Power distribution and drive control system of fuel cell heavy truck

Abstract

The invention discloses a power distribution and drive control system of a fuel cell heavy truck in the technical field of fuel cell electric vehicle energy management, which comprises a parameter acquisition module used for acquiring vehicle driving parameters, equipment operation parameters and map environment parameters; the power prediction module is used for predicting a future short-term power demand by using a neural network; the fuel cell optimization module is used for establishing an efficiency mapping model and a service life mapping model of the fuel cell and searching an optimal working point of the fuel cell; the power distribution module is used for establishing a power battery power distribution strategy; and the output module is used for controlling power output and speed change of the corresponding motor. The method is fast and effective, can perform accurate short-term prediction on the required power of the motor by using the memory neural network, finds the optimal working point of the battery by establishing the battery efficiency-service life joint optimization model, realizes reliable dynamic tracking and reasonable distribution of power requirements, improves the energy utilization rate, prolongs the service life of the fuel battery, and improves the reliability of the system. And the maintenance cost of the fuel cell is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy management for fuel cell electric vehicles, and specifically relates to a power distribution and drive control system for fuel cell heavy trucks. Background Art

[0002] As an energy converter that converts chemical energy into electrical energy, a fuel cell has the advantages of high efficiency, zero emissions, and energy conservation, and is an ideal clean energy source. However, when it is used as the sole energy source for a vehicle, due to the drastic changes in the operating conditions of the fuel cell stack, its performance deteriorates rapidly, resulting in unstable output power. In addition, the fuel cell cannot recover braking energy, and the lifespan of existing fuel cells is not ideal due to the drastic changes in operating conditions. To solve these problems, a hybrid drive of a fuel cell and an auxiliary energy source (power battery or supercapacitor) is generally used as the energy source mode for the vehicle. Under the premise of multiple energy sources, the vehicle needs a system to reasonably distribute the energy output between the fuel cell and other power sources to achieve complementary advantages and improve the operating stability and economic benefits of the vehicle.

[0003] The prior art has introduced an automatic control system that calculates the motor demand power based on the throttle, brake pedal, and vehicle speed data, uses the power battery to compensate for the output power fluctuations of the fuel cell in a peak shaving and valley filling manner, and uses a gradient algorithm based on deep learning to distribute the fuel cell power and the power battery power, so as to achieve stable power output of the vehicle and balance in economy, as shown in the patents with patent publication number CN114290916B and publication number CN119037243A.

[0004] Although the above patents have basically achieved stable power output of the fuel cell vehicle and reasonably distributed the energy output of different energy sources based on the power demand prediction results to improve the economic benefits of the fuel cell vehicle, they fail to consider the characteristics of the fuel cell itself as an energy converter. When the fuel cell is under low load or high load for a long time, its lifespan will decrease rapidly, and there are far more factors that change the power demand than the driver's driving control. This will cause the system to be difficult to accurately and dynamically predict the actual motor output power required, and may even cause fluctuations in the output power of the fuel cell due to the misleading of the prediction algorithm, affecting normal driving operations and battery lifespan.

[0005] Therefore, it is necessary to propose a method that can collect multiple parameters and use a memory neural network to accurately predict the motor demand power in the short term, establish an efficiency-life joint optimization model for the fuel cell, find the optimal operating point of the fuel cell, and generate a power distribution strategy for the battery to achieve reliable dynamic tracking and reasonable distribution of power demand, improve energy utilization efficiency, extend the lifespan of the fuel cell, and reduce the maintenance cost of the fuel cell. Summary of the Invention

[0006] To solve the above problems, the object of the present invention is to provide a power distribution and drive control system for a fuel cell heavy truck, which collects vehicle load parameters through a variety of sensors, introduces a bidirectional long short-term memory network with an attention mechanism to capture acceleration characteristics, predicts short-term power demand changes, establishes an efficiency-life joint optimization model of the fuel cell, finds the optimal operating point of the fuel cell, conducts power distribution and automatic control, extends the life of the fuel cell on the premise of the stable operation of the whole vehicle, optimizes the power output of multiple batteries, and reduces operation and maintenance costs.

[0007] To achieve the above object, the technical solution of the present invention is as follows: The power distribution and drive control system of a fuel cell heavy truck includes a parameter acquisition module, a power prediction module, a fuel cell optimization module, a power distribution module, and an output module.

[0008] The parameter acquisition module is used to collect vehicle driving parameters, equipment operation parameters, and map environment parameters, and perform data enhancement processing.

[0009] The power prediction module is used to predict future short-term power demand according to the enhanced parameters using a trained neural network.

[0010] The fuel cell optimization module is used to establish an efficiency mapping model and a life mapping model of the fuel cell based on a multi-objective optimization framework, and find the optimal operating point of the fuel cell.

[0011] The power distribution module is used to establish a power battery power distribution strategy based on the predicted short-term power demand and the optimal operating point of the fuel cell to compensate for the power fluctuation of the fuel cell.

[0012] The output module is used to control the power output of the corresponding motor based on the optimal operating point of the fuel cell and the power battery power distribution strategy, and control the AMT transmission to perform automatic shifting.

[0013] The principle of the basic solution is as follows: The parameter acquisition module is responsible for collecting vehicle driving parameters (such as accelerator pedal position, brake pedal position, vehicle speed, etc.), equipment operation parameters (such as fuel cell voltage, current, temperature, power battery SOC, temperature, etc.), and map environment parameters (such as slope, curvature, speed limit, etc.). The collected data is subjected to enhancement processing to improve the diversity and quality of the data, providing a reliable data basis for subsequent prediction and optimization.

[0014] The power prediction module uses a trained neural network (such as a bidirectional long short-term memory network with an attention mechanism) to analyze the enhanced parameters, capture acceleration characteristics, and predict future short-term power demand changes. This prediction helps to plan the power output of the fuel cell and the power battery in advance to meet the power demand of the vehicle.

[0015] Based on a multi-objective optimization framework, the fuel cell optimization module establishes an efficiency mapping model and a life mapping model for the fuel cell. Through these two models, the efficiency and life of the fuel cell at different operating points can be evaluated, so as to find the optimal operating point and achieve the joint optimization of efficiency and life.

[0016] According to the predicted short-term power demand and the optimal operating point of the fuel cell, the power distribution module formulates a power distribution strategy for the power battery. The power battery is used to compensate for the power fluctuations of the fuel cell, ensuring that the vehicle optimizes the power output of multiple batteries on the premise of stable operation.

[0017] According to the optimal operating point of the fuel cell and the power distribution strategy of the power battery, the output module controls the power output of the corresponding motor. At the same time, it controls the AMT transmission to perform automatic shifting to adapt to different driving conditions and power demands.

[0018] The beneficial effects of the basic solution are as follows: 1. By accurately predicting the short-term power demand, the system can plan the power output of the fuel cell and the power battery in advance, ensuring that the vehicle can obtain sufficient power support under various driving conditions. The power distribution strategy of the power battery helps to compensate for the power fluctuations of the fuel cell, improving the stability and response speed of the vehicle.

[0019] 2. The fuel cell optimization module realizes the joint optimization of the efficiency and life of the fuel cell by finding the optimal operating point. This helps to reduce the loss and failure rate of the fuel cell, extend its service life, and thus reduce the operation and maintenance costs of the vehicle.

[0020] 3. By optimizing the power output of multiple batteries, the system can improve the energy utilization efficiency of the vehicle, reduce energy consumption and emissions. At the same time, the extended life of the fuel cell also reduces the frequency and cost of replacing the fuel cell.

[0021] 4. The system adopts advanced neural network and multi-objective optimization technologies to realize the intelligent management of the vehicle power distribution and drive control. This helps to improve the automation level and driving comfort of the vehicle and reduce the labor intensity of the driver.

[0022] Furthermore, the parameter acquisition module includes a nine-axis gyroscope, a wheel speed sensor, a suspension pressure sensor, a wind speed sensor, a fuel cell voltage inspection circuit, and a battery pack temperature control sensor.

[0023] The nine-axis gyroscope is used to detect the driving attitude and direction of the vehicle.

[0024] The wheel speed sensor is used to detect the sequential vehicle speed of the vehicle.

[0025] The suspension pressure sensor is used to detect the suspension pressure of the vehicle.

[0026] A wind speed sensor for detecting the wind direction and wind speed of the vehicle when it is affected by the wind;

[0027] Fuel cell voltage inspection for detecting changes in the output voltage of the fuel cell;

[0028] A battery pack temperature control sensor for detecting changes in the temperature inside the fuel cell stack.

[0029] The beneficial effects of the basic solution are as follows: 1. Through the collaborative work of multiple sensors, the system can comprehensively and accurately collect various key parameters during vehicle driving, including driving attitude, direction, vehicle speed, suspension pressure, wind direction and speed, fuel cell output voltage, and battery pack temperature, etc. These parameters provide a solid data foundation for subsequent power prediction, fuel cell optimization, and power distribution, and help improve the decision-making accuracy and reliability of the system.

[0030] 2. The data of the nine-axis gyroscope and wheel speed sensor can be used for vehicle stability control to help the system promptly identify and respond to potential driving risks, such as skidding and slipping. The data of the suspension pressure sensor helps the system understand the vehicle's load condition and driving road conditions, so as to adjust the driving strategy and ensure driving stability and comfort.

[0031] 3. The data of the wind speed sensor can be used to evaluate the wind resistance situation when the vehicle is driving, providing a reference for the system to adjust the power output to reduce energy consumption. The data of the fuel cell voltage inspection and battery pack temperature control sensor help the system monitor the working status of the fuel cell and power battery in real time, ensure that they operate within the optimal working range, and improve energy utilization efficiency.

[0032] 4. By monitoring the working status of the fuel cell and power battery in real time, the system can promptly detect and handle potential fault hazards, and avoid component damage caused by overload, overheating, etc. At the same time, the fuel cell optimization module based on the multi-objective optimization framework can find the optimal working point of the fuel cell, further reduce its losses and failure rate, and extend its service life.

[0033] Furthermore, the vehicle driving parameters include the vehicle accelerator pedal travel, vehicle braking pressure, and suspension pressure; the equipment operation parameters include the sequential vehicle speed, fuel cell voltage, power battery voltage, and battery pack temperature; the map environment parameters include map navigation data, slope, and wind speed.

[0034] The beneficial effects of the basic solution are as follows: 1. By collecting driving parameters such as the vehicle's accelerator pedal travel, vehicle braking pressure, and suspension pressure, the system can accurately perceive the driver's intention and the current driving state of the vehicle. This helps the system more accurately predict future power demands and adjust the power outputs of the fuel cell and power battery accordingly. The acquisition of sequential vehicle speed provides continuous information on the vehicle's driving speed, which serves as an important basis for the system to analyze the vehicle's driving conditions and formulate driving strategies.

[0035] 2. The acquisition of device operation parameters such as fuel cell voltage, power battery voltage, and battery pack temperature enables the system to monitor the working states of the fuel cell and power battery in real time. This helps the system promptly detect and handle potential fault hazards, ensuring that the battery pack operates within the optimal working range, improving energy utilization efficiency and battery life. At the same time, these parameters also provide the necessary data support for the fuel cell optimization module, facilitating the finding of the optimal operating point of the fuel cell and achieving the joint optimization of efficiency and life.

[0036] 3. The acquisition of map environment parameters such as map navigation data, slope, and wind speed enables the system to understand in advance the route and environmental conditions that the vehicle is about to travel. This helps the system adjust driving strategies according to road conditions and environmental changes, such as increasing power output in advance when going uphill and adjusting the wind resistance control strategy when the wind speed is high. By intelligently adapting to environmental changes, the system can further improve the performance and economy of the entire vehicle, reducing energy consumption and emissions.

[0037] Furthermore, data augmentation processing is to add Gaussian noise to simulate sensor errors with a standard deviation σ = 5%, and perform time-series sliding window processing on the collected parameters with a window length of 5 s and a step size of 0.1 s.

[0038] The beneficial effects of the basic solution are as follows: 1. By adding Gaussian noise to simulate sensor errors, the system can more realistically reflect the data fluctuations of sensors during actual operation. This helps the system better cope with sensor errors in subsequent processing and analysis, improving the robustness and reliability of the data. This data augmentation processing enables the system to maintain a high prediction accuracy and decision-making accuracy when facing sensor errors during actual operation.

[0039] 2. Performing time-series sliding window processing on the collected parameters enables the system to capture the changing trends and periodic characteristics of the parameters over time. This helps the system more accurately predict future power demands and formulate more reasonable driving strategies accordingly. Time-series sliding window processing also provides rich historical data, offering more comprehensive data support for the multi-objective optimization and power distribution of the system.

[0040] 3. Through data augmentation processing, the system can learn more diverse data patterns, thereby improving the prediction accuracy of future power demands. This helps the system more accurately plan the power outputs of fuel cells and power batteries to meet the vehicle's power demands.

[0041] 4. Data augmentation processing enables the system to face more diverse data situations, enhancing the system's adaptability and generalization ability. This helps the system maintain high performance and stability in different vehicle models, different application scenarios, and different environmental conditions. The improvement of the system's adaptability and generalization ability also reduces the dependence on specific sensors or specific data sets, improving the system's versatility and portability.

[0042] Furthermore, the neural network is a bidirectional long short-term memory network with an attention mechanism, including a forward layer, a backward layer, and a hidden layer. The forward layer captures the acceleration trend features, the backward layer identifies the historical steady segments, and the attention mechanism weights the slope changes in the short-term future to predict the short-term power demand changes.

[0043] The beneficial effects of the basic solution are as follows: 1. By capturing the acceleration trend features, the forward layer can reflect the acceleration intention and power demand changes during the vehicle's driving in real time, providing an accurate basis for the system's short-term power prediction. The backward layer, on the other hand, helps the system understand the stable state during the vehicle's driving by identifying the historical steady segments, thereby more accurately predicting the steady part of the future power demand and reducing unnecessary power fluctuations.

[0044] 2. The attention mechanism can automatically learn and weight the impact of slope changes in the short-term future on power demands, enabling the system to more intelligently adjust the power output strategy. When facing complex road conditions, such as continuous uphill or downhill, the attention mechanism can perceive and adjust the power distribution in advance to ensure the vehicle runs smoothly and efficiently.

[0045] 3. The combination of the bidirectional long short-term memory network and the attention mechanism can capture the features in time series data more comprehensively, improving the prediction accuracy of short-term power demands. At the same time, this network structure also speeds up the system's response speed, enabling the system to adjust the power output more timely to meet the vehicle's power demands.

[0046] 4. The bidirectional long short-term memory network with an attention mechanism can adaptively process data under different road conditions and driving styles, enhancing the system's adaptability and robustness. When facing sensor noise, data missing, or outliers, the system can maintain high prediction accuracy and stability to ensure the reliable operation of the whole vehicle.

[0047] 5. Through more accurate power prediction and more reasonable power distribution, the system can utilize the resources of fuel cells and power batteries more optimally, improving energy utilization efficiency. This helps reduce the energy consumption and emissions of the whole vehicle, cut operating costs, while extending the service life of the battery pack and reducing maintenance costs.

[0048] Furthermore, the objective function of the multi-objective optimization framework is formulated as:

[0049]

[0050] where η sys is the combined efficiency of the fuel cell and the power battery, dV decay / dt is the fuel cell voltage decay rate, SOC is the state of charge of the power battery, the weight coefficients α + β + γ = 1, and the dynamic adjustment strategy is that when SOC < 30%, α = 0.6, β = 0.2, γ = 0.2; when 30% ≤ SOC ≤ 70%, α = 0.4, β = 0.4, γ = 0.2; when SOC > 70%, α = 0.2, β = 0.6, γ = 0.2.

[0051] Furthermore, the efficiency mapping model of the fuel cell is formulated as:

[0052]

[0053] where P fc,net is the net output power of the fuel cell, P bat,out is the discharge power of the power battery, P aux is the auxiliary power consumption of non-driving devices, m H2 is the hydrogen mass flow rate, and LHV is the lower heating value of hydrogen.

[0054] Furthermore, the life mapping model of the fuel cell is formulated as:

[0055]

[0056] where k1, k2, k3 are the experimentally calibrated fuel cell decay coefficients, T is the fuel cell electrode temperature, λ air is the air stoichiometry ratio, and ΔRH is the humidity difference across the fuel cell membrane.

[0057] The beneficial effects of the basic solution are as follows: 1. Through the dynamic adjustment of the weight coefficients α, β, and γ in the multi-objective optimization framework, a balance is achieved among the comprehensive efficiency of the fuel cell and the power battery, the voltage decay rate of the fuel cell, and the state of charge of the power battery. When the SOC is low, the system focuses more on improving the comprehensive efficiency (α is larger) to ensure sufficient power output of the vehicle; when the SOC is moderate, the system pays attention to both efficiency and lifespan (α and β are comparable); when the SOC is high, the system focuses more on extending the lifespan of the fuel cell (β is larger). The dynamic adjustment strategy of the weight coefficients enables the system to automatically adjust the optimization objectives according to different driving conditions and battery states, improving the adaptability and flexibility of the system. This dynamic adjustment helps the system maintain optimal performance under different conditions and extend the service life of the fuel cell and the power battery.

[0058] 2. By considering the net output power of the fuel cell, the discharge power of the power battery, and the auxiliary power consumption of non-driving devices, the efficiency mapping model accurately evaluates the actual efficiency of the fuel cell. This helps the system better understand the working state of the fuel cell and provides a basis for subsequent power distribution and optimization. Based on the efficiency mapping model, the system can formulate a more reasonable power distribution strategy to ensure that the fuel cell operates in the high-efficiency working area and improve the energy utilization efficiency of the whole vehicle.

[0059] 3. By considering factors such as the fuel cell electrode temperature, air stoichiometry, and humidity difference across the membrane, the lifespan mapping model predicts the voltage decay rate of the fuel cell. This helps the system anticipate the lifespan decay of the fuel cell in advance and provides a basis for subsequent maintenance and replacement plans. Based on the lifespan mapping model, the system can adjust the working conditions of the fuel cell (such as temperature, air stoichiometry, etc.) to slow down the voltage decay rate and extend the service life of the fuel cell.

[0060] 4. Through the optimization of the multi-objective optimization framework and the guidance of the fuel cell efficiency and lifespan mapping models, the system can more accurately predict power demands, more reasonably distribute power, and more effectively utilize energy, thereby improving the performance and economy of the whole vehicle. The application of the fuel cell lifespan mapping model helps the system slow down the voltage decay rate of the fuel cell and extend its service life; at the same time, the SOC management of the power battery also helps extend its service life. Through more accurate power prediction and more reasonable power distribution, the system can reduce unnecessary power fluctuations and overload situations, enhancing the reliability and safety of the system. The improvement of the whole vehicle's performance, the extension of the service life of key components, and the enhancement of system reliability all contribute to reducing the operation and maintenance costs of the whole vehicle and improving the cost performance of the whole vehicle.

[0061] Furthermore, the power battery power distribution strategy is based on the future short-term power demand sequence predicted by a bidirectional long short-term memory network and adopts steady-state component extraction and transient component calculation, which is formulated as:

[0062]

[0063] Among them, α = 0.2, the time constant is 5 s, ensuring that the power change rate of the fuel cell ≤ 10 kW / s.

[0064]

[0065] Set the output power of the fuel cell to 40% - 90%, the power change rate to -15 kW / s - 10 kW / s, and the current density to 0.6 A / cm 2 -1.2 A / cm 2 When exceeding the interval, adjust the EMA filtering coefficient α to ensure the safety of the fuel cell.

[0066] The charging conditions of the power battery are as follows:

[0067]

[0068] Set the power battery to continuously charge and discharge ≤ 1C, the instantaneous peak value ≤ 3C, the duration of the instantaneous peak value ≤ 10 s, and the power battery temperature > 0 °C to be eligible for charging, and > 45 °C to prohibit discharging.

[0069] The beneficial effects of the basic solution are as follows: 1. The power distribution strategy of the power battery is based on the future short-term power demand sequence predicted by the bidirectional long short-term memory network, which can capture the dynamic changes of power demand more accurately, so as to formulate a more reasonable power distribution plan. Through the extraction of the steady-state component and the calculation of the transient component, the system can distinguish the long-term trend and short-term fluctuations in the power demand, providing a scientific basis for the power distribution of the fuel cell and the power battery.

[0070] 2. The steady-state component extraction formula ensures the smoothness of the fuel cell power change, avoids the sharp fluctuations of power, and is conducive to extending the service life of the fuel cell. At the same time, restricting the fuel cell power change rate ≤ 10 kW / s further protects the fuel cell from damage caused by excessive power impact.

[0071] 3. The setting of the charging conditions of the power battery ensures that the power battery is charged at the appropriate time, avoiding overcharging and unnecessary charging losses. This setting helps to improve the charging efficiency of the power battery and extend its service life.

[0072] 4. Through precise power distribution and reasonable charge and discharge condition settings, the system can utilize the energy of fuel cells and power batteries more efficiently, improving the performance and economy of the whole vehicle. Smooth fuel cell power output, reasonable charge and discharge current limits, and scientific temperature management all contribute to extending the service life of fuel cells and power batteries. Dynamically adjusting the EMA filtering coefficient, setting reasonable charge and discharge conditions, and temperature management all enhance the reliability and safety of the system, reducing potential safety risks. Precise power distribution and smooth power output help improve driving comfort and reduce driving discomfort caused by power fluctuations. At the same time, reasonable charge and discharge condition settings also avoid driving interruptions or failures caused by battery problems, enhancing the user experience.

[0073] Furthermore, the output module includes an output motor unit and a speed change unit.

[0074] The output motor unit is used to control the speed and torque of the motor based on the power battery power distribution strategy and distribute the power output of the battery.

[0075] The speed change unit is used to cooperate with the motor for speed coupling control to dynamically respond to load mutations caused by road surface gradient and load changes predicted by a bidirectional long short-term memory network.

[0076] The beneficial effects of the basic solution are as follows: 1. The output motor unit can precisely control the speed and torque of the motor and make dynamic adjustments according to the power output of the battery. This helps ensure that the motor can maintain optimal performance under different driving conditions, improving the power response and driving comfort of the whole vehicle. Through precise control, the motor can utilize the power provided by the battery more effectively, reducing energy loss and improving energy utilization efficiency.

[0077] 2. The speed change unit can cooperate with the motor for speed coupling control to dynamically respond to load mutations caused by road surface gradient and load changes predicted by a bidirectional long short-term memory network. This helps the whole vehicle maintain stable driving performance under different road conditions and load conditions. For example, when going uphill or the load increases, the speed change unit can automatically adjust the transmission ratio to increase the output torque of the motor, ensuring that the vehicle can climb the slope smoothly or carry heavy loads.

[0078] 3. The coordinated operation of the output motor unit and the transmission unit enables the vehicle to maintain optimal performance under different driving conditions, improving power response, driving comfort, and stability. By precisely controlling the motor speed and torque distribution, and optimizing the transmission ratio of the transmission unit, the system can more effectively utilize the power provided by the battery, reduce energy losses, and improve energy utilization efficiency. The output module can dynamically respond to sudden load changes caused by road surface gradient and load variations, enabling the vehicle to maintain stable driving performance under different road conditions and load conditions. This enhances the adaptability and robustness of the vehicle and improves its ability to cope with complex environments. By optimizing power distribution and reducing energy losses, the output module helps extend the service life of key components such as the motor and battery. At the same time, the reasonable control of the transmission unit also reduces the wear and failure rate of the transmission system. Brief Description of the Drawings

[0079] Figure 1 It is a schematic diagram of the power distribution and drive control system of the fuel cell heavy truck in the embodiment of the present invention.

[0080] Figure 2 It is a schematic diagram of the parameter acquisition module in the embodiment of the present invention.

[0081] Figure 3 It is a schematic diagram of the fuel cell optimization module in the embodiment of the present invention.

[0082] Figure 4 It is a schematic diagram of the output module in the embodiment of the present invention. Detailed Description of the Embodiments

[0083] The following is a more detailed description through specific embodiments:

[0084] Embodiment 1

[0085] Basically as shown in the attached Figure 1 、 Figure 2 、 Figure 3 and Figure 4 shown: The power distribution and drive control system of the fuel cell heavy truck includes a parameter acquisition module for collecting vehicle driving parameters, equipment operation parameters, and map environment parameters, and performing data enhancement processing. The parameter acquisition module includes a nine-axis gyroscope for detecting the vehicle driving attitude and direction; a wheel speed sensor for detecting the sequential vehicle speed during driving; a suspension pressure sensor for detecting the suspension pressure of the vehicle; a wind speed sensor for detecting the wind direction and wind speed of the vehicle at the current time; a fuel cell voltage inspection for detecting the change in the fuel cell output voltage; a battery pack temperature control sensor for detecting the change in the temperature inside the fuel cell pack. The vehicle driving parameters include the vehicle accelerator pedal travel, vehicle braking pressure, and suspension pressure; the equipment operation parameters include the sequential vehicle speed, fuel cell voltage, power battery voltage, and battery pack temperature; the map environment parameters include map navigation data, gradient, and wind speed.

[0086] A power prediction module, which is used to predict the future short-term power demand using a trained neural network according to the enhanced parameters. The data augmentation process is to add Gaussian noise to simulate sensor errors, with a standard deviation σ = 5%, and perform a time series sliding window process on the collected parameters, with a window length of 5 s and a step size of 0.1 s. The neural network is a bidirectional long short-term memory network with an attention mechanism, including a forward layer, a backward layer, and a hidden layer. The forward layer captures the acceleration trend features, the backward layer identifies the historical stable segments, and the attention mechanism weights the slope changes in the short-term future to predict the changes in short-term power demand.

[0087] A fuel cell optimization module, which is used to establish an efficiency mapping model and a life mapping model of the fuel cell based on a multi-objective optimization framework to find the optimal operating point of the fuel cell;

[0088] A power distribution module, which is used to establish a power distribution strategy for the power battery based on the predicted short-term power demand and the optimal operating point of the fuel cell to compensate for the power fluctuations of the fuel cell;

[0089] An output module, which is used to control the power output of the corresponding motor based on the optimal operating point of the fuel cell and the power distribution strategy of the power battery, and control the AMT transmission to perform automatic shifting. The output module includes an output motor unit, which is used to control the rotational speed and torque of the motor based on the power distribution strategy of the power battery to distribute the power output of the battery; a speed change unit, which is used to cooperate with the motor to perform rotational speed coupling control to dynamically respond to load mutations caused by changes in road surface slope and load predicted by the bidirectional long short-term memory network.

[0090] The specific implementation process is as follows: As Figure 2 shown, during the operation of the fuel cell heavy truck, the parameter acquisition module continuously obtains the vehicle roll angle, pitch angle, and yaw angle data through a nine-axis gyroscope. The wheel speed sensors collect the rotational speeds of each wheel at a frequency of 10 Hz and convert them into vehicle speed. The suspension pressure sensors continuously monitor the load status of each axle. The environmental perception unit synchronously integrates the elevation information and slope parameters in the high-precision map navigation data, and the wind speed sensor measures the real-time wind speed and wind direction angle. The fuel cell stack uses a distributed voltage inspection device to monitor the voltage fluctuations of single cells, and the battery pack temperature control sensors track the temperature field distribution of the fuel cell stack with an accuracy of 0.5 °C. These data provide a solid and complete data basis for subsequent power demand prediction. In particular, non-active change parameters such as wind speed and wind direction not only have a greater impact on vehicle load but also change frequently, which is likely to affect the stability of the output power.

[0091] After the original data is standardized and outliers are removed by the preprocessing module, Gaussian white noise with a standard deviation of σ = 5% is injected to enhance the robustness of the model. The time series data is processed by a sliding window with a window length of 5 seconds and a step size of 0.1 seconds to generate a continuous sample sequence, which is input into a bidirectional LSTM neural network with an attention mechanism. The forward layer of this network analyzes the acceleration trend features in the historical driving data, the backward layer extracts the steady-state operation rules, and the attention mechanism dynamically allocates weight coefficients according to the future short-term slope changes, and outputs the predicted value of the future short-term driving power demand.

[0092] In the bidirectional LSTM with an attention mechanism, the forward layer processes the historical time series data from t -5s to t0, and the backward layer processes the reverse time series data from t0 to t -5s , and the hidden layer outputs the hidden state and The formula for calculating the attention weight of the attention mechanism can be expressed as:

[0093]

[0094] Context vector generation:

[0095]

[0096] The loss function training of the neural network is optimized by the main loss function with a smoothing constraint, and the prediction result is corrected by the Kalman filter. The correction process is expressed formulaically as:

[0097]

[0098] Among them, x k is the power state, z k is the predicted value, and K k is the Kalman gain, which dynamically adjusts the prediction residual.

[0099] Through the prediction and optimization of the bidirectional LSTM, the accurate future short-term power demand change is obtained, which improves the accuracy of the subsequent model calculation results, thereby realizing the accurate and reasonable power distribution of the fuel cell and the power battery, helping to improve the overall efficiency of the system and extend the life of the fuel cell.

[0100] Based on the current stack health state, temperature distribution and efficiency mapping curve, the fuel cell optimization module uses the multi-objective particle swarm algorithm to solve the Pareto optimal solution set in real time, and determines the best operating point in the efficiency-life trade-off space. The power distribution module generates a dynamic charge and discharge strategy according to the difference between the predicted power demand and the fuel cell output reference power, combined with the power battery SOC state and temperature limit: when the demand power exceeds the fuel cell regulation ability, the power battery compensates for the power gap at a preset slope; during regenerative braking, the energy is preferentially stored in the battery.

[0101] As Figure 4 shown, the output module decomposes the target power into motor dq-axis current commands through a vector control algorithm and generates drive signals using space vector pulse width modulation. The AMT controller calculates the target gear position in advance based on the predicted load mutation and intervenes in the shifting operation during the motor speed synchronization stage: during the torque reduction stage, a feedforward compensation is used to maintain the vehicle acceleration smoothly, and during the shifting execution, a double closed-loop control of the clutch position and pressure is completed through a hydraulic actuator. The system updates the global optimization parameters every 100 ms to achieve the dynamic coupling of fuel cell operating point tracking, battery power distribution, and driveline response. Through precise control, the motor can utilize the power provided by the battery more effectively, reduce energy loss, improve energy utilization efficiency, and help extend the service life of key components such as the motor and battery.

[0102] Embodiment 2

[0103] The difference from the above embodiment is that, as shown in Figure 1 , Figure 3 and Figure 4 : The objective function of the multi-objective optimization framework is formulated as:

[0104]

[0105] where η sys is the combined efficiency of the fuel cell and the power battery, dV decay / dt is the fuel cell voltage decay rate, SOC is the state of charge of the power battery, the weight coefficients α + β + γ = 1, and the dynamic adjustment strategy is that when SOC < 30%, α = 0.6, β = 0.2, γ = 0.2; when 30% ≤ SOC ≤ 70%, α = 0.4, β = 0.4, γ = 0.2; when SOC > 70%, α = 0.2, β = 0.6, γ = 0.2.

[0106] The efficiency mapping model of the fuel cell is formulated as:

[0107]

[0108] where P fc.net is the net output power of the fuel cell, P bat.out is the discharge power of the power battery, P aux is the auxiliary power consumption of non-drive devices, m H2 is the hydrogen mass flow rate, and LHV is the lower heating value of hydrogen.

[0109] The life mapping model of the fuel cell is formulated as:

[0110]

[0111] Among them, k1, k2, and k3 are the fuel cell attenuation coefficients calibrated through experiments, T is the temperature of the fuel cell electrodes, λ air is the air stoichiometry ratio, and ΔRH is the humidity difference across the fuel cell membrane.

[0112] The power distribution strategy for the power battery is based on the future short-term power demand sequence predicted by a bidirectional long short-term memory network, and uses steady-state component extraction and transient component calculation, which is formulated as:

[0113]

[0114] Among them, α = 0.2, and the time constant is 5 s, ensuring that the fuel cell power change rate ≤ 10 kW / s.

[0115]

[0116] Set the output power of the fuel cell to 40% - 90%, the power change rate to -15 kW / s - 10 kW / s, and the current density to 0.6 A / cm 2 - 1.2 A / cm 2 , when exceeds the range, adjust the EMA filtering coefficient α to ensure the safety of the fuel cell.

[0117] The charging conditions for the power battery are:

[0118]

[0119] Set the power battery to continuously charge and discharge ≤ 1C, the instantaneous peak value ≤ 3C, the instantaneous peak duration ≤ 10 s, the power battery temperature > 0°C to be eligible for charging, and > 45°C to prohibit discharging.

[0120] The specific implementation process is as follows: As Figure 3 shown, after the fuel cell optimization module receives the power demand prediction result, based on the current stack temperature field distribution and health status, it calculates the efficiency - life model frontier in real time: the efficiency model substitutes parameters such as hydrogen flow rate and air compressor energy consumption into the formula, and the life model evaluates the performance degradation risk through the voltage decay equation. The multi-objective optimizer dynamically adjusts the target function weights according to the SOC state: when the SOC drops to 28%, the α weight is increased to 0.6 to prioritize energy supply; when the SOC is at 50%, the efficiency and life are balanced (α = β = 0.4), and the optimal operating point is solved through the particle swarm algorithm. The dynamic adjustment strategy of the weight coefficients enables the system to automatically adjust the optimization objectives according to different driving conditions and battery states, improving the adaptability and flexibility of the system. This dynamic adjustment helps the system maintain better performance under different conditions and extend the service life of the fuel cell and the power battery.

[0121] The power distribution module decomposes the power demand predicted by LSTM into steady-state and transient components, and uses EMA filtering (α = 0.2) to extract the base load power of the fuel cell. Ensure that its output is limited within the rated power range of 40%-90% and the rate of change ≤ 10kW / s. When exceeding the threshold, the filtering coefficient is adaptively adjusted. The power battery charges and discharges according to the difference P. bat When it detects regenerative braking energy recovery and SOC < 65%, it charges; in case of an emergency acceleration demand (P. bat > 250kW), it triggers the 3C instantaneous discharge mode and starts the liquid cooling system for forced heat dissipation. The fuel cell safety module continuously monitors the current density and humidity gradient, and switches to the derating mode when abnormal. This power distribution algorithm can reduce unnecessary power fluctuations and overload situations, enhancing the reliability and safety of the system. The performance improvement of the whole vehicle, the extension of the service life of key components, and the enhancement of system reliability all contribute to reducing the operation and maintenance costs of the whole vehicle and improving the cost performance of the whole vehicle.

[0122] Vehicle verification experiment: The improvement of hydrogen consumption and fuel cell life loss by this system under different working conditions is shown in Table 1 below.

[0123] Table 1. Comprehensive effects of this system on energy consumption and battery life

[0124]

[0125] As can be seen from the table, this system has a certain effect on reducing energy consumption and extending battery life under different working conditions.

[0126] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0127] The above are only embodiments of the present invention. Specific structures and common knowledge such as characteristics that are well-known in the art are not described in detail herein. Those of ordinary skill in the art know all the common general technical knowledge in the technical field to which the invention pertains before the filing date or the priority date, are able to know all the prior art in this field, and have the ability to apply the conventional experimental means before this date. Those of ordinary skill in the art can, under the inspiration given in this application, perfect and implement this solution in combination with their own abilities. Some typical well-known structures or well-known methods should not become obstacles for those of ordinary skill in the art to implement this application. It should be noted that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can still be made, and these should also be regarded as the protection scope of the present invention, and these will not affect the implementation effect of the present invention and the practicality of the patent. The protection scope claimed in this application should be based on the content of its claims, and the specific implementation manners and the like recorded in the specification can be used to interpret the content of the claims.

Claims

1. The power distribution and drive control system of a fuel cell heavy truck is characterized by: It includes parameter acquisition module, power prediction module, fuel cell optimization module, power distribution module and output module. Parameter collection module, used to collect vehicle driving parameters, equipment operation parameters and map environment parameters, and perform data enhancement processing; A power prediction module, which is used to predict the future short-term power demand using the trained neural network based on the enhanced parameters; The fuel cell optimization module is used to establish the efficiency mapping model and life mapping model of the fuel cell based on the multi-objective optimization framework to find the optimal operating point of the fuel cell; A power allocation module is used to establish a power allocation strategy for the power battery based on the predicted short-term power demand and the optimal operating point of the fuel cell to compensate for the power fluctuation of the fuel cell; The output module is used to control the power output of the corresponding motor based on the optimal operating point of the fuel cell and the power distribution strategy of the power battery, and control the AMT transmission to perform automatic speed change.

2. The power distribution and drive control system of a fuel cell heavy truck according to claim 1 is characterized in that: The parameter acquisition module includes a nine-axis gyroscope, a wheel speed sensor, a suspension pressure sensor, a wind speed sensor, a fuel cell voltage inspection sensor, and a battery pack temperature control sensor. Nine-axis gyroscope, used to detect vehicle driving posture and direction; Wheel speed sensor, used to detect the sequential speed of the vehicle; Suspension pressure sensor, used to detect the suspension pressure of the vehicle; Wind speed sensor, used to detect the wind direction and wind speed that the vehicle is currently exposed to; Fuel cell voltage inspection, used to detect changes in fuel cell output voltage; The battery pack temperature control sensor is used to detect temperature changes within the fuel cell pack.

3. The power distribution and drive control system of a fuel cell heavy truck according to claim 1 is characterized in that: Vehicle driving parameters include vehicle accelerator pedal travel, vehicle brake pressure and suspension pressure; equipment operation parameters include sequential vehicle speed, fuel cell voltage, power battery voltage and battery pack temperature; map environment parameters include map navigation data, slope and wind speed.

4. The power distribution and drive control system of a fuel cell heavy truck according to claim 1, characterized in that: The data enhancement processing is to add Gaussian noise to simulate sensor error, with a standard deviation of σ = 5%, and perform time series sliding window processing on the acquisition parameters with a window length of 5s and a step length of 0.1s.

5. The power distribution and drive control system of a fuel cell heavy truck according to claim 1, characterized in that: The neural network is a bidirectional long short-term memory network with an attention mechanism, including a forward layer, a backward layer and a hidden layer. The forward layer captures the acceleration trend characteristics, the backward layer identifies the historical stable segment, and the attention mechanism weights the future short-term slope changes to predict short-term power demand changes.

6. The power distribution and drive control system of a fuel cell heavy truck according to claim 1, characterized in that: The objective function of the multi-objective optimization framework is formulated as: Among them, η sys is the combined efficiency of the fuel cell and power battery, dV decay / dt is the fuel cell voltage decay rate, SOC is the state of charge of the power battery, the weight coefficient α+β+γ=1, and the dynamic adjustment strategy is when SOC<30%, α=0.6, β=0.2, γ=0.2; when 30%≤SOC≤70%, α=0.4, β=0.4, γ=0.2; when SOC>70%, α=0.2, β=0.6, γ=0.

2.

7. The power distribution and drive control system of a fuel cell heavy truck according to claim 6, characterized in that: The efficiency mapping model of the fuel cell is formulated as: Among them, P fc，net is the net output power of the fuel cell, P bat，out is the power battery discharge power, P aux is the auxiliary power consumption of non-driving devices, m H2 is the mass flow rate of hydrogen, and LHV is the lower heating value of hydrogen.

8. The power distribution and drive control system of a fuel cell heavy truck according to claim 6, characterized in that: The fuel cell life mapping model is formulated as: Among them, k1, k2, k3 are the experimentally calibrated fuel cell attenuation coefficients, T is the fuel cell electrode temperature, λ air is the air stoichiometric ratio, and ΔRH is the humidity difference on both sides of the fuel cell membrane.

9. The power distribution and drive control system of a fuel cell heavy truck according to claim 1, characterized in that: The power allocation strategy of the power battery is based on the future short-term power demand sequence predicted by the bidirectional long short-term memory network, using steady-state component extraction and transient component calculation, and is formulated as follows: Wherein, α=0.2, the time constant is 5s, and the fuel cell power change rate is ensured to be ≤10kW / s. The output power of the fuel cell is set to 40%-90%, the power change rate is -15kW / s-10kW / s, and the current density is 0.6A / cm 2 -1.2A / cm 2 ,when When the range is exceeded, the EMA filter coefficient α is adjusted to ensure the safety of the fuel cell. The charging conditions of the power battery are: Set the power battery continuous charge and discharge to ≤1C, the instantaneous peak value to ≤3C, the instantaneous peak duration to ≤10s, the power battery temperature is greater than 0℃ for charging, and greater than 45℃ for discharging.

10. The power distribution and drive control system of a fuel cell heavy truck according to claim 1, characterized in that: The output module includes an output motor unit and a speed change unit. An output motor unit, used to control the speed and torque of the motor to distribute the power output of the battery based on the power distribution strategy of the power battery; The speed change unit is used to cooperate with the motor to perform speed coupling control and dynamically respond to sudden load changes caused by road slope and load changes predicted by the bidirectional long and short-term memory network.

Images