Ultralow-temperature cold start thermal management system for hydrogen fuel cell automobile

Abstract

The invention discloses a hydrogen fuel cell vehicle ultralow temperature cold start thermal management system, and relates to the technical field of hydrogen fuel cell vehicle thermal management, the hydrogen fuel cell vehicle ultralow temperature cold start thermal management system comprises a thermal management platform, the thermal management platform is in communication connection with the following modules: a temperature monitoring and predicting module, and the temperature prediction module is used for monitoring the temperature of key parts in the hydrogen fuel cell by using the temperature sensor and predicting the change trend of the temperature in the fuel cell in combination with the temperature prediction model. By utilizing the high-precision temperature sensor and the long-short-term memory network model, the internal temperature change trend of the fuel cell is pre-judged in advance, the heating power is dynamically adjusted, it is ensured that the temperature of the cell is rapidly increased to the proper working temperature in the ultralow-temperature environment, starting difficulty caused by local supercooling is avoided, the low-temperature starting time is remarkably shortened, and the starting efficiency is improved. The reliability and the stability of the system in an extreme environment are improved, and the normal operation of the hydrogen fuel cell vehicle in a cold region is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal management of hydrogen fuel cell vehicles, and in particular to an ultra-low temperature cold start thermal management system for hydrogen fuel cell vehicles. Background Art

[0002] Hydrogen fuel cells generate electricity through the chemical reaction of hydrogen and oxygen, which releases water and heat. The hydrogen is first stored and converted into electricity through the electrochemical reaction in the fuel cell stack. This process depends on suitable temperature conditions. Hydrogen fuel cell vehicles (FCEVs) are regarded as an important development direction for future automobiles due to their zero emissions, long driving range and high efficiency. However, the performance of hydrogen fuel cell vehicles in low temperature environments remains one of the challenges of technological development, especially the cold start problem.

[0003] In the existing technology, in an ultra-low temperature environment, it is difficult for the thermal management system to accurately control the temperature inside the fuel cell, resulting in local overheating or overcooling, affecting the battery performance and life. In addition, there is a delay in the system's response to temperature changes, and it is impossible to adjust the heating or cooling strategy in time, resulting in a decrease in battery performance in the initial stage of startup. Therefore, how to predict temperature changes, dynamically adjust the heating and cooling strategies, and shorten the system's response time to temperature changes is the problem to be solved by the present invention. To this end, a hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system is proposed. Summary of the Invention

[0004] The object of the present invention is to provide a hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system includes a thermal management platform, wherein the thermal management platform is communicatively connected to the following modules, wherein:

[0007] The temperature monitoring and prediction module is used to monitor the temperature of key parts inside the hydrogen fuel cell using a temperature sensor and, in combination with a temperature prediction model, predict the temperature change trend inside the fuel cell;

[0008] The thermal management control module is used to dynamically adjust the heating system and the cooling system to perform heating or cooling operations according to the output of the temperature monitoring and prediction module;

[0009] The temperature response adjustment module is used to adopt an adaptive control response algorithm to monitor and quickly adjust temperature changes in real time, thereby shortening the system's response time to temperature changes;

[0010] The actuator response module is used to optimize the response speed of the actuators of the heating and cooling systems, adjust the working state after receiving the control signal, reduce temperature fluctuations caused by actuator response delays, and improve the overall performance of the system.

[0011] A further improvement of the technical solution of the present invention is that: the temperature monitoring and prediction module includes a real-time temperature monitoring unit and a temperature prediction unit;

[0012] The real-time temperature monitoring unit is used to collect the temperature of key parts inside the hydrogen fuel cell using a high-precision temperature sensor, obtain real-time temperature data, and provide accurate input for the temperature prediction and control system;

[0013] The temperature prediction unit uses a long short-term memory network to predict the future temperature change trend of the fuel cell based on the data collected by the temperature sensor, combined with the thermal model and historical operation data of the fuel cell.

[0014] A further improvement of the technical solution of the present invention is that the real-time temperature monitoring unit specifically includes:

[0015] High-precision temperature sensors are distributed in key locations of hydrogen fuel cells and initialized and calibrated to ensure measurement accuracy. The connection status of the sensors is checked to ensure unobstructed data transmission channels to measure the temperature of the target area. Initialization calibration includes zero point calibration, gain calibration, and connection status check.

[0016] The sensor collects temperature data from key parts inside the hydrogen fuel cell in real time, transmits it to the thermal management platform in real time through a high-bandwidth communication interface (CAN bus), and preprocesses the collected temperature data, including filtering to eliminate noise interference, normalization to unify the dimension, and timestamp marking to ensure data timing while retaining the temperature change trend. The output voltage or resistance value of the temperature sensor is converted into a standard temperature value, the dimension is unified for subsequent analysis, and a high-precision timestamp is added to each data to ensure data timing and multi-sensor synchronization. The preprocessed temperature data is then stored in a temporary buffer for temporary storage.

[0017] A further improvement of the technical solution of the present invention is that the temperature prediction unit specifically includes:

[0018] Receive temperature data from key parts of the hydrogen fuel cell collected in real time by temperature sensors, and simultaneously load the fuel cell thermal model containing heat capacity, thermal resistance, and heat transfer coefficient parameters, as well as a historical operation database covering temperature-power mapping relationships under different operating conditions. Using timestamp alignment technology, map the timestamps of the sensor data, the thermal model calculation results, and historical data records to a unified time series coordinate system, extract multidimensional feature vectors, including the current temperature gradient, historical temperature fluctuation rate, power load change rate, and ambient temperature correction terms, and construct a feature matrix to characterize the dynamic characteristics of the system;

[0019] The feature matrix is ​​input into a temperature prediction model trained based on a long-short-term memory network model. Multi-layer gated units are used to capture long-term dependencies and short-term fluctuation patterns in the time series. Future temperature changes are predicted based on the maximum allowable temperature rise rate constraints in the thermal model. This prediction uses a sliding window mechanism, starting from the current moment, to generate a temperature trend curve for the next 5-10 seconds. The output is structured data containing the predicted value and its confidence interval.

[0020] The temperature prediction results output by the temperature prediction model are verified to check whether the prediction results meet the thermal model constraints, and the predicted temperature trend is compared with historical data under similar working conditions to evaluate the deviation range. The prediction results that pass the verification are marked as valid and transmitted to the thermal management control module for dynamic adjustment of the heating power strategy to reduce the temperature fluctuation amplitude.

[0021] A further improvement of the technical solution of the present invention is that: the thermal management control module includes a heating power adjustment unit and a cooling system adjustment unit;

[0022] The heating power adjustment unit is used to dynamically adjust the power output of the heating system according to the temperature prediction results to ensure that the fuel cell is quickly heated to a suitable operating temperature at the initial startup;

[0023] The cooling system regulating unit is used to automatically enable the cooling system and adjust the cooling parameters of the cooling system to reduce the temperature during the operation of the fuel cell according to the temperature prediction result, so as to prevent the battery from being damaged due to overheating.

[0024] A further improvement of the technical solution of the present invention is that the heating power adjustment unit specifically includes:

[0025] Receive the output data of the temperature monitoring and prediction module, including the trend curve of the fuel cell internal temperature change in the next 5-10 seconds and its confidence interval, and parse the output data of the temperature monitoring and prediction module to extract the slope and extreme points of the predicted temperature curve. Combined with the thermal model constraints, calculate the remaining temperature difference and time window from the current to the target temperature. The slope of the predicted temperature curve is the heating rate, which represents the current heating trend. The extreme point is the moment of lowest temperature, which identifies the local overcooling risk area. The thermal model constraint is that the maximum allowable temperature rise rate is ≤5℃ / s.

[0026] Based on the target temperature trajectory, the regulation unit dynamically adjusts the heating system power output through a closed-loop control strategy. A multi-zone collaborative control algorithm is used to distribute the total heating power to key parts of the hydrogen fuel cell as needed. Based on the deviation between the real-time temperature of each zone and the target trajectory, the power allocation weights are optimized using a weighted least squares method to ensure that the temperature difference between each zone is ≤2°C.

[0027] A multi-dimensional evaluation of the heating power regulation effect is performed, the deviation integral between the actual heating trajectory and the target trajectory is calculated, and the heating efficiency and energy utilization efficiency are quantified. If the evaluation result shows that the deviation exceeds the preset deviation threshold, the control parameters are optimized to improve the regulation accuracy. If a heating system anomaly is detected, that is, the power output deviates from the command by more than 10%, the redundant heating module switching or safety shutdown mechanism is triggered to ensure system reliability.

[0028] A further improvement of the technical solution of the present invention is that the cooling system adjustment unit specifically includes:

[0029] The temperature monitoring and prediction module receives the temperature trend curve and confidence interval for the next 5-10 seconds, extracts the heating rate and extreme points, and uses numerical differentiation to calculate the slope of the temperature curve to determine whether the cooling threshold (temperature rise rate > 3°C / s) has been triggered. For the extreme points, the highest temperature moment in the temperature curve is located to identify potential overheating risk areas. Based on the thermal model constraints, a decision is made as to whether the cooling system needs to be activated. If the predicted temperature exceeds the safety threshold of 75°C or the temperature rise rate exceeds the limit, the cooling system is immediately activated. The remaining overtemperature and time window from the current temperature to the target temperature are calculated. Using a fuzzy logic algorithm, these overtemperature and time windows are mapped to the cooling demand level, and the initial parameters of the cooling system, including fan speed, water pump flow, and coolant distribution, are determined.

[0030] An adaptive PID control algorithm is used to dynamically adjust cooling parameters based on the real-time deviation between the predicted and actual temperatures. For fan speed, the proportional term quickly responds to deviations and suppresses temperature overshoot. For water pump flow, the integral term eliminates steady-state errors to ensure long-term cooling efficiency. For coolant distribution, the differential term predicts deviation trends and optimizes coolant flow distribution in each region. Combined with multi-region temperature sensor data, regional temperature difference constraints and cooling efficiency constraints are imposed on cooling parameters. For regional temperature difference constraints, the maximum temperature difference between each region is ensured to be ≤3°C to avoid local thermal stress concentration. For cooling efficiency constraints, the difference between the coolant temperature and the battery temperature is limited to ≥10°C to ensure heat exchange efficiency.

[0031] The integral of the deviation between the actual temperature trajectory and the target temperature trajectory, namely the root mean square error (RMSE), is calculated, and the cooling efficiency index and energy consumption index are analyzed. For the cooling efficiency index, when RMSE ≤ 2°C, it is judged as high-efficiency cooling. For the energy consumption index, when the cooling energy consumption per unit temperature difference ≤ 0.05kWh / °C, it is judged as energy-saving operation. If the evaluation result shows that the deviation exceeds the threshold, that is, RMSE > 3°C, the strategy optimization mechanism is triggered.

[0032] A further improvement of the technical solution of the present invention is that the temperature response adjustment module specifically includes:

[0033] Integrate real-time data from multi-region temperature sensors and combine it with the trend curve for the next 5-10 seconds output by the temperature prediction module to build a dynamic temperature field model. Use a sliding window filtering algorithm to eliminate sensor noise and extract temperature change characteristics including heating rate and extreme points.

[0034] Temperature mutations are determined based on a dynamic threshold method, which is divided into heating rate mutations and extreme point shifts. For heating rate mutations, if the temperature change rate exceeds ±2 standard deviations of the historical mean, it is determined to be rapid heating or cooling. For extreme point shifts, if the deviation between the predicted temperature extreme point and the target trajectory exceeds the preset offset threshold, a strategy adjustment is triggered.

[0035] Automatically switches control modes based on the type of sudden temperature rise / fall. For sudden temperature rises, the rapid heating strategy is activated, prioritizing the increase in heating system power allocation weight. For sudden temperature falls, the enhanced cooling strategy is activated, prioritizing the increase in fan speed and water pump flow, and optimizing the coolant distribution ratio. The fuzzy PID algorithm is used to dynamically adjust control parameters, including proportional gain, integral time, and differential time.

[0036] Based on the integral of the deviation between the predicted temperature and the actual temperature, the control parameters are dynamically corrected through the gradient descent method. Combined with the regional temperature synchronization index, that is, the temperature difference between each area is ≤2℃, the heating / cooling resource allocation is optimized through the quadratic programming algorithm.

[0037] A further improvement of the technical solution of the present invention is that the specific process of optimizing the allocation of heating / cooling resources is as follows:

[0038] Continuously calculate the deviation between the predicted temperature and the actual temperature, and integrate the calculated deviation to obtain the deviation integral to quantify the cumulative error;

[0039] An objective function is constructed based on the deviation integral to evaluate the difference between the control parameter and the actual temperature. The control parameter is iteratively optimized using the gradient descent method. The gradient of the objective function with respect to the current control parameter is calculated to determine the parameter adjustment direction. The control parameter is adjusted along the gradient direction to gradually reduce the deviation integral and improve control accuracy. The temperature change after adjustment is also monitored in real time.

[0040] The regional temperature synchronization index, i.e., the temperature difference between regions is ≤ 2°C, is introduced as a constraint condition. At the same time, the allocation efficiency of heating / cooling resources is used as the objective function to construct a quadratic programming model.

[0041] The quadratic programming algorithm is used to solve the quadratic programming model and calculate the optimal heating / cooling resource allocation plan. Based on the solution, the heating power or cooling flow is allocated to each area to synchronize the temperature of each area and maintain the uniformity of the temperature field inside the fuel cell.

[0042] A further improvement of the technical solution of the present invention is that the actuator response module specifically includes:

[0043] The actuator response module receives control signals from the thermal management control module in real time, including power adjustment commands for the heating or cooling system, and pre-processes the received control signals. It also checks the integrity and validity of the signals and confirms whether the control signals conform to the expected command format and range. If the command fluctuation is less than 1% within five consecutive sampling cycles, the current command value is locked. Multi-source signals are synchronized based on the system clock to avoid command conflicts. If the temperature change rate is greater than 2°C / s, the heating / cooling intensity command is prioritized. If the regional temperature difference is greater than 1.5°C, the resource allocation command is prioritized.

[0044] The processed control signal is converted into a specific drive signal to adjust the working state of the actuators of the heating and cooling systems. For the actuators of the heating system, rapid temperature increase is achieved by changing the power output of the heating element. For the actuators of the cooling system, rapid temperature reduction is achieved by adjusting the fan speed and water pump flow parameters.

[0045] Continuously monitor the actual working status of the actuator, including power output, speed and flow parameters, and feed back the status information to the thermal management control module in real time through the feedback loop. Compare the difference between the actual status and the target status, and use the adaptive control algorithm to dynamically optimize the actuator drive parameters to ensure the actuator's response speed and accuracy. At the same time, monitor the actuator's operating status in real time, detect and handle any existing faults or abnormal conditions, and ensure the long-term stable operation of the actuator.

[0046] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0047] 1. The present invention provides a hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system. By utilizing high-precision temperature sensors and long-short-term memory network models, the system can predict the temperature change trend inside the fuel cell in advance and dynamically adjust the heating power to ensure that the battery is quickly heated to a suitable operating temperature in an ultra-low temperature environment, avoiding starting difficulties caused by local overcooling, significantly shortening the low-temperature start time, improving the reliability and stability of the system in extreme environments, and ensuring the normal operation of hydrogen fuel cell vehicles in cold areas.

[0048] 2. The present invention provides a hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system. According to the temperature prediction results, the power output of the heating system and the cooling system are accurately adjusted to avoid energy waste. Through the PID control parameter dynamic tuning algorithm and the multi-zone collaborative control algorithm, the system can allocate heating power on demand, ensuring that the fuel cell heats up quickly in the initial stage of startup while reducing unnecessary energy consumption. In addition, the adaptive PID control algorithm dynamically adjusts the cooling parameters during the cooling process, further improving energy utilization efficiency and extending the vehicle's cruising range. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0050] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0051] Figure 2 Schematic diagram of the working process of the temperature response adjustment module of the present invention. DETAILED DESCRIPTION

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] Example 1, as Figure 1 、 Figure 2 As shown, the present invention provides a hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system, including a thermal management platform, the thermal management platform is communicatively connected to the following modules, wherein:

[0054] The temperature monitoring and prediction module is used to monitor the temperature of key parts inside the hydrogen fuel cell using temperature sensors and, in combination with the temperature prediction model, predict the temperature change trend inside the fuel cell. The temperature monitoring and prediction module includes a real-time temperature monitoring unit and a temperature prediction unit;

[0055] Among them, the real-time temperature monitoring unit is used to collect the temperature of key parts inside the hydrogen fuel cell using high-precision temperature sensors, obtain real-time temperature data, and provide accurate input for the temperature prediction and control system. High-precision temperature sensors (platinum resistance or thermocouple) are distributedly deployed in key parts of the hydrogen fuel cell (membrane electrode, gas diffusion layer, cooling channel interface) to ensure that the sensor fits tightly with the target area to achieve accurate measurement. The temperature sensor is initialized and calibrated to ensure its measurement accuracy. At the same time, the connection status of the sensor is checked to ensure that the data transmission channel is unobstructed to measure the temperature of the target area. Among them, the initialization calibration includes zero point calibration, gain calibration and connection status check. Zero point calibration is to adjust the temperature sensor output to the theoretical zero point value under a known reference temperature environment to eliminate static errors. Gain calibration is to correct the temperature sensor output slope through multi-temperature point calibration. To ensure full-scale measurement accuracy, the connection status check verifies the physical connection and electrical parameters between the temperature sensor and the communication interface, and ensures that there is no open circuit or short circuit in the data transmission channel. The sensor collects temperature data of key parts inside the hydrogen fuel cell in real time, and transmits it to the thermal management platform in real time through the high-bandwidth communication interface (CAN bus). The collected temperature data is preprocessed, including filtering to eliminate noise interference, normalization to unify the dimension, and timestamp marking to ensure data timing. A first-order low-pass filtering algorithm (the cutoff frequency is set according to the dynamic response requirements of the system) is used to eliminate high-frequency noise while retaining the temperature change trend. The output voltage or resistance value of the temperature sensor is converted to a standard temperature value and the dimension is unified for subsequent analysis. A high-precision timestamp is added to each data to ensure data timing and multi-sensor synchronization. The preprocessed temperature data is then stored in a temporary buffer for temporary storage.

[0056] The temperature prediction unit, based on the data collected by the temperature sensor, combines the thermal model and historical operation data of the fuel cell, uses the long short-term memory network to predict the future trend of the temperature change inside the fuel cell, predicts the temperature change in advance, provides forward-looking guidance for the dynamic adjustment of the heating power, reduces temperature fluctuations, receives the temperature data of the key parts inside the hydrogen fuel cell collected by the temperature sensor in real time, and simultaneously loads the fuel cell thermal model containing heat capacity, thermal resistance and heat transfer coefficient parameters and the historical operation database covering the temperature-power mapping relationship under different working conditions. Through the timestamp alignment technology, the timestamp of the sensor data is mapped to the thermal model calculation results and the historical data records into a unified time series coordinate system, and the multi-dimensional feature vector is extracted, including the current Temperature gradient, historical temperature fluctuation rate, power load change rate and ambient temperature correction term are used to construct a feature matrix to characterize the dynamic characteristics of the system. The rows correspond to the time steps and the columns correspond to the feature dimensions. Among them, the current temperature gradient is the first / second order time derivative of the real-time temperature of the key parts, which characterizes the transient thermal change rate. The historical fluctuation rate is based on the standard deviation or range of the temperature in the past 30 seconds based on the sliding window statistics, which quantifies the thermal stability. The power load change rate is to analyze the slope of the battery stack output power and the associated heat generation rate. The environmental correction term is to introduce external variables such as ambient temperature and humidity, and adjust the feature weight through the thermal model correction factor. The feature matrix is ​​input into the temperature prediction model trained based on the long short-term memory network model as the basic architecture. The long-term dependency and short-term fluctuation pattern in the time series are captured through multiple layers of gated units, and the future temperature changes are predicted in combination with the constraint of the maximum allowable temperature rise rate in the thermal model. The future temperature change prediction adopts a sliding window mechanism, which takes the current moment as the starting point to generate a temperature trend curve within the next 5-10 seconds, and outputs structured data containing the predicted value and its confidence interval. The model architecture includes: input layer, hidden layer and fully connected layer. The input layer is a standardized feature matrix (mean is 0, variance is 1), and the hidden layer is a two-layer LSTM unit (64 / 128 neurons per layer). The long-term dependency (such as thermal inertia effect) and short-term fluctuation (such as power mutation) are captured through the forget gate, input gate and output gate. The fully connected layer The LSTM output is mapped to the temperature prediction value for the next 5-10 seconds. Dropout (0.2) is used to prevent overfitting. The temperature prediction results output by the temperature prediction model are verified to check whether the prediction results meet the thermal model constraints. The predicted temperature trend is compared with historical data under similar operating conditions to evaluate the deviation range. The prediction results that pass the verification are marked as valid and transmitted to the thermal management control module for dynamic adjustment of the heating power strategy to reduce the temperature fluctuation amplitude. The predicted temperature trend is aligned with the historical data under similar operating conditions through the dynamic time warping algorithm. The Euclidean distance deviation threshold is calculated. The comprehensive verification results are used to generate a confidence score. If it is below the threshold, model retraining or data supplementation is triggered.

[0057] A thermal management control module is used to dynamically adjust the heating system and cooling system to perform heating or cooling operations based on the output of the temperature monitoring and prediction module. The thermal management control module includes a heating power adjustment unit and a cooling system adjustment unit;

[0058] Among them, the heating power adjustment unit is used to dynamically adjust the power output of the heating system according to the temperature prediction results, to ensure that the fuel cell quickly heats up to the appropriate working temperature at the initial stage of startup, to avoid startup difficulties caused by local overcooling, and at the same time to reduce energy waste and improve energy utilization efficiency. It receives the output data of the temperature monitoring and prediction module, including the trend curve of the internal temperature of the fuel cell in the next 5-10 seconds and its confidence interval, and parses the output data of the temperature monitoring and prediction module to extract the slope and extreme points of the predicted temperature curve. Combined with the thermal model constraints, it calculates the remaining temperature difference and time window from the current to the target temperature. The slope of the predicted temperature curve is the heating rate, which characterizes the current heating trend. The extreme point is the lowest temperature moment, and the local overcooling risk area is identified. The thermal model constraint is that the maximum allowable temperature rise rate is ≤5℃ / s. Based on the residual temperature difference and the time window, the target heating trajectory is generated by the PID control parameter dynamic tuning algorithm to ensure that the fuel cell quickly and smoothly heats up to the appropriate working temperature at the initial stage of startup, to avoid local overcooling due to slow heating or thermal stress damage to the material due to fast heating. Based on the target heating trajectory, the regulation unit dynamically adjusts the heating system power output through a closed-loop control strategy. A multi-zone coordinated control algorithm is used to allocate total heating power to key components of the hydrogen fuel cell on demand. Based on the deviation between each zone's real-time temperature and the target trajectory, the power allocation weights are optimized using a weighted least squares method to ensure that the temperature difference between each zone is ≤2°C, ensuring temperature synchronization. Based on the real-time deviation between the predicted and actual temperatures, an adaptive PID algorithm dynamically adjusts the heating power to precisely match the heating rate with the target trajectory. Combining thermal models with temperature sensor data, multi-level constraints are imposed on the heating power to avoid energy waste or localized overheating due to power overload. The heating power regulation effect is evaluated from multiple dimensions, calculating the integral of the deviation between the actual and target heating trajectories (root mean square error (RMSE)) to quantify heating efficiency and energy utilization efficiency. If the evaluation results indicate a deviation exceeding a preset threshold, control parameters are optimized to improve regulation accuracy. If a heating system anomaly is detected (i.e., a power output deviation >10% from the command), a redundant heating module switchover or safety shutdown mechanism is triggered to ensure system reliability.

[0059] The cooling system adjustment unit is used to automatically enable the cooling system and adjust the cooling parameters of the cooling system to reduce the temperature during the operation of the fuel cell according to the temperature prediction results to prevent overheating from causing damage to the battery, ensure uniform temperature distribution inside the fuel cell, extend battery life, and improve system reliability. It receives the temperature trend curve and confidence interval of the next 5-10 seconds output by the temperature monitoring and prediction module, extracts the heating rate and extreme points, among which, for the heating rate, the slope of the temperature curve is calculated by numerical differentiation to determine whether the cooling threshold is triggered (temperature rise rate > 3°C / s). For the extreme points, the highest temperature moment in the temperature curve is located to identify the potential overheating risk area. Combined with the thermal model constraints, it is determined whether the cooling system needs to be enabled. If the predicted temperature exceeds the safety threshold of 75°C or the temperature rise rate exceeds the limit, the cooling system is immediately activated, and then the remaining overtemperature and time window from the current to the target temperature are calculated. Based on the fuzzy logic algorithm, the overtemperature and time window are mapped to the cooling demand level to determine the initial parameters of the cooling system, including fan speed, water pump flow and coolant distribution, using adaptive PI The D control algorithm dynamically adjusts cooling parameters based on the real-time deviation between the predicted and actual temperatures. For fan speed, the proportional term quickly responds to deviations and suppresses temperature overshoot. For water pump flow, the integral term eliminates steady-state errors to ensure long-term cooling efficiency. For coolant distribution, the differential term predicts the deviation trend and optimizes the coolant flow distribution in each region. Combined with multi-region temperature sensor data, regional temperature difference constraints and cooling efficiency constraints are imposed on the cooling parameters. The regional temperature difference constraint ensures that the maximum temperature difference between each region is ≤3°C to avoid local thermal stress concentration. The cooling efficiency constraint limits the difference between the coolant temperature and the battery temperature to ≥10°C to ensure heat exchange efficiency. The root mean square error (RMSE) is calculated as the integral of the deviation between the actual and target temperature trajectories. Cooling efficiency and energy consumption indicators are analyzed. For the cooling efficiency indicator, an RMSE ≤2°C indicates efficient cooling. For the energy consumption indicator, an energy consumption per unit temperature difference of ≤0.05 kWh / °C indicates energy-saving operation. If the evaluation result shows that the deviation exceeds a threshold, that is, an RMSE >3°C, the strategy optimization mechanism is triggered.

[0060] The temperature response adjustment module uses an adaptive control response algorithm to monitor and quickly adjust temperature changes in real time, shortening the system's response time to temperature changes and ensuring that heating or cooling strategies can be adjusted in a timely manner when the temperature suddenly changes to maintain battery performance.

[0061] The actuator response module is used to optimize the response speed of the actuators in the heating and cooling systems, adjust the working state after receiving the control signal, reduce the temperature fluctuation caused by the actuator response delay, and improve the overall performance of the system.

[0062] Example 2, as Figure 1 、 Figure 2As shown, based on Example 1, the present invention provides a technical solution: preferably, the temperature response adjustment module specifically includes:

[0063] The real-time data of multi-region temperature sensors are integrated and combined with the trend curve of the next 5-10 seconds output by the temperature prediction module to build a dynamic temperature field model. The sliding window filtering algorithm is used to eliminate sensor noise and extract temperature change characteristics including heating rate and extreme points. The temperature mutation is determined based on the dynamic threshold method, which is divided into heating rate mutation and extreme point offset. Among them, for heating rate mutation, if the temperature change rate exceeds the historical mean ±2 times the standard deviation, it is determined to be rapid heating or cooling. For extreme point offset, if the deviation between the predicted temperature extreme point and the target trajectory exceeds the preset offset threshold, the strategy adjustment is triggered, and the mutation signal is estimated through the Kalman filter to filter out false alarms. The control mode is automatically switched according to the type of temperature mutation of heating / cooling. For sudden changes, the rapid heating strategy is activated, and the power allocation weight of the heating system is prioritized. For sudden changes in temperature, the enhanced cooling strategy is activated, and the fan speed and water pump flow are prioritized, and the coolant distribution ratio is optimized. The fuzzy PID algorithm is used to dynamically adjust the control parameters, including proportional gain, integral time, and differential time. For the proportional gain, linear mapping is performed according to the sudden change amplitude to accelerate the response. For the integral time, dynamic adjustment is made according to the system inertia to suppress overshoot. For the differential time, adaptive changes are made according to the sudden change rate. The temperature change trend is predicted. Based on the integral of the deviation between the predicted temperature and the actual temperature, the control parameters are dynamically corrected through the gradient descent method. Combined with the regional temperature synchronization index, that is, the temperature difference between each region is ≤2℃, the heating / cooling resource allocation is optimized through the quadratic programming algorithm.

[0064] In addition, the specific process of optimizing heating / cooling resource allocation is:

[0065] The deviation between the predicted temperature and the actual temperature is continuously calculated and the calculated deviation is integrated to obtain the deviation integral to quantify the cumulative error. The deviation integral reflects the overall error degree in the temperature control process. An objective function is constructed based on the deviation integral to evaluate the difference between the control parameter and the actual temperature. The control parameter is iteratively optimized using the gradient descent method. The gradient of the objective function with respect to the current control parameter is calculated to determine the parameter adjustment direction. The control parameter is adjusted along the gradient direction to gradually reduce the deviation integral and improve the control accuracy. The adjusted temperature change is monitored in real time to ensure that the parameter correction is effective and the system is stable. The regional temperature synchronization index, that is, the temperature difference between each region is ≤2°C, is introduced as a constraint condition. At the same time, the allocation efficiency of heating / cooling resources is used as the objective function to construct a quadratic programming model. The goal of the model is to optimize resource allocation to ensure that the temperature of each region is uniform and consistent with the target temperature. The quadratic programming algorithm is used to solve the quadratic programming model and calculate the optimal heating / cooling resource allocation plan. Based on the solution, the heating power or cooling flow is allocated to each region to synchronize the temperature of each region and maintain the uniformity of the temperature field inside the fuel cell.

[0066] The actuator response module specifically includes:

[0067] The actuator response module receives the control signal from the thermal management control module in real time, including the power adjustment instruction of the heating or cooling system, and pre-processes the received control signal, including signal analysis and filtering to remove possible noise interference to ensure the accuracy and reliability of the signal. At the same time, it checks the integrity and validity of the signal and confirms whether the control signal conforms to the expected instruction format and range. Among them, if the instruction fluctuation is <1% within 5 consecutive sampling cycles, the current instruction value is locked, and the multi-source signal is synchronized based on the system clock to avoid instruction conflicts. If the temperature change rate is >2℃ / s, the heating / cooling intensity instruction is responded to first. If the regional temperature difference is >1.5℃, the resource allocation instruction is responded to first. The processed control signal is converted into a specific drive signal to adjust the working state of the actuator of the heating and cooling system. For the execution of the heating system For the actuator, rapid heating is achieved by changing the power output of the heating element. For the actuator of the cooling system, rapid cooling is achieved by adjusting the fan speed and water pump flow parameters. For the control of the fan speed, the target speed is converted into the motor drive voltage through the PID regulator. For the water pump flow control, the target flow is mapped to the water pump duty cycle by the lookup table method, and pressure feedforward compensation is introduced to continuously monitor the actual working status of the actuator, including power output, speed and flow parameters. The status information is fed back to the thermal management control module in real time through the feedback loop, and the difference between the actual state and the target state is compared. The drive parameters of the actuator are dynamically optimized using the adaptive control algorithm to ensure the response speed and accuracy of the actuator. At the same time, the operating status of the actuator is monitored in real time, and any faults or abnormal conditions are detected and handled to ensure the long-term stable operation of the actuator.

[0068] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A hydrogen fuel cell vehicle ultra-low temperature cold start thermal management system, including a thermal management platform, characterized by: The thermal management platform is communicatively connected to the following modules: The temperature monitoring and prediction module is used to monitor the temperature of key parts inside the hydrogen fuel cell using temperature sensors and, combined with the temperature prediction model, predict the temperature change trend inside the fuel cell; The thermal management control module is used to dynamically adjust the heating system and cooling system to perform heating or cooling operations based on the output of the temperature monitoring and prediction module; The temperature response adjustment module is used to adopt an adaptive control response algorithm to monitor and quickly adjust temperature changes in real time, thereby shortening the system's response time to temperature changes; The actuator response module is used to optimize the response speed of the actuators of the heating and cooling systems and adjust the working state after receiving the control signal.

2. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 1, characterized in that: The temperature monitoring and prediction module includes a real-time temperature monitoring unit and a temperature prediction unit; The real-time temperature monitoring unit is used to collect the temperature of key parts inside the hydrogen fuel cell using a temperature sensor to obtain real-time temperature data; The temperature prediction unit uses a long short-term memory network to predict the future temperature change trend of the fuel cell based on the data collected by the temperature sensor, combined with the thermal model and historical operation data of the fuel cell.

3. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 2, characterized in that: The real-time temperature monitoring unit specifically includes: High-precision temperature sensors are distributed in key locations of the hydrogen fuel cell, and initial calibration is performed on the temperature sensors. At the same time, the sensor connection status is checked to measure the temperature of the target area. Initial calibration includes zero point calibration, gain calibration, and connection status check. The sensor collects temperature data from key parts inside the hydrogen fuel cell in real time, transmits it to the thermal management platform in real time through a high-bandwidth communication interface, and preprocesses the collected temperature data, including filtering to eliminate noise interference, normalization to unify the dimension, and timestamp marking, while retaining the temperature change trend, converting the temperature sensor output voltage or resistance value into a standard temperature value, adding a timestamp to each data, and then storing the preprocessed temperature data in a temporary buffer for temporary storage.

4. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 2, characterized in that: The temperature prediction unit specifically includes: Receive temperature data from key parts of the hydrogen fuel cell collected in real time by temperature sensors, and simultaneously load the fuel cell thermal model containing heat capacity, thermal resistance, and heat transfer coefficient parameters, as well as a historical operation database covering temperature-power mapping relationships under different operating conditions. Using timestamp alignment technology, map the timestamps of the sensor data, the thermal model calculation results, and historical data records to a unified time series coordinate system, extract multidimensional feature vectors, including the current temperature gradient, historical temperature fluctuation rate, power load change rate, and ambient temperature correction terms, and construct a feature matrix to characterize the dynamic characteristics of the system; The feature matrix is ​​input into a temperature prediction model trained based on a long-short-term memory network model. Multi-layer gated units are used to capture long-term dependencies and short-term fluctuation patterns in the time series. Future temperature changes are predicted based on the maximum allowable temperature rise rate constraints in the thermal model. This prediction uses a sliding window mechanism, starting from the current moment, to generate a temperature trend curve for the next 5-10 seconds. The output is structured data containing the predicted value and its confidence interval. The temperature prediction results output by the temperature prediction model are verified to check whether the prediction results meet the thermal model constraints, and the predicted temperature trends are compared with historical data under similar working conditions to evaluate the deviation range. The prediction results that pass the verification are marked as valid and transmitted to the thermal management control module.

5. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 2, characterized in that: The thermal management control module includes a heating power adjustment unit and a cooling system adjustment unit; Wherein, the heating power adjustment unit is used to dynamically adjust the power output of the heating system according to the temperature prediction result; The cooling system regulating unit is used to automatically activate the cooling system and adjust the cooling parameters of the cooling system to reduce the temperature according to the temperature prediction result during the operation of the fuel cell.

6. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 5, characterized in that: The heating power adjustment unit specifically includes: Receive the output data of the temperature monitoring and prediction module, including the trend curve of the fuel cell internal temperature change in the next 5-10 seconds and its confidence interval, and parse the output data of the temperature monitoring and prediction module to extract the slope and extreme points of the predicted temperature curve. Combined with the thermal model constraints, calculate the remaining temperature difference and time window from the current to the target temperature. The slope of the predicted temperature curve is the heating rate, which represents the current heating trend. The extreme point is the moment of lowest temperature, which identifies the local overcooling risk area. The thermal model constraint is that the maximum allowable temperature rise rate is ≤5℃ / s. Based on the target temperature trajectory, the regulation unit dynamically adjusts the heating system power output through a closed-loop control strategy. A multi-zone collaborative control algorithm is used to distribute the total heating power to key parts of the hydrogen fuel cell as needed. Based on the deviation between the real-time temperature of each zone and the target trajectory, the power allocation weights are optimized using a weighted least squares method to ensure that the temperature difference between each zone is ≤2°C. The heating power regulation effect is evaluated in multiple dimensions, the deviation integral between the actual heating trajectory and the target trajectory is calculated, and the heating efficiency and energy utilization efficiency are quantified. If the evaluation result shows that the deviation exceeds the preset deviation threshold, the control parameters are optimized. If a heating system abnormality is detected, that is, the power output deviation from the command is greater than 10%, the redundant heating module switching or safety shutdown mechanism is triggered.

7. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 5, characterized in that: The cooling system adjustment unit specifically includes: The system receives the temperature trend curve and confidence interval for the next 5-10 seconds output by the temperature monitoring and prediction module, extracts the heating rate and extreme points, and uses numerical differentiation to calculate the slope of the temperature curve to determine whether the cooling threshold has been triggered. For the extreme points, the highest temperature moment in the temperature curve is located to identify potential overheating risk areas. Based on the thermal model constraints, the system determines whether the cooling system needs to be activated. If the predicted temperature exceeds the safety threshold of 75°C or the temperature rise rate exceeds the limit, the cooling system is immediately activated. The remaining overtemperature and time window from the current temperature to the target temperature are calculated, and the initial parameters of the cooling system, including fan speed, water pump flow, and coolant distribution, are determined. An adaptive PID control algorithm is used to dynamically adjust cooling parameters based on the real-time deviation between the predicted and actual temperatures. For fan speed, the proportional term quickly responds to deviations. For water pump flow, the integral term eliminates steady-state errors. For coolant distribution, the differential term predicts deviation trends. Combined with multi-region temperature sensor data, regional temperature difference constraints and cooling efficiency constraints are imposed on cooling parameters. The regional temperature difference constraint ensures that the maximum temperature difference between each region is ≤3°C. The cooling efficiency constraint limits the difference between the coolant temperature and the battery temperature to ≥10°C. The integral of the deviation between the actual temperature trajectory and the target temperature trajectory, namely the root mean square error (RMSE), is calculated, and the cooling efficiency index and energy consumption index are analyzed. For the cooling efficiency index, when RMSE ≤ 2°C, it is judged as high-efficiency cooling. For the energy consumption index, when the cooling energy consumption per unit temperature difference ≤ 0.05kWh / °C, it is judged as energy-saving operation. If the evaluation result shows that the deviation exceeds the threshold, that is, RMSE > 3°C, the strategy optimization mechanism is triggered.

8. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 7, characterized in that: The temperature response adjustment module specifically includes: Integrate real-time data from multi-region temperature sensors and combine it with the trend curve for the next 5-10 seconds output by the temperature prediction module to build a dynamic temperature field model. Use a sliding window filtering algorithm to eliminate sensor noise and extract temperature change characteristics including heating rate and extreme points. Temperature mutations are determined based on a dynamic threshold method, which is divided into heating rate mutations and extreme point shifts. For heating rate mutations, if the temperature change rate exceeds ±2 standard deviations of the historical mean, it is determined to be rapid heating or cooling. For extreme point shifts, if the deviation between the predicted temperature extreme point and the target trajectory exceeds the preset offset threshold, a strategy adjustment is triggered. Automatically switches control modes based on the type of sudden temperature rise / fall. For sudden temperature rises, the rapid heating strategy is activated, prioritizing the increase in heating system power allocation weight. For sudden temperature falls, the enhanced cooling strategy is activated, prioritizing the increase in fan speed and water pump flow, and optimizing the coolant distribution ratio. The fuzzy PID algorithm is used to dynamically adjust control parameters, including proportional gain, integral time, and differential time. Based on the integral of the deviation between the predicted temperature and the actual temperature, the control parameters are dynamically corrected through the gradient descent method. Combined with the regional temperature synchronization index, that is, the temperature difference between each area is ≤2℃, the heating / cooling resource allocation is optimized through the quadratic programming algorithm.

9. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 8, characterized in that: The specific process of optimizing heating / cooling resource allocation is as follows: Continuously calculate the deviation between the predicted temperature and the actual temperature, and integrate the calculated deviation to obtain the deviation integral to quantify the cumulative error; An objective function is constructed based on the deviation integral to evaluate the difference between the control parameter and the actual temperature. The control parameter is iteratively optimized using the gradient descent method. The gradient of the objective function with respect to the current control parameter is calculated, the parameter adjustment direction is determined, and the control parameter is adjusted along the gradient direction to gradually reduce the deviation integral. The temperature change after adjustment is monitored in real time. The regional temperature synchronization index, i.e., the temperature difference between regions is ≤ 2°C, is introduced as a constraint condition. At the same time, the allocation efficiency of heating / cooling resources is used as the objective function to construct a quadratic programming model. The quadratic programming algorithm is used to solve the quadratic programming model and calculate the optimal heating / cooling resource allocation plan. Based on the solution, the heating power or cooling flow is allocated to each zone to ensure that the temperature of each zone is adjusted synchronously.

10. The ultra-low temperature cold start thermal management system for a hydrogen fuel cell vehicle according to claim 9, characterized in that: The actuator response module specifically includes: The actuator response module receives control signals from the thermal management control module in real time, including power adjustment instructions for the heating or cooling system, and pre-processes the received control signals. At the same time, it checks the integrity and validity of the signals and confirms whether the control signals conform to the expected instruction format and range. Among them, if the instruction fluctuation is less than 1% within 5 consecutive sampling periods, the current instruction value is locked. If the temperature change rate is greater than 2°C / s, the heating / cooling intensity instruction is responded to first. If the regional temperature difference is greater than 1.5°C, the resource allocation instruction is responded to first. The processed control signal is converted into a specific drive signal to adjust the working state of the actuators of the heating and cooling systems. For the actuators of the heating system, rapid temperature increase is achieved by changing the power output of the heating element. For the actuators of the cooling system, rapid temperature reduction is achieved by adjusting the fan speed and water pump flow parameters. Continuously monitor the actual working status of the actuator, including power output, speed and flow parameters, and feed the status information back to the thermal management control module in real time through the feedback loop. Compare the difference between the actual status and the target status, and use the adaptive control algorithm to dynamically optimize the actuator drive parameters. At the same time, monitor the operating status of the actuator in real time to detect and handle any faults or abnormal conditions.

Images