An energy residual-based electric drive thermal management deviation compensation and early warning method

Abstract

The application discloses an energy residual-based electric drive thermal management deviation compensation early warning method, which comprises the following steps: establishing an electric drive system thermal management lumped parameter thermal network model to obtain thermal state variables; calculating a normalized energy residual absolute value; constructing a quasi-steady state judgment index; outputting a steady state permission signal when the index meets a preset entering condition; starting online reconstruction and compensation when the normalized energy residual absolute value meets a preset triggering condition; constructing a variable step normalized projection recursive deviation estimator to perform online reconstruction on a target sensor deviation; reconstructing the normalized energy residual absolute value based on the compensated measurement, combining a continuous counting judgment mechanism to output a fault early warning result; and taking the compensated measurement as a feedback input of an electric drive system thermal management controller. The application can solve the problems that the prior art is difficult to realize online reconstruction under the constraint of thermal management mechanism, is difficult to balance convergence speed and compensation stability, and is prone to misstarting the deviation estimation under a dynamic working condition.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for electric drive systems of new energy vehicles, and specifically to a method for early warning of deviation compensation in electric drive thermal management based on energy residual. Background Technology

[0002] Thermal management of electric drive systems is a crucial foundation for ensuring the safety, efficiency, and lifespan of the powertrain in new energy vehicles. Its control precision is highly dependent on the accuracy of sensor measurements such as coolant mass flow rate and inlet coolant temperature.

[0003] Existing methods for thermal management of electric drive systems often employ threshold judgment, empirical rules, fixed parameter observers, or fixed gain compensation strategies. When there are measurement deviations in the coolant mass flow sensor or inlet coolant temperature sensor, they can usually only achieve anomaly detection or rough correction. It is difficult to achieve online reconstruction of sensor deviations under the constraints of thermal management mechanisms. Especially when the operating conditions change rapidly, the thermal capacity storage term cannot be ignored, or the noise disturbance is strong, the measured quantities contaminated by the deviation will simultaneously affect thermal state estimation, fault diagnosis, and controller feedback, which can easily lead to false alarms, missed alarms, and the accumulation of control deviations.

[0004] In addition, existing partial deviation estimation methods mostly adopt fixed step size or fixed gain recursive update mechanism, which makes it difficult to balance compensation speed and compensation stability. They usually converge slowly in the early stage of the fault, and when the estimated value is close to the actual deviation, they are prone to oscillation, overshoot or jitter, which in turn affects the compensation accuracy and the reliability of subsequent early warning judgment.

[0005] Meanwhile, some existing methods lack reliable start-up conditions that are integrated with the thermal management mechanism. They often start the deviation estimation process only after an anomaly is detected, making it difficult to avoid false starts under dynamic and severe operating conditions, model mismatch conditions, or significant transient energy storage conditions. Furthermore, existing technologies usually fail to integrate sensor deviation compensation, fault warning, and thermal management closed-loop control, resulting in deviation-containing measurements continuously participating in control feedback, which reduces the robustness and control accuracy of the thermal management system. Summary of the Invention

[0006] In view of this, the present invention provides an electric drive thermal management deviation compensation and early warning method based on energy residual, in order to solve the problems in the prior art that the target sensor deviation is difficult to reconstruct online under the constraint of thermal management mechanism, fixed step size compensation is difficult to balance convergence speed and compensation stability, deviation estimation is prone to false start under dynamic operating conditions, and fault early warning and control feedback are easily affected by deviation-containing measurement quantities.

[0007] A method for early warning of deviation compensation in electric drive thermal management based on energy residual includes: Step S1: Establish a lumped parameter thermal network model for the thermal management of the electric drive system oriented to explicit mapping of target sensor deviation, so as to introduce the target sensor measurement into the energy balance relationship of the thermal management of the electric drive system, and obtain the thermal state quantity characterizing the heat transfer relationship of the thermal management of the electric drive system. The target sensor includes a coolant mass flow sensor and an inlet coolant temperature sensor. Step S2: Collect the operating parameters and measurement parameters of the thermal management of the electric drive system, construct the energy residual based on the lumped parameter thermal network model, and calculate the normalized absolute value of the energy residual; Step S3: Construct a quasi-steady-state determination index based on the rate of change of thermal state quantities. When the quasi-steady-state determination index meets the preset entry conditions, output a steady-state allow signal. When the absolute value of the normalized energy residual meets the preset trigger conditions, proceed to step S4 to start the online reconstruction and compensation of the target sensor deviation. Step S4: Construct a variable step size normalized projection recursive deviation estimator driven by residual sensitivity based on the energy residual and residual sensitivity, calculate the residual sensitivity of the energy residual to the target sensor deviation, and then reconstruct the target sensor deviation online to obtain the deviation estimate, and compensate the measured quantity based on the deviation estimate. Step S5: Based on the compensated measurement, reconstruct the normalized energy residual absolute value, compare the reconstructed normalized energy residual absolute value with the preset alarm threshold and the preset release threshold respectively, and output the fault warning result in combination with the continuous counting judgment mechanism; at the same time, use the compensated measurement to replace the original measurement with deviation as the feedback input of the electric drive system thermal management controller.

[0008] The method for early warning of deviation compensation in electric drive thermal management based on energy residual provided by the present invention has the following beneficial effects: 1. This invention establishes a lumped-parameter thermal network model for explicit mapping of target sensor bias, enabling coolant mass flow rate and inlet coolant temperature to explicitly enter the energy balance residual construction process, thus achieving online identification and reconstruction of target sensor bias under the constraints of thermal management mechanisms.

[0009] 2. This invention proposes a residual sensitivity-driven variable step-size normalized projection recursive bias estimator. Through the synergistic effect of residual sensitivity, variable step-size adjustment factor, residual sensitivity normalization term, residual dead zone function and projection constraint operator, it achieves rapid convergence in the stage of large bias and smooth approximation in the stage of small bias, thereby simultaneously improving compensation speed, compensation accuracy and compensation stability.

[0010] 3. This invention designs a reliable activation logic of "quasi-steady-state gating + residual triggering + continuous counting", and realizes fault early warning output and control feedback based on the compensated measurement, thereby reducing the risk of false updates, false alarms and missed alarms under dynamic operating conditions, and improving the robustness of thermal management and closed-loop control accuracy of electric drive system. Attached Figure Description

[0011] Figure 1 A flowchart illustrating the method for early warning compensation of electric drive thermal management deviation based on energy residual provided in an embodiment of the present invention; Figure 2 The figure shows a comparison of the online reconstruction and compensation results of the method of this invention and the basic compensation method under the condition of coolant mass flow sensor deviation failure. Figure 3 The figure shows a comparison of online reconstruction and compensation results using the method of this invention and the basic compensation method under the condition of inlet coolant temperature sensor deviation failure. Detailed Implementation

[0012] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain embodiments of the present invention, and should not be construed as limiting the present invention.

[0013] Please see Figure 1 The present invention provides a method for early warning of deviation compensation for electric drive thermal management based on energy residual, comprising steps S1 to S5: Step S1: Establish a lumped parameter thermal network model for the thermal management of the electric drive system oriented to explicit mapping of target sensor deviation, so as to introduce the target sensor measurement into the energy balance relationship of the thermal management of the electric drive system, and obtain the thermal state quantity characterizing the heat transfer relationship of the thermal management of the electric drive system. The target sensor includes a coolant mass flow sensor and an inlet coolant temperature sensor.

[0014] In this embodiment, the thermal management object of the electric drive system is the drive motor and its cooling circuit as the core, and includes related sensors, controllers and actuators for thermal status monitoring, sensor deviation compensation, fault early warning and closed-loop control.

[0015] In this embodiment, a four-node lumped parameter thermal network is used to describe the main heat transfer relationship between the drive motor and the cooling circuit. The four thermal nodes are the winding thermal node, the stator core thermal node, the housing thermal node, and the coolant thermal node.

[0016] Step S1 specifically includes: Define the thermal state variables as:

[0017] in, For discrete sampling time indexes, for Thermal state parameters at any given time for The temperature of the winding hot node at any given time. for The stator core hot junction temperature at any given moment. for The temperature of the casing thermal node at any given time. for The temperature of the coolant at the hot junction at any given time. This indicates transpose.

[0018] Define parameter vector for:

[0019] in, , , and These are the equivalent heat capacities of the windings, stator core, housing, and coolant hot nodes, respectively. , , and These are the contact thermal resistance between the winding and the stator core, the contact thermal resistance between the stator core and the housing, the convective and radiative thermal resistance between the housing and the external environment, and the reference heat transfer thermal resistance, respectively. As the baseline mass flow rate, For traffic-related indices, This refers to the specific heat capacity of the coolant.

[0020] Construct a lumped-parameter thermal network model, expressed as:

[0021]

[0022]

[0023]

[0024] in, To represent the differential, Indicates the continuous sampling time. , The sampling period is , , and These are the winding hot node temperature, stator core hot node temperature, housing hot node temperature, and coolant hot node temperature, respectively. and These are the heating power of the winding and the heating power of the core, respectively. For ambient temperature, This refers to the inlet coolant temperature. The thermal resistance for convective heat transfer between the casing and the coolant. This refers to the coolant mass flow rate.

[0025] In this embodiment, considering the impact of coolant flow rate changes on the heat transfer capacity between the casing and the coolant, the flow-related heat transfer resistance satisfies:

[0026] in, for The thermal resistance of convective heat transfer between the engine casing and the coolant. Based on the reference heat transfer thermal resistance, for The coolant mass flow rate at any given time. and This constitutes a flow-dependent heat transfer mapping, making the coolant mass flow rate and inlet coolant temperature explicit as they enter the subsequent coolant, carrying away heat and energy residuals to construct the process.

[0027] Thus, the relationships of the main heat flows are obtained:

[0028]

[0029]

[0030]

[0031] in, for The heat flow conducted between the windings and the stator core at all times. for The heat flow conducted between the stator core and the housing at all times. for The convective heat flow between the engine casing and the coolant. for The convective and radiative heat flow between the machine casing and the external environment.

[0032] Step S2: Collect the operating parameters and measurement parameters of the thermal management of the electric drive system, construct the energy residual based on the lumped parameter thermal network model, and calculate the normalized absolute value of the energy residual.

[0033] The operating parameters are the motor's operating condition parameters, including motor speed, motor output torque, and motor phase current. The measured parameters are physical quantities directly measured by sensors, including coolant mass flow rate, inlet coolant temperature, ambient temperature, and the critical node temperatures corresponding to the thermal state quantities (winding temperature, stator core temperature, casing temperature, and coolant temperature).

[0034] In step S2, for the coolant mass flow sensor, the measurement relationship satisfies:

[0035] in, for The measured output value of the coolant mass flow sensor at all times. for Measurement deviation of the coolant mass flow sensor at all times.

[0036] For the inlet coolant temperature sensor, the measurement relationship satisfies:

[0037] in, for The measured output value of the inlet coolant temperature sensor at any given time. for The inlet coolant temperature at any given time. for Measurement deviation of the inlet coolant temperature sensor at all times.

[0038] Including at least one of constant value deviation and slow drift deviation, similarly, It also includes at least one of constant deviation and slow drift deviation.

[0039] Then, based on the complete energy balance relationship, the energy residual is constructed, expressed as:

[0040]

[0041]

[0042]

[0043]

[0044] in, for Energy residual at any given moment for Monitor the total input heat power at all times. for Constantly monitor the side coolant to remove heat. for Constantly monitor the heat dissipation of the surrounding environment. for The change in the system's heat capacity and energy storage at any given time. and They are respectively Real-time monitoring of the heating power of the side winding and the heating power of the core. for The ambient temperature measurement at that moment.

[0045] The formula for calculating the absolute value of the normalized energy residual is:

[0046]

[0047] in, The normalized absolute value of the energy residual. To normalize the energy residual, the absolute energy residual is divided by the absolute value of the input heat power. This eliminates the influence of the input power magnitude on the residual amplitude under different operating conditions, allowing the residual threshold to be set uniformly across the entire operating range. It is a pre-defined positive number, specifically a very small positive number.

[0048] Step S3: Construct a quasi-steady-state determination index based on the rate of change of thermal state quantities. When the quasi-steady-state determination index meets the preset entry conditions, output a steady-state allow signal. When the absolute value of the normalized energy residual meets the preset trigger conditions, proceed to step S4 to start the online reconstruction and compensation of the target sensor deviation.

[0049] Specifically, step S3 includes: The rate of change of thermal state quantities is defined as:

[0050] in, for The rate of change of thermal state variables at time t; The maximum rate of temperature rise is defined as:

[0051] in, for The maximum rate of temperature rise at any given time, This indicates taking the maximum value; The quasi-steady-state criterion is constructed as follows:

[0052]

[0053] in, express Quasi-steady-state determination index at time t, express Quasi-steady-state determination index at time t, For smoothing coefficients, This is the smoothing time constant.

[0054] It should be noted that before starting the estimator in step S4, the initial values ​​of the deviation estimate, steady-state allowable signal, steady-state judgment index, and duration counter need to be initialized.

[0055] When continuous Each sampling period satisfies When this happens, the steady-state enable signal is output, i.e. , for The steady-state allowable signal value at time t; when When this happens, the steady-state enable signal is revoked. ; The preset first continuous counting threshold, To reach the quasi-steady-state threshold, To exit the quasi-steady-state threshold, and ; When continuous Each sampling period satisfies When the absolute value of the normalized energy residual is determined to meet the preset triggering condition, The second continuous counting threshold is preset. This is the preset normalization threshold.

[0056] when ,and Only if this condition is met will the process proceed to step S4, initiating online reconstruction and compensation of the target sensor bias. Otherwise, the current bias estimate will remain unchanged.

[0057] Thus, step S3 forms a complete activation logic of "initialization + steady state permission + residual threshold trigger + continuous counting".

[0058] Step S4: Construct a variable step size normalized projection recursive deviation estimator driven by residual sensitivity based on the energy residual and residual sensitivity, calculate the residual sensitivity of the energy residual to the target sensor deviation, and then reconstruct the target sensor deviation online to obtain the deviation estimate, and compensate the measured quantity based on the deviation estimate.

[0059] The working process of the variable step-size normalized projection recursive bias estimator is as follows: definition The target sensor measurement at time is The real physical quantity is The compensated measurement quantity is The target sensor deviation is The deviation estimate is Then we have:

[0060]

[0061] To minimize the normalized energy residual after compensation, a target cost function is constructed with the objective of minimizing the absolute value of the normalized energy residual reconstructed based on the compensated measurement. This defines the residual sensitivity, and the expression for the target cost function is:

[0062] in, Let the objective cost function be... It represents the absolute value of the normalized energy residual reconstructed based on the compensated measurement.

[0063] In this embodiment, residual sensitivity is used to determine the decreasing direction of the recursive update of the deviation estimate. Residual sensitivity characterizes the direction and degree of influence of the target sensor deviation on the normalized energy residual, and serves as the core driving information for the recursive update of the deviation estimate in step S4.

[0064] For coolant mass flow sensors, the residual sensitivity is defined as:

[0065] in, for The residual sensitivity of the coolant mass flow sensor at any given time. Indicates partial derivative, for Estimated deviation of coolant mass flow rate measurement at any given time.

[0066] For the inlet coolant temperature sensor, the residual sensitivity is defined as:

[0067] in, for The residual sensitivity of the inlet coolant temperature sensor at any given time. for Estimated deviation of inlet coolant temperature measurement at any given time.

[0068] Define the variable step size adjustment factor as:

[0069] in, for The step size adjustment factor at time intervals, For saturation constraint functions, and These are the minimum and maximum values ​​of the variable step size adjustment factor, respectively. Based on step size, and This is the step size adjustment coefficient. for The residual sensitivity of the target sensor at any given time characterizes the negative value of the change in normalized energy residual caused by a unit deviation change. Its sign determines the update direction of the deviation estimate, and its magnitude determines the degree of influence of the unit deviation change on the residual.

[0070] Through the above variable step size mechanism, the update step size can be automatically increased to accelerate convergence when the deviation is large, and the update step size can be automatically decreased to suppress oscillation and overshoot when the deviation gradually decreases.

[0071] Set the residual sensitivity normalization term, with the expression:

[0072] in, for The residual sensitivity normalization term at time step is used to suppress the amplification or reduction effect of changes in the residual sensitivity amplitude under different operating conditions on the error estimation update. It is a preset positive number.

[0073] Set the residual dead zone function, the expression is:

[0074] in, For residual dead zone function, This is the preset dead zone threshold.

[0075] The residual dead zone function is used to reduce the impact of noise disturbances in small residual intervals on recursive updates.

[0076] Set the projection constraint operator, with the expression:

[0077] in, For projection constraint operators, As an intermediate variable, and These represent the lower and upper bounds of the physically feasible range of the target sensor deviation, respectively.

[0078] The projection constraint operator limits the bias estimate to a physically feasible range. This prevents the deviation estimate from diverging or exceeding the actual possible deviation range of the sensor under abnormal disturbances, model mismatch, or noise.

[0079] The bias estimate is calculated using the update law of the variable step-size normalized projection recursive bias estimator, expressed as:

[0080] in, for The estimated deviation at time.

[0081] It should be noted that when the target sensors include a coolant mass flow sensor and an inlet coolant temperature sensor, the parameters of the basic step size, upper limit of step size, step size adjustment coefficient, residual dead zone threshold and projection constraint range corresponding to different sensors need to be set separately in order to achieve differentiated adaptive matching for the dynamic characteristics of different physical parameters.

[0082] It should be further explained that the basic step size, lower step size limit, upper step size limit, residual dead zone threshold, quasi-steady-state entry threshold, quasi-steady-state exit threshold, algorithm trigger threshold, and projection constraint range are set based on the sampling period, thermal network model parameter calibration results, measurement noise level, and the physical feasibility range of the target sensor. Specifically, the quasi-steady-state entry threshold and quasi-steady-state exit threshold are preferably set to different values ​​with hysteresis to avoid frequent switching of the steady-state allowable signal near the threshold. The projection constraint range is preferably set according to the allowable deviation range of the coolant mass flow sensor and the inlet coolant temperature sensor in the actual system. The residual dead zone threshold is preferably set according to the noise level or the normalized residual fluctuation range under normal operating conditions to reduce noise interference on the recursive update of the deviation.

[0083] In a preferred embodiment, since the target sensor includes both a coolant mass flow sensor and an inlet coolant temperature sensor, sequential compensation or time-sharing recursive update methods can be used to reduce the coupling effect between the two deviation variables. Specifically, within the same monitoring window, the coolant mass flow sensor deviation can be updated first based on the current inlet coolant temperature measurement value or its compensation value to obtain a coolant mass flow compensation value, and then the inlet coolant temperature sensor deviation can be updated based on the updated coolant mass flow compensation value; or, recursive updates of the two types of target sensor deviations can be performed alternately in adjacent sampling windows. The above-mentioned sequential compensation or time-sharing recursive update methods are beneficial for improving the distinguishability, estimation stability, and compensation reliability when dual sensor deviations coexist.

[0084] In this embodiment, during the process of compensating for the measured quantity based on the deviation estimate, the following equation is satisfied:

[0085]

[0086] in, for Coolant mass flow rate after real-time compensation. for The inlet coolant temperature after real-time compensation.

[0087] After compensation and Used respectively to replace the original biased and This serves as feedback input to the thermal management controller of the electric drive system, generating updated actuator control commands to maintain the stability of the thermal management control closed loop even when deviations exist. The actuator control commands include adjustments to at least one of the following: electric water pump speed, cooling fan duty cycle, and cooling valve opening, thereby reducing the accumulation of control errors caused by sensor bias.

[0088] Step S5: Based on the compensated measurement, reconstruct the normalized energy residual absolute value, compare the reconstructed normalized energy residual absolute value with the preset alarm threshold and the preset release threshold respectively, and output the fault warning result in combination with the continuous counting judgment mechanism; at the same time, use the compensated measurement to replace the original measurement with deviation as the feedback input of the electric drive system thermal management controller.

[0089] In the process of reconstructing the normalized absolute value of the energy residual based on the compensated measurement, the following equation is satisfied:

[0090]

[0091]

[0092] in, To reconstruct the heat removal of the coolant based on the compensated measurement, The normalized energy residual is obtained by reconstructing the measured values ​​after compensation. This represents the absolute value of the normalized energy residual after reconstruction.

[0093] The specific mechanism for determining continuous counting is as follows: when ,and continuous All sampling periods satisfy: At that time, a fault warning signal is output, that is , for The fault warning signal value at any given time; when and continuous Each sampling period satisfies At that time, cancel the fault warning signal. , for The fault warning signal value at any given time.

[0094] in, This is the alarm threshold. To remove the threshold, and , and These are the continuous counting thresholds for entering and clearing the alarm, respectively.

[0095] The continuous counting determination mechanism also outputs the counter status. , and when At the same time, maintain the current warning status. Do not update, and maintain or reset the counter value. ;when At that time, according to and , The comparison result is related to the value of the counter. Accumulation or clearing is performed to achieve fault early warning judgment with hysteresis and continuous counting constraints.

[0096] To verify the technical effectiveness of the method of the present invention, a basic compensation method is set as a comparative method. The basic compensation method adopts the same lumped parameter thermal network model, energy residual construction method, quasi-steady-state determination logic and algorithm triggering conditions as the present invention, but its deviation estimation update adopts a fixed step size recursive method, without adaptively adjusting the update step size according to the residual sensitivity direction, and without setting differentiated parameter matching for different target sensors.

[0097] In this embodiment, the deviation estimation update law of the basic compensation method is expressed as: when and continuous Each sampling period satisfies At that time, the following conditions are met:

[0098] in, To maintain a fixed step size, for The deviation estimate of the time-based compensation method. for The deviation estimate of the time-based compensation method. This is the normalized energy residual obtained by reconstructing the measured quantity after compensation under the basic compensation method.

[0099] Please see Figure 2 At t=600s, a negative measurement deviation fault was injected into the coolant mass flow sensor to verify the online reconstruction and compensation performance of the method of the present invention for coolant mass flow sensor deviation. Figure 2 It can be seen that after a fault occurs, the measured fault value deviates significantly from the true value. Although the basic compensation method can gradually reduce the impact of the deviation, the convergence speed is slow and there are obvious residual oscillations during the convergence process. In contrast, the method of this invention can pull the compensation result back to the vicinity of the true value more quickly after the fault is injected, and maintain a smaller fluctuation amplitude and a more stable convergence process after approaching the true value, thereby effectively reducing the level of normalized residual after compensation and improving the accuracy and stability of subsequent early warning judgment.

[0100] Please see Figure 3 At t=600s, a positive measurement deviation fault was injected into the inlet coolant temperature sensor to verify the online reconstruction and compensation performance of the method of the present invention for the inlet coolant temperature sensor deviation. Figure 3 It can be seen that after the fault occurs, the measured value of the fault is generally higher than the true value. The basic compensation method can achieve deviation correction to a certain extent, but the compensation response is slow and there is still a significant residual deviation for a long time. In contrast, the method of the present invention can complete the online reconstruction of the inlet coolant temperature sensor deviation more quickly, and make the compensated result approach the true value more quickly. Moreover, the compensation process is more stable, which helps to reduce the risk of false alarms and missed alarms caused by temperature sensor deviation, and improve the reliability of thermal management early warning and control of electric drive system.

[0101] In summary, the electric drive thermal management deviation compensation and early warning method based on energy residuals according to the above embodiments has the following beneficial effects: 1. This invention establishes a lumped-parameter thermal network model for explicit mapping of target sensor bias, enabling coolant mass flow rate and inlet coolant temperature to explicitly enter the energy balance residual construction process, thus achieving online identification and reconstruction of target sensor bias under the constraints of thermal management mechanisms.

[0102] 2. This invention proposes a residual sensitivity-driven variable step-size normalized projection recursive bias estimator. Through the synergistic effect of residual sensitivity, variable step-size adjustment factor, residual sensitivity normalization term, residual dead zone function and projection constraint operator, it achieves rapid convergence in the stage of large bias and smooth approximation in the stage of small bias, thereby simultaneously improving compensation speed, compensation accuracy and compensation stability.

[0103] 3. This invention designs a reliable activation logic of "quasi-steady-state gating + residual triggering + continuous counting", and realizes fault early warning output and control feedback based on the compensated measurement, thereby reducing the risk of false updates, false alarms and missed alarms under dynamic operating conditions, and improving the robustness of thermal management and closed-loop control accuracy of electric drive system.

[0104] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for early warning compensation of electric drive thermal management deviation based on energy residual, characterized in that, include: Step S1: Establish a lumped parameter thermal network model for the thermal management of the electric drive system oriented to explicit mapping of target sensor deviation, so as to introduce the target sensor measurement into the energy balance relationship of the thermal management of the electric drive system, and obtain the thermal state quantity characterizing the heat transfer relationship of the thermal management of the electric drive system. The target sensor includes a coolant mass flow sensor and an inlet coolant temperature sensor. Step S2: Collect the operating parameters and measurement parameters of the thermal management of the electric drive system, construct the energy residual based on the lumped parameter thermal network model, and calculate the normalized absolute value of the energy residual; Step S3: Construct a quasi-steady-state determination index based on the rate of change of thermal state quantities. When the quasi-steady-state determination index meets the preset entry conditions, output a steady-state allow signal. When the absolute value of the normalized energy residual meets the preset trigger conditions, proceed to step S4 to start the online reconstruction and compensation of the target sensor deviation. Step S4: Construct a variable step size normalized projection recursive deviation estimator driven by residual sensitivity based on the energy residual and residual sensitivity, calculate the residual sensitivity of the energy residual to the target sensor deviation, and then reconstruct the target sensor deviation online to obtain the deviation estimate, and compensate the measured quantity based on the deviation estimate. Step S5: Based on the compensated measurement, reconstruct the normalized energy residual absolute value, compare the reconstructed normalized energy residual absolute value with the preset alarm threshold and the preset release threshold respectively, and output the fault warning result in combination with the continuous counting judgment mechanism; at the same time, use the compensated measurement to replace the original measurement with deviation as the feedback input of the electric drive system thermal management controller.

2. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 1, characterized in that, Step S1 specifically includes: Define the thermal state variables as: in, For discrete sampling time indexes, for Thermal state parameters at any given time for The temperature of the winding hot node at any given time. for The stator core hot junction temperature at any given moment. for The temperature of the casing thermal node at any given time. for The temperature of the coolant at the hot junction at any given time. Indicates transpose; Define parameter vector for: in, , , and These are the equivalent heat capacities of the windings, stator core, housing, and coolant hot nodes, respectively. , , and These are the contact thermal resistance between the winding and the stator core, the contact thermal resistance between the stator core and the housing, the convective and radiative thermal resistance between the housing and the external environment, and the reference heat transfer thermal resistance, respectively. As the baseline mass flow rate, For traffic-related indices, This refers to the specific heat capacity of the coolant. Construct a lumped-parameter thermal network model, expressed as: in, To represent the differential, Indicates the continuous sampling time. , The sampling period is , , and These are the winding hot node temperature, stator core hot node temperature, housing hot node temperature, and coolant hot node temperature, respectively. and These are the heating power of the winding and the heating power of the core, respectively. For ambient temperature, This refers to the inlet coolant temperature. The thermal resistance for convective heat transfer between the casing and the coolant. This refers to the coolant mass flow rate; Furthermore, the flow-related heat transfer thermal resistance satisfies: in, for The thermal resistance of convective heat transfer between the engine casing and the coolant. Based on the reference heat transfer thermal resistance, for Coolant mass flow rate at any given time; The corresponding heat flow satisfies: in, for The heat flow conducted between the windings and the stator core at all times. for The heat flow conducted between the stator core and the housing at all times. for The convective heat flow between the engine casing and the coolant. for The convective and radiative heat flow between the machine casing and the external environment.

3. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 2, characterized in that, In step S2, for the coolant mass flow sensor, the measurement relationship satisfies: in, for The measured output value of the coolant mass flow sensor at all times. for Measurement deviation of the coolant mass flow sensor at all times; For the inlet coolant temperature sensor, the measurement relationship satisfies: in, for The measured output value of the inlet coolant temperature sensor at any given time. for The inlet coolant temperature at any given time. for Measurement deviation of the inlet coolant temperature sensor at all times; Then, the energy residual is constructed, expressed as: in, for Energy residual at any given moment for Monitor the total input heat power at all times. for Constantly monitor the side coolant to remove heat. for Constantly monitor the heat dissipation of the surrounding environment. for The change in the system's heat capacity and energy storage at any given time. and They are respectively Real-time monitoring of the heating power of the side winding and the heating power of the core. for The ambient temperature measurement at that moment.

4. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 3, characterized in that, In step S2, the formula for calculating the absolute value of the normalized energy residual is: in, The normalized absolute value of the energy residual. To normalize the energy residual, It is a preset positive number.

5. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 4, characterized in that, Step S3 specifically includes: The rate of change of thermal state quantities is defined as: in, for The rate of change of thermal state variables at time t; The maximum rate of temperature rise is defined as: in, for The maximum rate of temperature rise at any given time, This indicates taking the maximum value; The quasi-steady-state criterion is constructed as follows: in, express Quasi-steady-state determination index at time t, express Quasi-steady-state determination index at time t, For smoothing coefficients, The smoothing time constant; When continuous Each sampling period satisfies When, output steady-state enable signal; when When this happens, the steady-state enable signal is revoked; The preset first continuous counting threshold, To reach the quasi-steady-state threshold, To exit the quasi-steady-state threshold, and ; When continuous Each sampling period satisfies When the absolute value of the normalized energy residual is determined to meet the preset triggering condition, The second continuous counting threshold is preset. This is the preset normalization threshold.

6. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 5, characterized in that, In step S4, the working process of the variable step size normalized projection recursive bias estimator is as follows: definition The target sensor measurement at time is The real physical quantity is The compensated measurement quantity is The target sensor deviation is The deviation estimate is Then we have: With the objective of minimizing the absolute value of the normalized energy residual reconstructed based on the compensated measurement, a target cost function is constructed, thereby defining the residual sensitivity. The expression for the target cost function is: in, Let the objective cost function be... The normalized absolute value of the energy residual reconstructed based on the compensated measurement; For coolant mass flow sensors, the residual sensitivity is defined as: in, for The residual sensitivity of the coolant mass flow sensor at any given time. Indicates partial derivative, for Estimated deviation of coolant mass flow rate measurement at any given time; For the inlet coolant temperature sensor, the residual sensitivity is defined as: in, for The residual sensitivity of the inlet coolant temperature sensor at any given time. for Estimated deviation of inlet coolant temperature measurement at any given time; Define the variable step size adjustment factor as: in, for The step size adjustment factor at time intervals, For saturation constraint functions, and These are the minimum and maximum values ​​of the variable step size adjustment factor, respectively. Based on step size, and This is the step size adjustment coefficient. for The residual sensitivity of the target sensor at any given time; Set the residual sensitivity normalization term, with the expression: in, for The residual sensitivity normalization term at time step, The preset positive number; Set the residual dead zone function, the expression is: in, For residual dead zone function, The preset dead zone threshold; Set the projection constraint operator, with the expression: in, For projection constraint operators, As an intermediate variable, and These represent the lower and upper bounds of the physically feasible range of the target sensor deviation, respectively. The bias estimate is calculated using the update law of the variable step-size normalized projection recursive bias estimator, expressed as: in, for The estimated deviation at time.

7. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 6, characterized in that, In step S4, during the process of compensating for the measured quantity based on the deviation estimate, the following equation is satisfied: in, for Coolant mass flow rate after real-time compensation. for The inlet coolant temperature after real-time compensation.

8. The method for early warning compensation of electric drive thermal management deviation based on energy residual as described in claim 7, characterized in that, In step S5, during the process of reconstructing the normalized absolute value of the energy residual based on the compensated measurement, the following equation is satisfied: in, To reconstruct the heat removal of the coolant based on the compensated measurement, The normalized energy residual is obtained by reconstructing the measured values ​​after compensation. This represents the absolute value of the normalized energy residual after reconstruction.

Images