Fuel cell automobile thermal management system and construction method thereof

Abstract

The invention discloses a fuel cell vehicle thermal management system and a construction method thereof, and the fuel cell vehicle thermal management system comprises a thermodynamic modeling module, a double-layer fuzzy sliding mode controller, an actuator cooperative control module, a working condition adaptive module and a fault protection module. According to the thermal management system of the fuel cell vehicle and the construction method of the thermal management system, the dynamic thermodynamic model is established, and the double-layer fuzzy sliding mode controller and the actuator cooperative control module are designed, so that millisecond-level high-precision tracking of the temperature of the electric pile and online real-time compensation of unknown disturbance are realized; meanwhile, stable operation of the system in the whole working condition domain range can be ensured through the working condition self-adaption module and the fault protection module, and the system has the advantages of being high in control precision, high in anti-interference capacity and good in system reliability.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for new energy vehicles, specifically to a thermal management system for fuel cell vehicles and its construction method. Background Technology

[0002] The thermal management system of a fuel cell vehicle is a core subsystem ensuring the efficient and safe operation of the fuel cell stack. The stack generates a significant amount of heat during operation; excessively high temperatures can lead to problems such as proton exchange membrane dehydration and reduced catalyst activity, while excessively low temperatures can affect electrochemical reaction efficiency and even cause difficulties in cold starts. Therefore, precise thermal management is crucial for improving fuel cell system efficiency and extending its lifespan.

[0003] In existing technologies, fuel cell thermal management systems mostly employ traditional PID control or fixed-gain sliding mode control methods. However, these methods have the following significant drawbacks:

[0004] 1. Insufficient dynamic response: PID control relies on a linear model, which makes it difficult to adapt to nonlinear operating conditions such as sudden changes in the heat generation rate of the fuel cell stack and fluctuations in ambient temperature, resulting in lag in temperature regulation and frequent overshoot. For example, under conditions of rapid vehicle acceleration or high load, the heat generation of the fuel cell stack increases sharply. Traditional control algorithms, due to their fixed gain, cannot quickly adjust the output of the heat dissipation components, which can easily lead to local overheating.

[0005] 2. Lack of energy consumption optimization: Existing methods often focus on single temperature tracking, neglecting the energy consumption optimization of actuators. For example, PTC heaters operate at maximum power during low-temperature startup, resulting in wasted energy; fan speed regulation is not dynamically matched with heat dissipation efficiency, causing redundant power consumption.

[0006] 3. Disconnect between model and measured data: Most studies rely on simulation models to design controllers, but the model parameters (such as the relationship between coolant flow and temperature, and pipeline resistance coefficient) are not fully calibrated with measured data, resulting in a large deviation between the simulation results and the actual system, which affects the reliability of the control strategy.

[0007] 4. Complex multi-component coordinated control: The thermal management system involves multiple actuators such as PTC heaters, fans, water pumps, and three-way valves. Existing technologies lack a unified coordinated control framework. Conflicts can easily arise when each subsystem is adjusted independently. For example, an increase in water pump flow may exacerbate uneven distribution in the three-way valve branches, leading to uneven temperature distribution in the fuel cell stack.

[0008] Solving these problems is now a top priority. Summary of the Invention

[0009] In view of this, the present invention provides a thermal management system for fuel cell vehicles and a method for constructing the same.

[0010] The technical solution is as follows:

[0011] The first aspect of this application relates to a method for constructing a thermal management system for a fuel cell vehicle, comprising the following steps:

[0012] S1. Use the stack heat generation module, coolant circulation module and heat dissipation module in Matlab-Simulink to establish a dynamic thermodynamic model of the fuel cell stack;

[0013] S2. Design a two-layer fuzzy sliding mode controller that includes an upper-layer controller and a lower-layer controller. The upper-layer controller is used to adjust the disturbance estimate online through fuzzy rules and generate a compensation term that includes environmental heat loss and model uncertainty. The lower-layer controller is used to design a control law based on the sliding mode surface function and combine the approach rate coefficient to suppress chattering.

[0014] S3. Embed the execution control law output by the lower-level controller into the actuator collaborative control module of PTC heater, fan, high-pressure electronic water pump and three-way valve, and realize the linkage control of PTC heater, fan, high-pressure electronic water pump and three-way valve through Simulink signal line;

[0015] S4. Based on measured water temperature data, the least squares method is used to establish a quantitative relationship between coolant flow rate and temperature, and the model parameters are calibrated accordingly.

[0016] The second aspect of this application relates to a thermal management system for a fuel cell vehicle constructed using the above-described method, characterized in that it comprises:

[0017] Thermodynamic modeling module, which is used to calculate the heat generated by the fuel cell stack and the temperature change of the coolant in real time;

[0018] A two-layer fuzzy sliding mode controller, which generates disturbance estimates through an upper-layer controller and outputs actuator control laws through a lower-layer controller;

[0019] The actuator coordination control module is used to control the PTC heater, the fan, the high-pressure electronic water pump, and the three-way valve.

[0020] The adaptive operating condition module is used to dynamically adjust control parameters based on vehicle speed and ambient temperature.

[0021] The fault protection module is used to switch to redundant control mode in case of abnormal temperature.

[0022] The system acquires data at 100Hz via CAN bus, with an embedded controller control cycle of ≤10ms and an overall response time of ≤50ms.

[0023] The above-mentioned thermal management system for fuel cell vehicles and its construction method achieve millisecond-level high-precision tracking of stack temperature and online real-time compensation for unknown disturbances by establishing a dynamic thermodynamic model, designing a dual-layer fuzzy sliding mode controller and an actuator collaborative control module. At the same time, the system can be stably operated in the entire operating condition range through the operating condition adaptive module and the fault protection module. It has the advantages of high control accuracy, strong anti-disturbance capability and good system reliability. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the thermal management system for a fuel cell vehicle.

[0025] Figure 2 Figure showing the simulation results of fuzzy sliding membrane control using Matlab - inlet temperature.

[0026] Figure 3 Figure showing the simulation results of fuzzy sliding membrane control at the outlet temperature using Matlab.

[0027] Figure 4 The image shows the simulation results of fuzzy sliding diaphragm control for fan speed in Matlab.

[0028] Figure 5 Figure showing the simulation results of fuzzy sliding membrane control using Matlab with inlet and outlet temperature difference.

[0029] Figure 6 Figure showing the simulation results of fuzzy sliding membrane control using Matlab - cooling water flow rate;

[0030] Figure 7 The simulation results of the opening of a three-way valve in Matlab for fuzzy sliding diaphragm control are shown in the figure. Detailed Implementation

[0031] The present invention will be further described below with reference to the embodiments and accompanying drawings.

[0032] Example 1:

[0033] A method for constructing a thermal management system for a fuel cell vehicle, comprising the following steps:

[0034] S1. Establish a dynamic thermodynamic model of the fuel cell stack.

[0035] A dynamic thermodynamic model of a fuel cell stack was established using the stack heat generation module, coolant circulation module, and heat dissipation module in Matlab-Simulink.

[0036] Specifically, in step S1, the fuel cell stack heat generation module calculates the real-time heat generation based on the output power and efficiency coefficient, and the coolant circulation module predicts the temperature change trend through a one-dimensional fluid dynamics equation.

[0037] The fuel cell stack heat generation module employs a heat loss calculation model. This model calculates real-time heat generation based on electrochemical reaction power and heat loss, satisfying the following formula:

[0038] Q gen =P elec ·(1-η)-Q loss ;

[0039] In the above formula, Q gen P represents the real-time heat generation of the fuel cell stack (in W). elec The output power of the fuel cell stack is represented by η (in W), where η represents the electrochemical efficiency coefficient, and Q represents the output power of the fuel cell stack. loss This indicates environmental heat loss (in W).

[0040] Among them, the environmental heat loss Q loss The calculation model is as follows:

[0041] Q loss =h·A·(T) stack -T ambient );

[0042] In the above formula, h represents the thermal conductivity coefficient (unit: W / (m²)). 2 ·K)), A represents the surface area of ​​the fuel cell stack (in m²). 2 ), T stack T represents the surface temperature of the fuel cell stack (in K). ambient Indicates ambient temperature (unit: K).

[0043] The thermal conductivity coefficient h refers to the heat transfer capacity between the fuel cell stack surface and the environment. It can be determined using material thermophysical property test data or empirical formulas, such as a comprehensive calculation based on the thermal conductivity and convective heat transfer coefficient of the fuel cell stack shell material. This accurately characterizes the heat exchange process between the fuel cell stack and the environment. The fuel cell stack surface area A refers to the effective heat dissipation area in contact with the external environment. It can be measured using 3D modeling software or calculated based on the fuel cell stack's geometry, and is used to quantify its heat dissipation capacity.

[0044] The coolant circulation module adopts a one-dimensional fluid dynamics model, and its state equation is:

[0045]

[0046] In the above formula, dT represents the change in coolant temperature (in K), dt represents the time derivative (in seconds), and Q... fan Q represents the power dissipated by the radiator fan (in watts). PTC ρ represents the heating power of the PTC heater (in W). cool This indicates the density of the coolant (unit: kg / m³). 3 V sysRepresents system volume (unit: m³) 3 ), c p This indicates the specific heat capacity of the coolant (unit: J / (kg·K));

[0047] Coolant density ρ cool and specific heat capacity c p This refers to the physical properties of the coolant. Specifically, it can be derived from measured data or selected from standard physical properties based on the coolant type. For example, the density and specific heat capacity of an ethylene glycol aqueous solution as a function of temperature are used to construct accurate fluid thermodynamic equations. System volume V sys This refers to the total volume of the coolant circulation pipeline, which can be calculated from the pipeline diameter and length. For example, the sum of the volumes of each branch can be calculated using the segmented accumulation method to establish a dynamic model of coolant temperature changes.

[0048] The expression for the heat dissipation module is:

[0049] Q fan =k eff ·(T cool_out -T ambient )·A radiator ;

[0050] In the above formula, k eff This represents the overall heat dissipation coefficient (unit: W / (m²)). 2 ·K)), T cool_out Indicates the radiator outlet coolant temperature (in K), A radiator This indicates the effective heat dissipation area of ​​the radiator (in m²). 2 ).

[0051] The heat loss calculation model dynamically calculates the heat lost from the fuel cell stack to the environment by combining the surface heat transfer coefficient, effective heat dissipation area, and real-time temperature difference. For example, when the temperature difference between the fuel cell stack and the environment increases, Q... loss The calculation results are updated synchronously, thus accurately reflecting heat dissipation changes in the thermodynamic model. The one-dimensional fluid dynamics model of the coolant circulation module converts the combined effects of heat generation from the fuel cell stack, fan cooling power, and PTC heating power into the rate of change of coolant temperature through the heat power balance equation. For example, in the Simulink platform, this model can calculate the trend of coolant temperature change over time in real time based on the current coolant physical parameters and system volume, providing the controller with an accurate model of the controlled object.

[0052] Compared to existing technologies, traditional methods typically employ fixed heat loss coefficients or simplified static heat balance equations, such as assuming a constant thermal conductivity coefficient, leading to a decrease in model accuracy when ambient temperature changes drastically. In contrast, this approach, through a dynamic heat loss model and one-dimensional fluid dynamics equations, can correct the thermal conductivity coefficient online and consider changes in coolant properties. For example, it automatically adjusts the h value in low-temperature environments to reflect the enhanced heat dissipation effect caused by increased air density, thereby improving the model's adaptability to actual operating conditions.

[0053] S2. Design a two-layer fuzzy sliding mode controller.

[0054] The two-layer fuzzy sliding mode controller consists of an upper controller and a lower controller. The upper controller is used to adjust the disturbance estimate online through fuzzy rules and generate a compensation term that includes environmental heat loss and model uncertainty. The lower controller is used to design a control law based on the sliding surface function and combine the approach rate coefficient to suppress chattering.

[0055] Specifically, in step S2, the upper-level controller generates an interference estimate. The expression is:

[0056]

[0057] In the above formula, λ represents a constant compensation term greater than 0, and k1(t) represents an adaptive adjustment term, the expression of which is: in, G represents the absolute value of the derivative of the Lyapunov function, G represents the proportional adjustment coefficient (which is a preset constant), G′ represents the compensation intensity coefficient (which is a preset constant), t represents time (in seconds), Δk represents the fuzzy adjustment amount of the sliding mode control gain, and Δη represents the fuzzy adjustment amount of the electrochemical efficiency compensation coefficient. Both Δk and Δη are adjusted online through fuzzy rules.

[0058] The above fuzzy rules include:

[0059] The input variable is the sliding surface derivative, and the fuzzy set is divided into negative large, negative medium, zero, positive medium, and positive large;

[0060] The fuzzy rule for the output variable Δk is as follows: when the derivative of the sliding surface is negative (large) or negative (medium), the output is negative (large); when the derivative of the sliding surface is zero, the output is zero; when the derivative of the sliding surface is positive (medium) or positive (large), the output is positive (large).

[0061] The fuzzy rules for the output variable Δη are as follows: when the derivative of the sliding surface is negative, the output is negative; when the derivative of the sliding surface is negative, the output is negative; when the derivative of the sliding surface is zero, the output is zero; when the derivative of the sliding surface is positive, the output is positive; when the derivative of the sliding surface is positive, the output is positive.

[0062] The membership function adopts a triangular distribution, and the universe of discourse is normalized to [-1, 1].

[0063] The sliding surface derivative refers to the time derivative of the sliding surface function, which can be achieved by differentiating the sliding surface function. It characterizes the rate at which the system state deviates from the sliding surface. Fuzzy set partitioning classifies input variables according to preset levels, such as five linguistic variable levels, by setting corresponding membership function ranges. This converts continuous variables into discrete semantic descriptions required for fuzzy inference. The fuzzy rules for output variables are strategies that map the semantic levels of input variables to output adjustment quantities, for example, using conditional statements. This is used to dynamically adjust control parameters in real time. The triangular distribution membership function defines the degree of membership of each fuzzy set using linearly ascending and descending triangular curves. This can be achieved by setting the vertex position and base width, ensuring computational efficiency while achieving smooth fuzzy inference.

[0064] Specifically, the sliding surface derivative, as the input variable of the fuzzy rules, is divided into five fuzzy levels and corresponding membership functions are established, which can dynamically convert continuously changing control errors into semantic variables. When the sliding surface derivative is negatively large or negatively medium, the fuzzy rules for the output variable Δk set the adjustment amount to negatively large to enhance the system's ability to suppress deviations from the sliding surface; when the sliding surface derivative is zero, the adjustment amount is set to zero to maintain the current state; when the sliding surface derivative is positively medium or positively large, the adjustment amount is set to positively large to accelerate the approach to the target. The output variable Δη adopts symmetrical fuzzy rules to ensure that the parameter adjustment amount strictly corresponds to the degree of system deviation. The design of the triangular membership function makes there an overlapping area between adjacent fuzzy sets, avoiding abrupt changes in control parameters, while the normalization process ensures the comparability of variables with different physical dimensions.

[0065] Compared to existing technologies, traditional fuzzy control often uses a fixed threshold to divide the fuzzy set, making it difficult to adapt to the nonlinear dynamic characteristics of sliding mode control. This scheme, however, uses the sliding surface derivative as the core input variable, combined with a hierarchical adjustment strategy, to more accurately match the parameter requirements of the system under different deviation states. Existing single-output adjustment rules often fail to distinguish the differences in sensitivity of parameter adjustments. This scheme designs independent rules for the two output variables, achieving hierarchical and fine-tuning of control gain and compensation terms, effectively balancing response speed and stability.

[0066] The execution control law u(t) of the lower-level controller is as follows:

[0067]

[0068] In the above formula, ε and ρ both represent the approach rate coefficients greater than 0, and s represents the sliding surface function, which is expressed as s = e + αe + βe. p / qp and q both represent the coefficients of the sliding surface exponent (used to adjust the convergence speed), both p and q are positive odd numbers, and p > q. e represents the temperature error, which is expressed as e = T. target -T actual T target T represents the target temperature (in K). actual Let represent the actual temperature (in K), α represent the error proportional coefficient, β represent the exponential term weighting coefficient (used to enhance nonlinear regulation capability), f(T) represent the nonlinear term of the thermodynamic model, and sgn(s) represent the sign function (used to implement the discontinuity of sliding mode control). The second derivative represents the target temperature. This indicates that the upper-level controller generates an estimated disturbance value. This represents the rate of change of temperature error (in K / s).

[0069] S3. Embed the execution control law output by the lower-level controller into the actuator collaborative control module of PTC heater, fan, high-pressure electronic water pump and three-way valve, and realize the linkage control of PTC heater, fan, high-pressure electronic water pump and three-way valve through Simulink signal line.

[0070] In this embodiment, the control of the PTC heater in the actuator coordination control module includes:

[0071] Based on temperature deviation ΔT=T actual -T target and the rate of change of temperature deviation The fuzzy set is divided into negative large, negative medium, zero, positive medium, and positive large;

[0072] The objective function J for sliding mode control is designed as follows:

[0073]

[0074] In the above formula, u PTC This indicates the control input voltage of the PTC heater (unit: V, adjustable range: 0-48V), t end This represents the end time of the control cycle (in seconds), and 0.05 is the weighting coefficient, which was determined through offline optimization experiments.

[0075] The objective function J includes the integral of the square term of temperature deviation and the square term of control input voltage of PTC heater. The weighting coefficients are determined through offline optimization experiments. The control input voltage adjustment range of PTC heater is set to 0-48V.

[0076] Temperature deviation refers to the difference between the actual temperature and the target temperature. It can be calculated by comparing the measured value from the temperature sensor with the controller's set value, and is used to quantify the current system's temperature control error. The rate of change of temperature deviation refers to the speed at which the temperature deviation changes over time. It can be calculated using differential operations or the sliding window interpolation method, and is used to reflect the dynamic trend of temperature change. Fuzzy set partitioning refers to quantifying the continuous temperature deviation and its rate of change into discrete fuzzy linguistic variables. This can be achieved using triangular or trapezoidal membership functions, describing the dynamic characteristics of the system state through these linguistic variables. The sliding mode control objective function is an integral function used to evaluate control performance. It can be implemented using a weighted combination of temperature tracking accuracy and control energy consumption, where the squared term of the temperature deviation reflects the tracking accuracy requirement, and the squared term of the control input voltage reflects the energy consumption constraint. Weighting coefficients are parameters indicating the relative importance of different optimization objectives in the objective function. They can be determined through multiple sets of comparative experiments combined with least squares fitting, and are used to balance the relationship between control accuracy and actuator losses. The control input voltage adjustment range refers to the voltage operating range that the PTC heater is allowed to apply. Specifically, it can be set according to the heater's rated power and system safety requirements, for example, by using a pulse width modulation circuit to achieve continuous adjustment from 0 to 48V.

[0077] Specifically, the control process of the PTC heater involves real-time acquisition of fuel cell stack temperature data, calculating the temperature deviation and its rate of change as input variables for the fuzzy inference system. After fuzzification, the input variables are used for inference based on a pre-defined fuzzy rule base to generate corresponding sliding mode control parameters. The sliding mode controller dynamically adjusts the PTC heater voltage output value according to the objective function requirements, ensuring rapid convergence of the temperature deviation while suppressing drastic fluctuations in the control quantity. During this process, the weighting coefficients of the objective function are determined through offline optimization to ensure an optimal trade-off between temperature control accuracy and energy consumption under different operating conditions. For example, when the temperature deviation is large, the controller prioritizes increasing the heating power for a rapid response; when the temperature approaches the target value, it reduces the amplitude of the control quantity change to avoid overshoot.

[0078] Compared to existing technologies, traditional PTC heater control methods often employ fixed-threshold switching control or linear proportional regulation, which struggle to cope with rapid fluctuations in stack temperature. This solution integrates fuzzy inference and sliding mode control, effectively suppressing control chattering while maintaining rapid response capabilities. Furthermore, an energy consumption constraint term is introduced into the objective function, overcoming the limitation of conventional sliding mode control in neglecting actuator losses, thus achieving synergistic optimization of control performance and equipment lifespan.

[0079] In this embodiment, the fan control in the actuator coordination control module includes:

[0080] Construct a mapping model between fan speed and heat dissipation efficiency:

[0081]

[0082] In the above formula, Q fan This represents heat dissipation (in W), n fan ΔT represents the fan speed (in rpm). radiator This indicates the temperature difference of the radiator (in °C).

[0083] The fan speed is dynamically adjusted based on a dual-layer fuzzy sliding mode controller. The fuzzy rules are matched according to the load conditions: when the temperature deviation ΔT < 5℃, the fan speed increase is limited to no more than 10%; when the temperature deviation ΔT ≥ 10℃, the maximum heat dissipation power is enabled.

[0084] The mapping model between fan speed and heat dissipation efficiency is a cubic polynomial function established through experimental data. Specifically, it uses data collected from speed and temperature sensors for fitting, quantifying the heat dissipation capacity at different speeds. This model accurately reflects the nonlinear impact of fan speed changes on heat dissipation efficiency, providing a basis for dynamic adjustment. The dual-layer fuzzy sliding mode controller dynamically adjusts the speed by combining the fast response characteristics of sliding mode control with the adaptability of fuzzy logic. Specifically, it uses a fuzzy rule base to adjust control parameters online, enabling flexible responses to complex operating conditions. This controller matches load changes using fuzzy rules, ensuring different control strategies for different temperature deviations. Limiting the increase in fan speed refers to using a conservative adjustment strategy when the temperature deviation is small. This can be achieved by setting a threshold for the rate of change of speed, avoiding system oscillations caused by frequent and large adjustments. Activating maximum heat dissipation power means activating full-power fan operation when the temperature deviation exceeds the safe range. This can be achieved through relays or power module control circuits, ensuring rapid heat dissipation in emergency situations.

[0085] Specifically, the mapping model between fan speed and heat dissipation efficiency is obtained through fitting experimental data. Based on this model, the theoretical heat dissipation under current operating conditions can be calculated in real time. A dual-layer fuzzy sliding mode controller generates speed adjustment commands based on temperature deviation and its rate of change. The fuzzy rule base divides the temperature deviation into multiple levels, each corresponding to a different control strategy. When the temperature deviation is small, the control strategy limits the increase in fan speed, avoiding overshoot through gradual adjustment. When the temperature deviation exceeds a critical value, the controller switches to maximum heat dissipation mode, rapidly reducing the temperature by running the fan at full speed. During this process, sliding mode control ensures the system's strong robustness to disturbances, while fuzzy logic effectively suppresses control signal chattering.

[0086] Compared to existing technologies, traditional fan control typically employs fixed thresholds or linear adjustment strategies, making it difficult to adapt to nonlinear heat dissipation demands under dynamic operating conditions. For example, existing technologies may lead to increased energy consumption due to overly aggressive adjustments when temperature deviations are small, while insufficient response and overshoot may occur when temperature deviations are high. This solution combines a precise mapping model with fuzzy sliding mode control, ensuring both adjustment accuracy and differentiated strategies under different operating conditions, effectively balancing heat dissipation efficiency and energy consumption.

[0087] In this embodiment, the control of the high-pressure electronic water pump in the actuator collaborative control module includes:

[0088] The control target is the coolant temperature difference ΔT coolant =T out -T in It stabilizes within ±2℃ of the set range, where T out T represents the outlet coolant temperature of the fuel cell stack (in K). in This indicates the temperature of the coolant at the fuel cell stack inlet (in K).

[0089] Real-time estimation of pipeline resistance using a sliding mode observer:

[0090]

[0091] In the above formula, K represents the estimated value of the pipeline resistance. p and K d Both are dynamic adjustment coefficients, with values ​​ranging from [0.8, 1.2] to [0.1, 0.3].

[0092] The sliding mode observer is a dynamic state estimator designed based on sliding mode control theory. Specifically, it can be implemented by linearly combining the current temperature difference and its rate of change. An observation model is constructed by real-time acquisition of coolant inlet and outlet temperature data to compensate for the impact of pipeline resistance changes on flow control. The dynamic adjustment coefficient refers to the parameters in the sliding mode observer used to adjust the proportional and derivative actions. It can be updated using an online adaptive algorithm, adjusting the observer gain through real-time feedback to ensure accurate estimation of pipeline resistance under different operating conditions.

[0093] Specifically, in the control process of the high-pressure electronic water pump, the stability of the temperature difference between the coolant inlet and outlet is achieved through dynamic compensation by a sliding mode observer. When the coolant flow rate fluctuates, the sliding mode observer adjusts the proportional and derivative coefficients online using a linear combination of the current temperature difference value and the rate of change of the temperature difference, thereby generating a real-time estimate of the pipeline resistance. This estimate is further used to correct the pump speed control command, enabling the coolant flow rate to quickly respond to changes in actual demand and maintain the temperature difference within a preset range.

[0094] Compared to existing technologies, conventional high-pressure electronic water pump control typically employs observers with fixed parameters or resistance estimation based on empirical formulas, making it difficult to adapt to the nonlinear characteristics caused by changes in coolant viscosity with temperature. This solution introduces a sliding mode observer and a dynamic adjustment coefficient to achieve online tracking of pipeline resistance, effectively solving the problem of insufficient estimation accuracy under variable temperature conditions using traditional methods, while also avoiding the risk of control instability caused by parameter drift.

[0095] In this embodiment, the control of the three-way valve in the actuator coordination control module includes:

[0096] Calculate the temperature distribution nonuniformity σ of the fuel cell stack T :

[0097]

[0098] In the above formula, T i This represents the temperature of the i-th section of the fuel cell stack (in °C). The average temperature of the fuel cell stack is expressed in °C, and N represents the number of fuel cell stack sections.

[0099] Based on the temperature distribution non-uniformity σ T And coolant flow deviation ΔQ=Q actual -Q target Dynamically allocate branch flow, and optimize the objective function as follows: Among them, Q actual Q represents the actual flow rate (in L / min). target α represents the target flow rate (in L / min). c and β c All are weighting coefficients for three-way valve control, with weighting coefficient α. c and β c Satisfy: α c ∈[0.6,0.8]、β c ∈[0.2,0.4], and adjusted online using fuzzy rules.

[0101] Temperature distribution non-uniformity refers to the dispersion of temperature between each section of the fuel cell stack and the average temperature. This can be achieved using a standard deviation calculation formula. Real-time temperature data is acquired by temperature sensors placed in each section of the stack, and the standard deviation of the temperature distribution is calculated to quantify the uniformity of the temperature field. Coolant flow deviation refers to the difference between the actual flow rate and the target flow rate. This can be achieved by collecting branch flow data using flow sensors and comparing it in real time with a preset target flow rate value to evaluate the accuracy of flow control. Dynamic allocation of branch flow refers to adjusting the flow ratio of each branch according to the current temperature distribution. This can be achieved by adjusting the opening of a three-way valve. An optimal allocation scheme is generated using an objective function optimization algorithm to balance temperature uniformity and system energy consumption.

[0102] Specifically, temperature sensors located in each section of the fuel cell stack periodically collect temperature data. The temperature distribution non-uniformity is calculated using the standard deviation formula. Simultaneously, flow sensors acquire the deviation between the actual and target flow rates in each branch. These two parameters are input into a fuzzy rule base, and the weighting coefficients are adjusted online to adapt to the current operating conditions. The optimization algorithm solves for the optimal flow distribution ratio based on a weighted objective function. By controlling the opening of the three-way valve, the flow rate in each branch is adjusted, improving temperature distribution uniformity while maintaining flow control accuracy. For example, when temperature distribution non-uniformity increases, the algorithm automatically increases its weighting coefficient, prioritizing the optimization of temperature uniformity; when the flow deviation exceeds a threshold, the flow deviation weight is increased to ensure system flow stability.

[0103] Compared to existing technologies, traditional three-way valve control typically employs fixed weight allocation strategies or single-parameter feedback control, failing to adapt simultaneously to dynamic changes in temperature distribution and flow deviation. Existing three-way valve control methods largely rely on preset rule bases for opening adjustment, which can easily lead to localized overheating or flow oscillations under complex operating conditions. This solution achieves multi-parameter collaborative optimization by adjusting weight coefficients online and dynamically optimizing the objective function, effectively improving temperature field uniformity and actuator control stability.

[0104] S4. Based on measured water temperature data, the least squares method is used to establish a quantitative relationship between coolant flow rate and temperature, and the model parameters are calibrated accordingly.

[0105] Compared with existing technologies, traditional PID controllers cannot handle temperature fluctuations at the hundred-millisecond level, while the dual-layer fuzzy sliding mode architecture of this solution significantly improves the dynamic response speed through online disturbance estimation and parameter adaptation. Existing centralized control strategies struggle to coordinate the actions of multiple actuators, while the cooperative control model of this solution effectively solves the coupling problem between heating and heat dissipation equipment through a signal line linkage mechanism. Furthermore, the parameter calibration method based on measured data is better suited to the nonlinear characteristics of coolant physical parameters changing with temperature compared to fixed parameter models.

[0106] Example 2:

[0107] Please see Figure 1 A fuel cell vehicle thermal management system constructed using the construction method of Example 1 includes:

[0108] Thermodynamic modeling module, which is used to calculate the heat generated by the fuel cell stack and the temperature change of the coolant in real time;

[0109] A two-layer fuzzy sliding mode controller, which generates disturbance estimates through an upper-layer controller and outputs actuator control laws through a lower-layer controller;

[0110] The actuator coordination control module is used to control the PTC heater, the fan, the high-pressure electronic water pump, and the three-way valve.

[0111] The adaptive operating condition module is used to dynamically adjust control parameters based on vehicle speed and ambient temperature.

[0112] The fault protection module is used to switch to redundant control mode in case of abnormal temperature.

[0113] The system acquires data at 100Hz via CAN bus, with an embedded controller control cycle of ≤10ms and an overall response time of ≤50ms.

[0114] The thermodynamic modeling module refers to a mathematical model that calculates real-time heat generation based on electrochemical reaction power and heat loss. It can be implemented using the Matlab-Simulink platform by integrating the fuel cell stack heat generation module, coolant circulation module, and heat dissipation module. By calculating the fuel cell stack heat generation and coolant temperature changes in real time, it provides accurate thermodynamic parameters for subsequent control. The dual-layer fuzzy sliding mode control module is a composite controller that includes upper-layer disturbance estimation and lower-layer actuator control. It can be implemented by adjusting parameters online using fuzzy rules and generating control signals using sliding mode control laws. By dynamically estimating disturbances and adjusting control parameters, it improves the system's anti-interference capability. The actuator collaborative control module refers to the linkage control model for the PTC heater, fan, water pump, and three-way valve. It can be implemented through Simulink signal lines to achieve multi-actuator command coordination. By coordinating the actions of each actuator, it avoids control conflicts and optimizes energy consumption. The driving condition adaptive module is an algorithm that dynamically adjusts the control strategy according to driving conditions. It can be implemented by switching control modes using vehicle speed and ambient temperature as input parameters. By matching the control objectives under different driving conditions, it maintains the system's stability across the entire driving range. The fault protection module refers to the safety control mechanism that is activated when the temperature is abnormal. It can be implemented by using redundant control strategies and preset safety parameters. By quickly switching to a safe mode, it prevents the fuel cell stack from overheating or components from being damaged.

[0115] Specifically, the thermodynamic modeling module establishes a dynamic thermodynamic model of the fuel cell stack and outputs real-time data on heat generation and coolant temperature; the upper controller of the dual-layer fuzzy sliding mode control module adjusts the disturbance estimation parameters online based on fuzzy rules, while the lower controller generates actuator control signals based on the sliding mode surface function; the actuator collaborative control module distributes control signals to the PTC heater, fan, water pump, and three-way valve, achieving synchronized actuator actions through signal lines; the operating condition adaptive module automatically adjusts the target temperature range and control parameters when the vehicle enters special operating conditions such as low-temperature start-up or high-speed operation; the fault protection module continuously monitors temperature deviations and immediately switches to a preset safety mode when continuous abnormalities are detected; the system ensures that the control cycle and response time meet real-time requirements through high-speed data acquisition and embedded controllers.

[0116] Compared to existing technologies, current thermal management systems typically use independent controllers to manage actuators separately. The lack of coordination strategies leads to response delays and increased energy consumption, and they also lack integrated adaptive operating conditions and fault protection mechanisms. This solution achieves multi-actuator collaborative control, dynamic operating condition matching, and rapid fault response through modular integration, solving the problems of temperature fluctuations and instability caused by decentralized control in traditional systems.

[0117] In this embodiment, the strategy of the working condition adaptive module includes:

[0118] Low-temperature start-up condition: Control the power of the PTC heater to increase to more than 80% of the rated power, and limit the fan speed to no more than 2000 rpm;

[0119] High-speed operation: Narrow the target range of coolant temperature difference to ±1℃;

[0120] The fuzzy rule base is based on vehicle speed v and ambient temperature T. amb Switch control parameters when vehicle speed v > 80 km / h and ambient temperature T amb When the temperature is <0℃, the approach rate coefficient ρ is adjusted to 1.2 times the original value.

[0121] Specifically, the power increase of the PTC heater in low-temperature start-up conditions refers to accelerating the coolant heating process by increasing the input voltage. This can be achieved by dynamically adjusting the voltage duty cycle using a proportional-integral algorithm, which can quickly establish thermal equilibrium during the cold start phase. Fan speed limiting involves setting an upper speed limit to avoid excessive heat dissipation. This can be achieved using a PWM signal duty cycle limiting module to prevent coolant overcooling and subsequent performance degradation of the fuel cell stack in low-temperature environments. Narrowing the target range for coolant temperature difference refers to adjusting the controller's error tolerance range. This can be achieved by modifying the error threshold parameter in the sliding mode surface function, thereby enhancing temperature tracking accuracy under high-speed conditions. The approach rate coefficient adjustment refers to dynamically changing the controller's convergence speed based on environmental parameters. This can be achieved by querying a preset fuzzy rule table to update parameters online, addressing thermal disturbances under high-speed and low-temperature combined conditions.

[0122] Specifically, during vehicle startup in low-temperature environments, the power of the PTC heater is increased to over 80% of its rated value. Combined with the fan speed limit of no more than 2000 rpm, this ensures that the fuel cell stack quickly reaches operating temperature while avoiding excessive heat dissipation and energy waste. When the vehicle enters high-speed operation, the coolant temperature difference control target range is adjusted from ±2℃ to ±1℃. By increasing the tracking accuracy requirements of the sliding mode controller, sudden changes in heat dissipation efficiency caused by enhanced airflow are suppressed. The fuzzy rule base monitors vehicle speed and ambient temperature parameters in real time. When the vehicle speed exceeds 80 km / h and the ambient temperature is below 0℃, the approach rate coefficient is automatically increased to 1.2 times the original value, enhancing the controller's anti-interference capability under high wind speed and low temperature conditions.

[0123] Compared to existing technologies, traditional methods typically employ fixed control parameters or single-condition optimization strategies, which cannot effectively cope with complex and ever-changing driving environments. For example, conventional control strategies may simultaneously increase heating and cooling power during low-temperature starts, leading to energy waste; and they fail to consider the nonlinear impact of airflow on heat dissipation efficiency under high-speed, low-temperature conditions. This solution achieves precise matching between the control strategy and environmental conditions by establishing a multi-dimensional condition identification mechanism and dynamic parameter adjustment rules.

[0124] In this embodiment, the fault protection module's mechanism includes:

[0125] Real-time monitoring of fuel cell stack temperature sensor data; if the absolute value of the temperature deviation exceeds 5°C within three consecutive sampling cycles, a fault alarm will be triggered.

[0126] In case of failure, switch to redundant control mode: fix the speed of the high-pressure electronic water pump at a preset safe value and increase the cooling power of the fan by 20%.

[0127] Real-time monitoring of fuel cell stack temperature sensor data refers to collecting temperature signals through water temperature sensors arranged in each section of the fuel cell stack. Specifically, this can be achieved using NTC thermistors or thermocouples. The sampling frequency is synchronized with the control cycle to dynamically track the temperature distribution of the fuel cell stack. The preset safety value refers to a water pump speed threshold pre-set based on the minimum flow requirements of the cooling system. This threshold can be obtained through experimental calibration to ensure the maintenance of basic coolant circulation capacity under fault conditions. The redundant control mode is a backup control strategy activated when the main controller fails or the temperature is abnormal. It is implemented through hardware circuitry independent of the main controller and can quickly take over actuator control.

[0128] Specifically, when the absolute value of the temperature deviation exceeds the set threshold three times consecutively, the system determines that there is a risk of temperature runaway. At this time, the fault protection module immediately interrupts the output command of the main controller and switches the speed of the high-pressure electric water pump to a preset safe speed value, which is preset according to the minimum circulation flow requirement of the cooling system. At the same time, the fan drive module receives a command to increase power by 20%, accelerating heat dissipation by increasing the airflow to the radiator. This control strategy is implemented through hardware-level signal switching, ensuring that the control mode transition is completed within 50ms.

[0129] Anomaly detection of temperature sensor data can be verified using a sliding window algorithm, for example, triggering an alarm when three consecutive sampling points exceed the threshold range. The preset safe operating speed can be calculated based on coolant properties, such as the operating speed required to maintain a coolant flow rate of no less than 0.5 m / s within a temperature range of -20℃ to 80℃.

[0130] Compared to existing technologies, traditional fault handling methods typically employ a single threshold-triggered protection mechanism, but lack continuous monitoring, making them prone to false triggering due to momentary interference. Existing redundant control systems often use a fixed power output mode, neglecting dynamic matching between actuators. This solution reduces the false alarm rate through continuous periodic monitoring and, combined with the coordinated adjustment of pump speed and fan power, ensures basic heat dissipation capacity while preventing actuator overload.

[0131] Comparative example: Application of thermal management system in fuel cell vehicle under low-temperature start-up conditions.

[0132] 1. Scene setting

[0133] Environmental conditions:

[0134] Ambient temperature T amb = -10℃, vehicle speed v = 0km / h (cold start).

[0135] Control objective:

[0136] Stack temperature T stack It rapidly rises from an initial value of -5℃ to 70℃.

[0137] Coolant temperature difference ΔT coolant It remains stable at ±2℃.

[0138] Key parameters:

[0139] fuel cell output power P elec =5000W, electrochemical efficiency coefficient η=0.6.

[0140] Coolant density ρ cool =1050kg / m 3 Specific heat capacity c p = 3400 J / (kg·K).

[0141] System volume V sys =0.005m 3 .

[0142] 2. Calculation of dynamic thermodynamic model of fuel cell stack

[0143] Calculation of heat generation from fuel cell stack:

[0144] Q gen =P elec ·(1-η)-Q loss

[0145] Q loss =h·A·(T) stack -T ambient ).

[0146] Substituting the parameters (h = 5W / (m)) 2 ·K),A=2m 2 ):

[0147] Q loss =5×2×[-5-(-10)]=50W,Q gen =5000×(1-0.6)-50=1950W.

[0148] Coolant temperature change (one-dimensional fluid dynamics model):

[0149]

[0150] Initial heat dissipation power Q fan =0, Q PTC =0, solving for:

[0151]

[0152] 3. Dual-layer fuzzy sliding mode controller

[0153] Upper-level controller (interference estimation):

[0154]

[0155] in

[0156] set up Δη = 0, fuzzy rule output:

[0157] Input sliding surface derivative (Normalized to [-1,1])→

[0158] Δk = negative large = -0.9 (rule: (When the value is negative, output negative)

[0159] Δη = negative maximum = -0.9 (rule: (When the value is negative, output negative)

[0160] Take G=1, G′=1, λ=0.1:

[0161]

[0162] When t = 1s:

[0163]

[0164] Lower-level controller (executes control law):

[0165]

[0166] Parameter settings: b = 1, ε = 0.5, ρ = 1.2, f(T) = 0, α = 0.1, β = 0.1, p = 3, q ​​= 1 (p > q and is a positive odd number).

[0167] Sliding surface function:

[0168] s=e+αe+βe p / q = e + 0.1e + 0.1e 3 =1.1e + 0.1e 3

[0169] Current e = T target -T stack =70-(-5)=75, (Based on thermodynamic models):

[0170] s = 1.1 × 75 + 0.1 × 75 3 =82.5 + 42,187.5 = 42,270

[0171] Control output:

[0172] u(1)=0.5·1+1.2×42,270-0.055+0.1×0.11+0.1×3×75 2 ×0.11≈50,910

[0173] 4. Actuator-coordinated control objective function for PTC heater:

[0174]

[0175] Low-temperature startup strategy: u PTC =38.4V (80% of the rated 48V).

[0176] Fan-controlled heat dissipation model:

[0177]

[0178] Temperature deviation ΔT = 75 > 10℃, activate maximum heat dissipation power → n fan =3000rpm (upper limit). High-pressure electronic water pump control, sliding mode observer estimates pipeline resistance:

[0179]

[0180] Take K p =1.0, K d =0.2, ΔT coolant =5℃, :

[0181]

[0182] Three-way valve control:

[0183] Temperature distribution non-uniformity (3-cell fuel cell stack, T1 = -4.5, T2 = -5.0, T3 = -5.5):

[0184] Optimization goal:

[0185]

[0186] 5. Coolant flow rate and temperature calibration

[0187] Fitting relationship:

[0188] ΔT coolant =a·ln(Q) flow )+b

[0189] Calibration results: a = -2.0, b = 15, flow rate Q flow =10L / min:

[0190] ΔTcoolant =-2.0·ln(10)+15≈10.39℃

[0191] 6. Adaptive operating conditions and fault protection

[0192] Low-temperature start-up strategy:

[0193] PTC power > 80% → u PTC =38.4V.

[0194] Fan speed limit ≤2000rpm.

[0195] Fault protection:

[0196] For three consecutive cycles, |ΔT| > 5℃ → Redundancy mode:

[0197] Water pump speed n pump =3000rpm (preset safe value).

[0198] Fan power increased by 20%.

[0199] Output result:

[0200] 1. Heat production model: Q gen =P elec ·(1-η)-Q loss .

[0201] 2. Control Law: Two-layer fuzzy sliding mode (disturbance estimation) +Exponential term e p / q ).

[0202] 3. Actuator coordination:

[0203] PTC objective function includes energy consumption constraints

[0204] Fan model

[0205] 4. Simulation verification:

[0206] The inlet temperature rises to 68°C in 120 seconds, with an overshoot of <3°C.

[0207] The fan speed remains stable below 2000 rpm.

[0208] The system response time is ≤50ms, verifying millisecond-level response and multi-executor collaborative optimization.

[0209] Figure 2 The inlet temperature curve of the cooling water for the fuel cell stack reflects its dynamic response characteristics, demonstrating the ability to rapidly track the target setpoint within milliseconds.

[0210] Initial start-up phase: The temperature rises rapidly from the ambient temperature to the target value, and the rise time is fast, indicating the high efficiency of the actuator collaborative control module of the fuel cell vehicle thermal management system in controlling the high-pressure electronic water pump and PTC heater.

[0211] Steady-state phase: The curve fluctuates within a very small range, demonstrating the ability of the two-layer fuzzy sliding mode controller to suppress disturbances during sudden load changes or ambient temperature fluctuations.

[0212] Disturbance immunity: Simulating a decrease in fan efficiency at the point of external disturbance injection: the temperature recovers quickly after a brief deviation, verifying the real-time compensation effect of the lower-level controller. Internal disturbances, such as step changes in fuel cell power: the curve shows no overshoot, demonstrating the robustness of the lower-level controller.

[0213] Figure 3 The cooling water outlet temperature profile of the fuel cell stack reflects:

[0214] The thermodynamic coupling relationship shows that the outlet temperature is always higher than the inlet temperature, which is consistent with the heat generation model of the fuel cell stack. Dynamic hysteresis: the change in outlet temperature is delayed compared to the inlet temperature (depending on the coolant flow rate), and the phase difference in the curves verifies the accuracy of the thermodynamic modeling.

[0215] Operating condition adaptive verification, operating condition switching point (such as switching from urban operating condition to high-speed operating condition): The outlet temperature rises briefly due to the increase in fuel cell power, but the control parameters are adjusted by the operating condition adaptive module and it returns to the set range in a short time.

[0216] Fault protection trigger: An abnormal rise in temperature triggers the protection mechanism, and the curve is forcibly limited to below the safety threshold for a short period of time.

[0217] Figure 4 The fan speed curve of the fuel cell radiator reflects:

[0218] The control strategy demonstrates that the fan speed is dynamically adjusted according to the heat load of the fuel cell stack: the speed increases sharply in the high-power range, and the response time is short, reflecting the real-time calculation capability of the actuator collaborative control module.

[0219] Low power / low temperature conditions: The rotational speed is maintained at the lowest threshold to verify the energy consumption optimization strategy of the adaptive module.

[0220] Disturbance resistance characteristics, ambient temperature change (simulating high temperature): the speed step increase compensates for the decrease in heat dissipation efficiency, proving the effectiveness of the feedforward compensation of the dual-layer fuzzy sliding mode controller, fault protection linkage: speed over-limit triggers safety lock (red line area), reflecting a multi-level protection mechanism.

[0221] Figure 5 The temperature difference curve between the inlet and outlet temperatures of the fuel cell stack reflects:

[0222] Thermodynamic model verification shows that the steady-state temperature difference is stable within the design range, conforming to the thermal balance equation, reflecting that the accuracy of the dynamic thermodynamic model of the fuel cell stack meets the requirements.

[0223] Transient response: The temperature fluctuation amplitude is small and recovers quickly, verifying the ability of the dual-layer fuzzy sliding mode controller to suppress heat generation disturbances. System reliability indicators, fault boundary control (coolant leakage): A sudden increase in temperature difference triggers a threshold alarm, and a fault power reduction strategy is activated within a short period of time, with the temperature difference returning to the safe range.

[0224] Figure 6 The cooling water flow rate curve in the fuel cell thermal management system reflects:

[0225] Actuator collaborative intelligence, flow rate and fan decoupling control: When there is a high demand for heat dissipation, flow rate increase takes precedence over fan speed increase, reflecting the optimal energy consumption collaborative strategy.

[0226] Low-temperature start-up phase: The flow rate gradually increases to avoid thermal shock, proving that the adaptive operating condition module optimizes parameters for cold start conditions.

[0227] The control accuracy and robustness are excellent, with small flow setpoint tracking error and no overshoot, reflecting the high-precision driving capability of the dual-layer fuzzy sliding mode controller for the high-pressure electronic water pump. Under sudden load changes, the instantaneous flow fluctuation is small, and disturbance suppression is superior to traditional PID control.

[0228] Figure 7 The opening curve of the three-way valve in the fuel cell thermal management system reflects:

[0229] Temperature hybrid control logic, low temperature condition: 100% opening (full heating circuit) to achieve rapid warm-up.

[0230] High-temperature operating conditions: 0% opening (full heat dissipation circuit) to maximize heat dissipation efficiency.

[0231] Transition zone: The opening degree is continuously adjustable, reflecting the precise mixing control of the proportional valve.

[0232] Fail-safe strategy for sensor failure: Lock the opening at 50% safe position to ensure basic circulation flow and prevent fuel cell overheating.

[0233] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention. Those skilled in the art, under the guidance of the present invention, can make various similar representations without departing from the spirit and claims of the present invention, and such modifications all fall within the protection scope of the present invention.

Claims

1. A method for constructing a thermal management system for a fuel cell vehicle, characterized in that, Follow these steps: S1. Use the stack heat generation module, coolant circulation module and heat dissipation module in Matlab-Simulink to establish a dynamic thermodynamic model of the fuel cell stack; S2. Design a two-layer fuzzy sliding mode controller that includes an upper-layer controller and a lower-layer controller. The upper-layer controller is used to adjust the disturbance estimate online through fuzzy rules and generate a compensation term that includes environmental heat loss and model uncertainty. The lower-layer controller is used to design a control law based on the sliding mode surface function and combine the approach rate coefficient to suppress chattering. S3. Embed the execution control law output by the lower-level controller into the actuator collaborative control module of PTC heater, fan, high-pressure electronic water pump and three-way valve, and realize the linkage control of PTC heater, fan, high-pressure electronic water pump and three-way valve through Simulink signal line; S4. Based on measured water temperature data, the least squares method is used to establish a quantitative relationship between coolant flow rate and temperature, and the model parameters are calibrated accordingly.

2. The construction method according to claim 1, characterized in that, In step S1, the fuel cell stack heat generation module uses a heat loss calculation model. This heat loss calculation model calculates the real-time heat generation based on the electrochemical reaction power and heat loss, satisfying the formula: Q gen =P elec ·(1-η)-Q loss ; In the above formula, Q gen P represents the real-time heat generation of the fuel cell stack. elec The output power of the fuel cell stack is represented by η, and the electrochemical efficiency coefficient is represented by Q. loss Indicates environmental heat loss; Among them, the environmental heat loss Q loss The calculation model is as follows: Q loss =h·A·(T stack -T ambient ); In the above formula, h represents the thermal conductivity coefficient, A represents the surface area of ​​the fuel cell stack, and T... stack T represents the surface temperature of the fuel cell stack. ambient Indicates ambient temperature; The coolant circulation module adopts a one-dimensional fluid dynamics model, and its state equation is: In the above formula, dT represents the change in coolant temperature, dt represents the time derivative, and Q fan Q represents the power dissipated by the radiator fan. PTC This represents the heating power of the PTC heater, ρ cool V represents the density of the coolant. sys c represents the system volume. p This indicates the specific heat capacity of the coolant; The expression for the heat dissipation module is: Q fan =k eff ·(T cool_out -T ambient )·A radiator ; In the above formula, k eff T represents the overall heat dissipation coefficient. cool_out Indicates the coolant temperature at the radiator outlet, A radiator This indicates the effective heat dissipation area of ​​the radiator.

3. The construction method according to claim 2, characterized in that, In step S2, the upper-level controller generates an interference estimate. The expression is: In the above formula, λ represents a constant compensation term greater than 0, and k1(t) represents an adaptive adjustment term, the expression of which is: in, Let represent the absolute value of the derivative of the Lyapunov function, G represent the proportional adjustment coefficient, G′ represent the compensation intensity coefficient, t represent time, Δk represent the fuzzy adjustment amount of the sliding mode control gain, and Δη represent the fuzzy adjustment amount of the electrochemical efficiency compensation coefficient. Both Δk and Δη are adjusted online through fuzzy rules. The execution control law u(t) of the lower-level controller is as follows: In the above formula, ε and ρ both represent the approach rate coefficients greater than 0, and s represents the sliding surface function, which is expressed as s = e + αe + βe. p / q p and q both represent the coefficients of the sliding surface exponent, both p and q are positive odd numbers, and p > q. e represents the temperature error, which is expressed as e = T. target -T actual T target T represents the target temperature. actual Let represent the actual temperature, α represent the error proportionality coefficient, β represent the weighting coefficient of the exponential term, f(T) represent the nonlinear term of the thermodynamic model, and sgn(s) represent the sign function. The second derivative represents the target temperature. This indicates that the upper-level controller generates an estimated disturbance value. This indicates the rate of change of temperature error.

4. The construction method according to claim 3, characterized in that, The fuzzy rules include: The input variable is the sliding surface derivative, and the fuzzy set is divided into negative large, negative medium, zero, positive medium, and positive large; The fuzzy rule for the output variable Δk is as follows: when the derivative of the sliding surface is negative (large) or negative (medium), the output is negative (large); when the derivative of the sliding surface is zero, the output is zero; when the derivative of the sliding surface is positive (medium) or positive (large), the output is positive (large). The fuzzy rules for the output variable Δη are as follows: when the derivative of the sliding surface is negative, the output is negative; when the derivative of the sliding surface is negative, the output is negative; when the derivative of the sliding surface is zero, the output is zero; when the derivative of the sliding surface is positive, the output is positive; when the derivative of the sliding surface is positive, the output is positive. The membership function adopts a triangular distribution, and the universe of discourse is normalized to [-1, 1].

5. The construction method according to claim 4, characterized in that, In step S3, the control of the PTC heater in the actuator collaborative control module includes: Based on temperature deviation ΔT=T actual -T target and the rate of change of temperature deviation The fuzzy set is divided into negative large, negative medium, zero, positive medium, and positive large; The objective function J for sliding mode control is designed as follows: In the above formula, u PTC t represents the control input voltage of the PTC heater. end This indicates the end time of the control cycle, and 0.05 is the weighting coefficient, which was determined through offline optimization experiments.

6. The construction method according to claim 4, characterized in that, In step S3, the control of the fan in the actuator coordination control module includes: Construct a mapping model between fan speed and heat dissipation efficiency: In the above formula, Q fan Indicates heat dissipation, n fan Indicates fan speed, ΔT radiator Indicates the temperature difference of the radiator; The fan speed is dynamically adjusted based on a dual-layer fuzzy sliding mode controller. The fuzzy rules are matched according to the load conditions: when the temperature deviation ΔT < 5℃, the fan speed increase is limited to no more than 10%; when the temperature deviation ΔT ≥ 10℃, the maximum heat dissipation power is enabled.

7. The construction method according to claim 4, characterized in that, In step S3, the control of the high-pressure electronic water pump in the actuator collaborative control module includes: The control target is the coolant temperature difference ΔT coolant =T out -T in It stabilizes within ±2℃ of the set range, where T out T represents the outlet coolant temperature of the fuel cell stack. in Indicates the temperature of the coolant at the fuel cell stack inlet; Real-time estimation of pipeline resistance using a sliding mode observer: In the above formula, K represents the estimated value of the pipeline resistance. p and K d Both are dynamic adjustment coefficients, with values ​​ranging from [0.8, 1.2] to [0.1, 0.3].

8. The construction method according to claim 4, characterized in that, In step S3, the control of the three-way valve in the actuator coordination control module includes: Calculate the temperature distribution nonuniformity σ of the fuel cell stack T : In the above formula, T i This represents the temperature of the i-th section of the fuel cell stack. The average temperature of the fuel cell stack is represented by N, which represents the number of fuel cell stack sections. Based on the temperature distribution non-uniformity σ T And coolant flow deviation ΔQ=Q actual -Q target Dynamically allocate branch flow, and optimize the objective function as follows: Among them, Q actual Q represents the actual traffic volume. target Represents the target flow, α c and β c All are weighting coefficients for three-way valve control, with weighting coefficient α. c and β c Satisfy: α c ∈[0.6,0.8]、β c ∈[0.2,0.4], and adjusted online using fuzzy rules.

9. A thermal management system for a fuel cell vehicle constructed using the construction method described in any one of claims 1-8, characterized in that, include: Thermodynamic modeling module, which is used to calculate the heat generated by the fuel cell stack and the temperature change of the coolant in real time; A two-layer fuzzy sliding mode controller, which generates disturbance estimates through an upper-layer controller and outputs actuator control laws through a lower-layer controller; The actuator coordination control module is used to control the PTC heater, the fan, the high-pressure electronic water pump, and the three-way valve. The adaptive operating condition module is used to dynamically adjust control parameters based on vehicle speed and ambient temperature. The fault protection module is used to switch to redundant control mode in case of abnormal temperature. The system acquires data at 100Hz via CAN bus, with an embedded controller control cycle of ≤10ms and an overall response time of ≤50ms.

10. The fuel cell vehicle thermal management system according to claim 9, characterized in that, The strategies of the adaptive operating condition module include: Low-temperature start-up condition: Control the power of the PTC heater to increase to more than 80% of the rated power, and limit the fan speed to no more than 2000 rpm; High-speed operation: Narrow the target range of coolant temperature difference to ±1℃; The fuzzy rule base is based on vehicle speed v and ambient temperature T. amb Switch control parameters when vehicle speed v > 80 km / h and ambient temperature T amb When the temperature is <0℃, the approach rate coefficient ρ is adjusted to 1.2 times its original value; The fault protection module's mechanism includes: Real-time monitoring of fuel cell stack temperature sensor data; if the absolute value of the temperature deviation exceeds 5°C within three consecutive sampling cycles, a fault alarm will be triggered. In case of failure, switch to redundant control mode: fix the speed of the high-pressure electronic water pump at a preset safe value and increase the cooling power of the fan by 20%.

Images