefficient for the fuel consumption item. The weights are for the soft constraint relaxation terms. These are slack variables.

[0049] Furthermore, the functional safety monitoring and fault safety response module includes a first-level response to determine computing power exceeding limits, a second-level response to determine that there is no feasible solution, and a third-level response to determine continuous anomalies.

[0050] The beneficial effects of this invention are as follows:

[0051] 1) By calculating online and applying a dynamic physical envelope as a hard constraint, this invention ensures that the controller's output command is always within the actual capability range of the heat dissipation system, avoiding control failure and deviation caused by command overshoot, and improving the actual control effect and accuracy.

[0052] 2) By introducing a continuous thermal energy supply and demand tension index to dynamically adjust the cost function weight, this invention enables the control strategy to smoothly transition between different operating conditions such as thermal tension and thermal abundance, effectively avoiding frequent actuator actions caused by logical abrupt changes near the critical point, enhancing system stability and helping to extend hardware life.

[0053] 3) By incorporating catalyst temperature into system-level closed-loop control through virtual sensors, this invention enables the controller to proactively manage the thermal state of the catalyst under hybrid-specific operating conditions such as pure electric coasting and engine restart, which helps maintain its catalytic activity and thus improves the vehicle's transient emission performance.

[0054] 4) By designing a multi-level fault-safe response mechanism, this invention provides a robust safety guarantee for the application of computationally intensive advanced control algorithms on vehicle-mounted platforms with limited computing power and variable environments. When a solution anomaly occurs, the system can automatically and smoothly degrade to a safe or backup operating mode, greatly improving the engineering robustness and mass production application value of this technical solution. Attached Figure Description

[0055] Figure 1 This is a flowchart of the hybrid vehicle thermal energy coordinated control method based on dynamic physical envelope as described in this invention. Detailed Implementation

[0056] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the scope of protection of the present invention is not limited thereto.

[0057] like Figure 1 As shown, a method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope is presented. This method employs hierarchical, multi-timescale model predictive control logic, taking vehicle operating state, thermal management state, and catalytic converter state as inputs, and engine torque, motor torque, coolant pump speed, and multi-way valve opening as control objects. Dynamic physical envelope constraints of the cooling module are injected during the optimization process. This method can be implemented according to the following steps.

[0058] Step 1: Collect vehicle status, thermal management status, and environmental requirements parameters.

[0059] The system collects vehicle operating status, thermal management system status, and environmental requirements parameters. The collected vehicle status includes, but is not limited to, instantaneous vehicle speed, drive motor output power, and battery state of charge. Thermal management status includes engine coolant temperature, coolant pump speed, valve opening, and relevant measurements used for online reconstruction of the three-way catalytic converter substrate temperature. These relevant measurements preferably include one or more of the following: engine speed, real-time fuel injection quantity, catalytic converter inlet exhaust temperature, oxygen sensor signal, and estimated exhaust flow rate. Environmental requirements parameters include ambient temperature, passenger compartment heat load requirements, and vehicle frontal airflow. This collected data serves as input for subsequent catalytic converter temperature reconstruction, dynamic physical envelope calculation, and cost function construction, ensuring the controller can perform closed-loop optimization control based on the current actual operating conditions.

[0060] Step 2: Construct an augmented isomeric state-space model incorporating the substrate temperature of the three-way catalytic converter.

[0061] A nonlinear discrete dynamic model equation incorporating the longitudinal dynamics and thermophysical characteristics of the vehicle is established, namely, an augmented heterogeneous state-space model:

[0062]

[0063] in, This indicates the slow dynamic rolling optimization step number, corresponding to the discrete update time of the model predictive control; For the first The system state vector for each slow dynamic control step; For the first The control input vector for each slow dynamic control step; For the first The external disturbance vector for each slow dynamic control step; It is a discretized nonlinear state transition function used to characterize the comprehensive evolution relationship of vehicle dynamics, thermal management state, and catalyst thermal state.

[0064] To reflect the dual time-scale characteristics of fast actuator refresh and hot state rolling optimization, a fast dynamic update step sequence number is further introduced. This is used to represent discrete moments in actuator control, rapid heat distribution, and catalyst temperature state estimation. Preferably, the slow dynamic time step is denoted as . The fast dynamic time step is recorded as And satisfy More preferably, The value ranges from 10 milliseconds to 50 milliseconds and is used for actuator control and rapid heat distribution updates; The value ranges from 100 milliseconds to 1000 milliseconds and is used for slower hot refresh processes and scrolling optimization updates.

[0065] The system state vector of the model Defined as:

[0066]

[0067] in, The instantaneous speed of the vehicle. This refers to the real-time state of charge of the power battery. This refers to the engine coolant temperature. This refers to the substrate temperature of the three-way catalytic converter. (This is for...) To address the difficulty of direct measurement, a virtual sensor based on extended Kalman filtering is introduced. Using known parameters such as engine speed and real-time fuel injection quantity, the temperature of the three-way catalytic converter substrate is reconstructed and estimated online.

[0068] The online reconstruction and estimation process of the three-way catalytic converter substrate temperature is as follows:

[0069] The catalyst lumped heat balance prediction model is written as follows:

[0070]

[0071] in, This indicates a fast dynamic update of the step number; For the first The catalyst substrate temperature is updated rapidly in a single dynamic step; The predicted catalyst substrate temperature for the next rapid dynamic update step; For fast dynamic time steps; The equivalent heat capacity of the catalyst; For the first A rapid dynamic update step flows into the equivalent heat input of the catalyst; For the first The rapid dynamic update step reduces heat loss from the catalytic converter. This is process noise, used to characterize unmodeled dynamics, parameter perturbations, and external disturbances.

[0072] in, The optimal value is determined by the equivalent thermal mass of the catalyst and the average specific heat capacity of the material, and can be written as:

[0073]

[0074] in, The equivalent thermal mass of the catalyst support and encapsulation; This represents the average specific heat capacity of the catalyst material.

[0075] Preferably, the heat input item can be written as:

[0076]

[0077] in, This represents the real-time fuel injection quantity or fuel consumption per unit time in the j-th fast dynamic update step. The lower heating value of fuel oil. For the engine speed in the j-th fast dynamic update step, The catalytic converter inlet exhaust temperature or its estimated value; , , The coefficients to be calibrated represent the contributions of fuel chemical exothermics, engine speed-dependent exhaust enthalpy flow, and exhaust temperature to the catalytic converter's heat input, respectively. To ensure the heat input term has a clear physical meaning and avoid estimation divergence, the following conditions are preferably met: > 0、 ≥ 0、 ≥ 0, and makes it work across the entire operating range. ≥ 0.

[0078] Preferably, the catalyst heat loss item Written as:

[0079]

[0080] in, The overall heat transfer coefficient of the catalyst to the outside world; This is the equivalent heat exchange area of ​​the catalyst. For the first The ambient temperature is updated rapidly.

[0081] Within the extended Kalman filter framework, based on the aforementioned catalyst lumped heat balance prediction model, the catalyst temperature prediction equation can be written as:

[0082]

[0083] in, This is the predicted catalyst temperature value for the (j+1)th fast dynamic update step. This is the catalyst temperature estimate after correction in the j-th fast dynamic update step. Let be the nonlinear state transition function constructed from the lumped heat balance prediction model. This is the control input corresponding to the current fast dynamic step. This represents the external disturbance corresponding to the current fast dynamic step.

[0084] In the state prediction step of the extended Kalman filter, the prediction error covariance propagation equation corresponding to the catalyst temperature prediction equation can be written as:

[0085]

[0086] in, For the first The prediction error covariance matrix of each fast dynamic update step; For the first The error covariance matrix after a fast dynamic update step is updated; The Jacobian matrix is ​​the local linearization of the state transition function with respect to the state variables. Let be the process noise covariance matrix.

[0087] In the measurement update step of the extended Kalman filter, the nonlinear measurement equation corresponding to the catalyst temperature estimate can be written as:

[0088]

[0089] in, For measurement vectors; To map the catalyst temperature to a nonlinear observation function for each measurement; This is for measurement noise. The measurement vector It can be composed of engine speed, fuel injection quantity, catalytic converter inlet exhaust temperature, oxygen sensor signal, exhaust flow rate estimate or a combination thereof.

[0090] Based on the aforementioned covariance propagation equation and measurement equation, the gain matrix of the extended Kalman filter can be further obtained:

[0091]

[0092] in, For the first The Kalman gain matrix for each fast dynamic update step; The Jacobian matrix is ​​the locally linearized version of the observation function with respect to the state variables. This is the measurement noise covariance matrix.

[0093] The catalyst temperature is corrected using the gain matrix to obtain the catalyst temperature estimate for the current fast dynamic update step:

[0094]

[0095] The values ​​within square brackets represent the measurement residuals, indicating the difference between actual measurements and model-predicted measurements. A larger Kalman gain indicates that the current estimate relies more on real-time measurements; a smaller Kalman gain indicates that it relies more on model predictions.

[0096] Finally, with the corrected This serves as the estimated temperature of the three-way catalytic converter substrate in the current fast dynamic update step, and is then mapped or updated to the state variables corresponding to the slow dynamic rolling optimization time. This allows the controller to directly use the estimated catalyst temperature instead of the catalyst substrate temperature, which cannot be directly measured in real time, during subsequent rolling optimization processes, thus incorporating the catalyst thermal state into the closed loop of vehicle thermal energy coordinated control.

[0097] The control input vector of the augmented heterogeneous state-space model Defined as:

[0098]

[0099] in, For the output torque of the internal combustion engine, To drive the motor to output torque, This refers to the speed of the cooling water pump. This refers to the opening degree of a multi-way proportional valve.

[0100] External disturbance vector It includes at least one or more of the following: ambient temperature, passenger compartment heat load requirements, driver torque or power requirements, and incoming air conditions caused by vehicle movement.

[0101] Step 3: Calculate the extreme value envelope online based on the cross-flow microscopic heat transfer mechanism and inject hard constraints.

[0102] In this embodiment, the vehicle's front-end cooling module is considered a dynamically heat-limited unit. Within each control cycle, the maximum achievable heat transfer power limit is calculated online based on the current air-side and coolant-side heat capacity rates, the number of heat transfer units, and the temperature difference. The process is as follows:

[0103] First, calculate the air heat capacity factor. With the heat capacity ratio of the coolant :

[0104]

[0105]

[0106] in, Air mass flow rate, This refers to the coolant mass flow rate. The specific heat capacity of air at constant pressure. The specific heat capacity at constant pressure of the coolant.

[0107] Further definition:

[0108]

[0109]

[0110] in, For minimum heat capacity, This is the heat capacity ratio.

[0111] Next, calculate the number of heat transfer units. :

[0112]

[0113] in, The transient overall heat transfer coefficient is... For effective heat exchange area.

[0114] For cross-flow heat transfer conditions where the air and coolant sides are not mixed, the heat transfer efficiency factor The following approximation relationship can be used:

[0115]

[0116] The above formula is an engineering approximation expression for the cross-flow heat transfer efficiency-number of heat transfer units method, used to give an upper limit estimate of the heat transfer capacity under current operating conditions. The exponents 0.78 and -0.22 are empirical exponents in this approximation relationship.

[0117] This leads to the dynamic physical envelope:

[0118]

[0119] in, This represents the upper limit of the maximum heat transfer power that the heat dissipation module can achieve within the k-th slow dynamic control step. This refers to the inlet temperature of the coolant in the heat dissipation module. The ambient temperature.

[0120] Incorporate this dynamic physical envelope into the optimization problem as a hard constraint:

[0121]

[0122] in, In order to achieve the current slow dynamic control step The future time predicted at each moment The target heat exchange requirement for a slow dynamic control step; In order to achieve the current slow dynamic control step The future time predicted at each moment The maximum heat exchange power that can be achieved in one slow dynamic control step; To predict the length of the time domain.

[0123] The purpose of the above inequality constraints is to limit the optimization process of the controller to the current physical realizable boundary of the heat dissipation module, and to prevent the solver from giving control quantities that do not have actual heat transfer support conditions.

[0124] Step 4: Construct an adaptive cost function that includes a continuously adjusting index and a cold start penalty multiplier.

[0125] First, in order to achieve continuous adjustment between heat shortage and heat abundance conditions, this embodiment defines a system heat energy supply and demand tension index. :

[0126] First, define the total predicted heat demand within the k-th slow dynamic control step. :

[0127]

[0128] in, To meet the thermal needs of the crew cabin, To meet the thermal requirements of power battery temperature control, For engine thermal management requirements, To meet the requirements of maintaining or restoring the thermal state of the catalyst, to These are non-negative weight parameters.

[0129] Further define the indicators of heat energy supply and demand tension :

[0130]

[0131] in, For the first one defined above The total predicted heat demand within one slow dynamic control step This refers to the upper limit of the achievable maximum heat exchange power obtained in step three. To prevent the minimum regularization constant where the denominator is zero; This serves as the lower bound for the indicator of the tension between heat energy supply and demand. This represents the upper bound of the heat energy supply-demand tension index. This embodiment preferably adopts: .

[0132] In order to The adjustment relationship for the subsequently constructed adaptive cost function is clear. In this embodiment, the state error weight matrix is ​​preferably used. and control increment weight matrix It can be written in the following continuous mapping form:

[0133]

[0134]

[0135] in, The basic state error weight matrix; This is the incremental weight matrix added to the state error term when the heat becomes tighter; Based on the control increment weight matrix; This is the incremental weight matrix added to the control smoothing term when there is sufficient heat.

[0136] The above mapping relationship reveals the interaction between parameters: when When it increases, An increase indicates that the system places greater emphasis on temperature-related state deviations; simultaneously A relative decrease indicates that the actuator is allowed to make larger control adjustments. When When the error term is reduced, the state error term is relatively lessened, while the control increment smoothing term is relatively heavier, and the control strategy is more biased towards smoothness and energy consumption constraints.

[0137] Secondly, to smoothly handle low-temperature start-up scenarios, this embodiment introduces an exponential cold start penalty multiplier. :

[0138]

[0139] in, This is the penalty amplitude coefficient, used to control the order of magnitude of the additional penalty intensity at low temperatures; This is the penalty attenuation coefficient, used to control the rate at which the penalty decreases as the coolant temperature approaches the target temperature; The target operating temperature for the engine; This refers to the current coolant temperature. Preferably, , .

[0140] Finally, the adaptive cost function for predicting and optimizing the objective is obtained. :

[0141]

[0142] in, For the first Predictive optimization cost function for each slow dynamic control step; To predict the length of the time domain; To control the time domain length, and preferably satisfy ; In order to achieve the current slow dynamic control step The next step towards the future The state prediction vector for each slow dynamic control step. The reference trajectory for the state prediction vector; For the future The control increment of a slow dynamic control step; The baseline fuel consumption rate for the predicted time; This refers to the weighting coefficient for the fuel consumption item. The weights are for the soft constraint relaxation terms. These are slack variables.

[0143] Preferably, , This range is derived from a comprehensive consideration of the thermal system hysteresis and the real-time performance of the on-board controller.

[0144] Step 5: Model Dimensionality Reduction and Optimization Solution

[0145] Since the system model in step two is a nonlinear model, this embodiment performs a first-order Taylor expansion around the current operating point in each control cycle to form a locally linear time-varying model:

[0146]

[0147] in, The state Jacobian matrix; The input is the Jacobian matrix; It is a compensation vector consisting of a linearized intercept term and a perturbation term.

[0148] The linearization form described above is used to transform the nonlinear optimization problem of each control cycle into a local quadratic programming problem, which is easier to execute on the vehicle controller.

[0149] In the control time domain, in addition to the dynamic physical envelope constraint from step three, the optimization problem also includes the following input and rate-of-change constraints:

[0150]

[0151]

[0152] in, , These are the lower and upper bounds of the control quantity, respectively. , These are the lower and upper bounds of the control increment, respectively. In order to achieve the current slow dynamic control step Predicting the future at any moment A slow dynamic control step controls the input vector.

[0153] The physical boundaries of each control variable are described below:

[0154] Limited by engine torque limits; Limited by the peak torque of the motor and the power limit of the battery; Limited by the upper limit of the pump's structural rotational speed and the minimum stable rotational speed; The opening range is usually between fully closed and fully open.

[0155] In each control cycle, the solver minimizes the cost function constructed in step four. The length is obtained as Optimal control increment sequence :

[0156]

[0157] in, For the current slow dynamic control step For the future The optimal control increment sequence obtained step by step; For the first The optimal control increment.

[0158] According to the rolling execution principle of model predictive control, only the first step of the control increment is executed, and this control increment is added to the control quantity at the previous moment to obtain the current control quantity:

[0159]

[0160] in, This is the actual control input issued for the current slow dynamic control step; The control input has been executed in the previous slow dynamic control step; This is the first optimal control increment obtained in the current step.

[0161] The above strategy ensures that the controller can recalculate the control sequence based on the latest operating conditions in each control cycle and maintain closed-loop updates.

[0162] Step Six: Configure the Functional Safety Monitoring and Fault Safety Response Module

[0163] In order to enable the control scheme to adapt to the real-time control environment of the vehicle, this embodiment sets up a functional safety monitoring and fault safety response module in parallel outside the optimized solver. This module is used to monitor the solution time, solution status and number of consecutive anomalies, and trigger different levels of degradation actions according to the anomaly level.

[0164] (1) Level 1 response: translation vector continuation control when computing power exceeds limit

[0165] When the solver's computation time in the current control cycle exceeds a preset safe clock cycle limit, it is determined that the computing power has exceeded the limit. In this case, the current unconverged solution result is not used; instead, the remaining portion of the optimal control increment sequence from the previous control cycle is extracted, time-shifted, and used as the current control output reference. Preferably, the solution timeout threshold can be set to 0.6 to 0.9 times the fast dynamic time step to ensure that the solution result still has execution timeliness within the current control cycle. This method is used to maintain the continuity of the control sequence even when a single solution timeout occurs but the system can still continue running.

[0166] (2) Secondary response: When no feasible solution is found, switch to rule feedback control.

[0167] When the solver returns a state of no feasible solution, or when the condition number of the Jacobian matrix deteriorates due to parameter mutations, causing the current optimization problem to fail to converge, the current Model Predictive Control (MPC) output is immediately terminated, and the system switches to a preset rule feedback controller.

[0168] The rule feedback controller can output minimum pump speed and valve position commands based on the coolant temperature, catalytic converter temperature estimate, and passenger compartment thermal requirements.

[0169] Its purpose is not to pursue global optimization, but to maintain the basic availability of thermal management and avoid control interruptions.

[0170] (3) Level 3 response: Trigger the underlying hardware to prioritize risk avoidance when there are continuous anomalies.

[0171] When multiple Level 1 or Level 2 anomalies occur consecutively, or when the underlying diagnostic and optimization control thread continuously crashes, a higher-priority hardware avoidance mode is triggered. This mode directly issues safety priority commands to the heat dissipation-related actuators, such as elevating the heat dissipation actuators to a preset minimum protection state, allowing the vehicle's thermal system to return to a conservative and safe range. Preferably, the threshold for the number of consecutive anomalies can be set to 2 to 5 times.

[0172] Step 7: Execute the control cycle in a rolling time domain loop.

[0173] After completing the above steps, the system enters the next control cycle. The system re-collects operating conditions, updates state estimates, calculates dynamic physical envelopes, refreshes cost functions, executes optimization solutions and safety monitoring, thereby forming a closed-loop control process.

[0174] The embodiments described above are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Any obvious improvements, substitutions or modifications that can be made by those skilled in the art without departing from the essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope, characterized in that, Includes the following steps: Step 1: Collect vehicle status, thermal management status, and environmental requirements parameters; Step 2: Construct an augmented isomeric state-space model that includes the temperature of the three-way catalytic converter substrate; Step 3: Based on the cross-microflow heat transfer mechanism, calculate the maximum heat transfer power limit that the front-end heat dissipation module can achieve online according to the current air-side and coolant-side heat capacity rates, number of heat transfer units and temperature difference, and use this limit as a hard constraint for the optimization problem; Step 4: First, continuously adjust the state error weight matrix and control increment weight matrix based on the thermal energy supply and demand tension index. Then, design an exponential cold start penalty multiplier related to coolant temperature. Finally, construct an adaptive cost function that includes the weight matrix and the cold start penalty multiplier. Step 5: Perform a first-order Taylor expansion on the augmented heterogeneous state-space model from Step 2 to form a locally linear time-varying model. Combine the hard constraints, input constraints, and rate of change constraints from Step 3 to minimize the adaptive cost function constructed in Step 4, thereby obtaining the optimal control increment sequence. In accordance with the rolling execution principle of model predictive control, the control increment from the first step in the sequence is superimposed with the control quantity from the previous time step to obtain the current control quantity. Step 6: Configure the Functional Safety Monitoring and Fault Safety Response module to monitor solution time, solution status, and number of consecutive anomalies, and trigger different levels of degradation actions based on the anomaly level. Step 7: Execute steps 1 to 6 in a rolling time-domain loop to form a closed-loop control process.

2. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 1, characterized in that, The vehicle status includes instantaneous vehicle speed, drive motor output power, and battery state of charge; the thermal management status includes engine coolant temperature, coolant pump speed, valve opening, and related measurements used for online reconstructing of the three-way catalytic converter substrate temperature. Environmental requirements parameters include ambient temperature, passenger compartment heat load requirements, and vehicle frontal airflow.

3. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 1, characterized in that, The augmented heterogeneous state-space model is expressed as follows: Where k is the slow dynamic scrolling optimization step number. Let be the system state vector. To control the input vector, This represents the external disturbance vector. This is the discretized nonlinear state transition function; The instantaneous speed of the vehicle. This refers to the real-time state of charge of the power battery. This refers to the engine coolant temperature. This refers to the substrate temperature of the three-way catalytic converter. For internal combustion engine output torque To drive the motor output torque Cooling water pump speed This refers to the opening degree of a multi-way proportional valve.

4. The hybrid vehicle thermal energy coordinated control method based on dynamic physical envelope according to claim 3, characterized in that, The temperature of the three-way catalytic converter substrate The estimation formula is expressed as: in, This is an estimated value for the substrate temperature of the three-way catalytic converter. This is the predicted catalyst temperature value for the (j+1)th fast dynamic update step. For the first The Kalman gain matrix for each fast dynamic update step; For measurement vectors; To map the catalyst temperature to a nonlinear observation function for each measurement; This is the control input corresponding to the current fast dynamic step. This represents the external disturbance corresponding to the current fast dynamic step.

5. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 1, characterized in that, The maximum heat exchange power limit The formula is expressed as: in, This refers to the inlet temperature of the coolant in the heat dissipation module. Ambient temperature; For minimum heat capacity, The ratio of heat capacity factor, The number of heat transfer units, This refers to the heat exchange efficiency factor.

6. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 5, characterized in that, The hard constraint formula is expressed as: in, In order to achieve the current slow dynamic control step The future time predicted at each moment The target heat exchange requirement for a slow dynamic control step; In order to achieve the current slow dynamic control step The future time predicted at each moment The maximum heat exchange power that can be achieved in one slow dynamic control step; To predict the length of the time domain.

7. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 1, characterized in that, The state error weight matrix and control increment weight matrix The formula is expressed as: in, The basic state error weight matrix; This is the incremental weight matrix added to the state error term when the heat becomes tighter; Based on the control increment weight matrix; This is the incremental weight matrix added to the control smoothing term when heat is abundant; As an indicator of the tension between heat energy supply and demand; No. Total predicted heat demand within one slow dynamic control step; This refers to the upper limit of the achievable maximum heat exchange power obtained in step three. It is the minimum regularization constant; This serves as the lower bound for the indicator of the tension between heat energy supply and demand. This is the upper limit of the indicator for the tension between heat energy supply and demand.

8. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 7, characterized in that, The exponential cold start penalty multiplier The formula is expressed as: in, This is the penalty amplitude coefficient; The penalty attenuation coefficient; The target operating temperature for the engine; This is the current coolant temperature.

9. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 8, characterized in that, The adaptive cost function The formula is expressed as: in, To predict the length of the time domain; To control the length of the time domain; In order to achieve the current slow dynamic control step The next step towards the future The state prediction vector for each slow dynamic control step. The reference trajectory for the state prediction vector; For the future The control increment of a slow dynamic control step; The baseline fuel consumption rate for the predicted time; This refers to the weighting coefficient for the fuel consumption item. The weights are for the soft constraint relaxation terms. These are slack variables.

10. The method for coordinated thermal energy control of hybrid vehicles based on dynamic physical envelope according to claim 1, characterized in that, The functional safety monitoring and fault safety response module includes a level 1 response to determine computing power exceeding limits, a level 2 response to determine that there is no feasible solution, and a level 3 response to determine continuous anomalies.

Images