Lithium battery hybrid power motorcycle performance improvement control strategy generation method

Abstract

The invention provides a lithium battery hybrid power motorcycle performance improvement control strategy generation method, which comprises the steps that vehicle driving states comprise a sliding state and a braking state, if the current vehicle is in the sliding state, a preset sliding energy recovery mode is activated, and if the current vehicle is in the braking state, a preset braking energy recovery mode is activated; different energy recovery strategies are combined and optimized according to the subdivision result of the energy recovery strategies in the sliding state and the braking state, and the vehicle speed, the battery charging state and the energy recovery potential value parameters are input to obtain a target energy recovery strategy combination scheme; the intention of acceleration, deceleration or constant speed of a driver is judged according to the operation characteristics of an accelerator and a brake, the operation characteristics refer to the force of stepping on an accelerator pedal and a brake pedal, and the driver intention obtained through judgment serves as the basis for dynamic adjustment of the energy recovery strategy.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery. Background Art

[0002] In the energy recovery control of hybrid motorcycles, a key technical issue is how to deal with the energy recovery needs in two different states: coasting and braking. Specifically, when the vehicle is in the coasting state, since the vehicle has kinetic energy but no active braking, it is necessary to accurately calculate the size of the motor's reverse torque to ensure the energy recovery efficiency and not affect the driver's coasting experience. In the braking state, because the driver applies active braking force, it is necessary to reasonably allocate the proportion of mechanical braking force and electric motor force under the premise of ensuring braking safety. There is a conflict between the energy recovery strategies in the two states, that is, the contradiction between the pursuit of smoothness in the coasting state and the pursuit of recovery efficiency in the braking state. At the same time, due to the randomness and diversity of the driver's operation, there is often a frequent switch between coasting and braking, which requires the control system to quickly identify the current state and adjust the recovery strategy in real time. In addition, the charging characteristics of lithium batteries also impose constraints on the energy recovery strategy. How to dynamically adjust the recovery power according to the battery's state of charge, temperature and other parameters further increases the complexity of strategy switching. Summary of the invention

[0003] The present invention provides a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery, which mainly includes:

[0004] Vehicle dynamics analysis is performed based on the real-time acquired lithium battery capacity status, motor energy recovery efficiency, vehicle driving status and driver operation data to obtain instantaneous energy consumption, and combined with the actual speed of the vehicle, the energy recovery potential value of the current vehicle is calculated;

[0005] The vehicle driving state includes a coasting state and a braking state. If the vehicle is currently in a coasting state, a preset coasting energy recovery mode is activated; if the vehicle is currently in a braking state, a preset braking energy recovery mode is activated;

[0006] The vehicle speed and braking force are collected through sensors, and the energy recovery strategies in the gliding and braking states are subdivided according to the speed and braking force. In the gliding state, the recovery strategy is divided into light, medium or heavy recovery strategies according to the length of the gliding distance, and in the braking state, the recovery strategy is divided into gradual, stable or sharp recovery strategies according to the braking force;

[0007] According to the subdivision results of energy recovery strategies under the coasting and braking states, different energy recovery strategies are combined and optimized, and the vehicle speed, battery charging state, and energy recovery potential value parameters are input to obtain the target energy recovery strategy combination scheme;

[0008] Determine the driver's intention to accelerate, decelerate or maintain a constant speed based on the operating characteristics of the accelerator and brake pedals, where the operating characteristics refer to the strength of the accelerator and brake pedals, and use the determined driver's intention as the basis for dynamic adjustment of the energy recovery strategy;

[0009] Dynamically adjust the intensity of the energy recovery strategy based on the real-time battery capacity status and driver intention. If the battery capacity status is higher than the preset threshold, the recovery intensity is reduced. At the same time, the driver's intention is determined by the operating characteristics of the accelerator and brake. If the driver intends to accelerate, the recovery intensity is increased.

[0010] The dynamic adjustment method of the energy recovery strategy intensity is applied to the target energy recovery strategy combination scheme, and the battery's state of charge and temperature parameters are monitored at the same time to calculate the recovery power of the energy recovery strategy before and after the dynamic adjustment of the intensity, obtain the change in the recovery power, and switch the strategy in the target energy recovery strategy combination scheme in real time.

[0011] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0012] The present invention discloses a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery. The method calculates instantaneous energy consumption and energy recovery potential values ​​by collecting lithium battery capacity, motor efficiency, vehicle driving status, and driver operation data in real time. According to the vehicle's coasting or braking status, the corresponding energy recovery mode is activated, and the recovery strategy is segmented and optimized based on the vehicle speed and braking force. At the same time, the energy recovery intensity is dynamically adjusted by analyzing the driver's operating intention and combining the battery capacity status. In addition, the battery parameters are monitored, and the recovery power changes are calculated in real time to achieve intelligent switching of strategies. The present invention ensures the smoothness and comfort of the driving experience through multi-dimensional data analysis and strategy optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flow chart of a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery.

[0014] Figure 2 It is a schematic diagram of a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery according to the present invention.

[0015] Figure 3 This is another schematic diagram of a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery according to the present invention. DETAILED DESCRIPTION

[0016] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with the accompanying drawings and embodiments. The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant inventions, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0017] like Figure 1-3 In this embodiment, a method for generating a control strategy for improving the performance of a hybrid motorcycle using a lithium battery may specifically include:

[0018] S101. Perform vehicle dynamics analysis based on the real-time acquired lithium battery capacity status, motor energy recovery efficiency, vehicle driving status and driver operation data to obtain instantaneous energy consumption, and calculate the energy recovery potential value of the current vehicle in combination with the acquired actual speed of the vehicle.

[0019] A first power value is obtained according to the product of the battery terminal voltage and the charging current, and the first power value is fitted by the least square method to obtain the battery capacity status value; a road slope value is calculated by using global positioning satellite data, and a linear Kalman filter operation is performed on the road slope value and the load mass data collected by the vehicle-mounted gravity sensor to obtain a second power value; if the motor recovery efficiency is higher than a preset threshold, a power recovery ratio value is obtained by a discrete integral operation according to the battery capacity status value and the vehicle speed data in the vehicle-mounted controller bus; the power recovery ratio value is subjected to deviation normalization processing to obtain standardized data, a long short-term memory neural network is used to construct a time series prediction model for the standardized data to obtain an energy consumption prediction value, and the energy recovery potential value is calculated by combining the instantaneous vehicle speed and the acceleration value.

[0020] Exemplarily, multiple sensors are used to continuously collect the battery terminal voltage and charging current, the body moment of inertia is calculated according to the vehicle mass parameter and the tire radius value, the power value is obtained by multiplying the battery terminal voltage and the charging current, the power value is fitted by the least square method and stored in the battery capacity comparison database, the battery capacity status record value is read from the database, and the vehicle speed sensor data and the engine power output data are collected to obtain the real-time power consumption value. The vehicle attitude quaternion is calculated according to the three-axis acceleration and angular velocity data collected by the inertial measurement unit, the vehicle center of mass height and the driving road slope value are calculated in combination with the global positioning satellite data, the road slope value and the vehicle load mass collected by the vehicle gravity sensor are linearly Kalman filtered, and the instantaneous power value is calculated using the drag coefficient and the real-time vehicle speed data. If the motor recovery efficiency is higher than 0.85, the power recovery controller is activated, the battery capacity status value and the vehicle speed data in the vehicle controller communication bus are used, the real-time power recovery ratio is calculated by discrete integration, and the voltage control command is executed on the vehicle braking device and the battery pack charging circuit to obtain the recovered electric energy value. The deviation between the power consumption per mileage value stored in the vehicle database and the currently collected data is normalized, and a time series prediction model is constructed using a long short-term memory neural network for the normalized data. The energy consumption value of the future road section is obtained from the prediction result, and the energy recovery potential value is calculated by comparing the instantaneous vehicle speed and acceleration value. The energy recovery potential can be evaluated by calculating the kinetic energy difference between the initial and final instantaneous vehicle speeds.

[0021]

[0022] , E rec represents the energy recovery potential value, m represents the vehicle mass, v i represents the initial instantaneous vehicle speed, v fIndicates the final instantaneous vehicle speed. The battery terminal voltage and charging current obtained by the on-board multi-sensor combination reflect the direct battery working state. When the battery is at an ambient temperature of 25 degrees Celsius, the terminal voltage of 3.7 volts corresponds to an 80% capacity state. As the temperature drops to minus 10 degrees Celsius, the capacity state drops to 65% at the same terminal voltage. The capacity state comparison database is constructed through the three-dimensional data of battery voltage, current and temperature to achieve accurate estimation of battery capacity. When the vehicle mass is 1500 kg and the tire radius is 0.35 meters, the vehicle's moment of inertia is about 156 kg / m2, the engine output power is 85 kWh, and the real-time power consumption is about 65 kW. When the vehicle is driving on an uphill section with a slope of 5 degrees, the longitudinal acceleration of 0.5 meters per square second, the lateral acceleration of 0.2 meters per square second, and the vertical acceleration of 9.8 meters per square second are obtained through the inertial measurement unit. Combined with the vehicle elevation data obtained when the global positioning satellite positioning accuracy is 0.5 meters, the vehicle's center of mass height of 1.2 meters is calculated. When the vehicle is loaded with 600 kg of cargo, the total mass of the vehicle is 2100 kg, the air resistance coefficient is 0.31, and the instantaneous power consumption reaches 72 kilowatts under the condition of 80 kilometers per hour measured by the gravity sensor. When the motor recovery efficiency reaches 0.85 or above, such as the vehicle slowing down on a downhill section or braking on a flat road, the power recovery controller calculates the braking recovery power of 45 kilowatts based on the battery 70% capacity state and the 80 kilometers per hour speed data in the vehicle bus. The controller sends a 400 volt voltage control command to the vehicle-mounted braking device to execute the battery charging operation to obtain 40 kilowatts of recovered electrical energy. The on-board database records the historical data of 0.9 degrees of power consumption per kilometer, and after standardization with the currently collected 1.2 degrees per kilometer data, it is input into the time series model trained by the long short-term memory neural network, predicting that the average power consumption of the next 5 kilometers of road section will be 1.1 degrees per kilometer. When the current speed drops from 80 kilometers per hour to 60 kilometers per hour and the acceleration is negative 2 meters per square second, the energy recovery potential value of 50 kilowatts is calculated. In the stop-and-go scenario during peak hours on urban roads, frequent changes in vehicle speed increase the opportunities for braking energy recovery, but the heat loss of the motor increases, causing the recovery efficiency to drop from 0.85 to 0.75. At this time, the recovery power should be reduced to 35 kilowatts. On unobstructed sections of suburban roads, when the vehicle speed is stable at 90 kilometers per hour, the number of braking times decreases, but the air resistance loss increases, and the power consumption per kilometer rises to 1.4 degrees.

[0023] S102: The vehicle driving state includes a coasting state and a braking state. If the vehicle is currently in a coasting state, a preset coasting energy recovery mode is activated; if the vehicle is currently in a braking state, a preset braking energy recovery mode is activated.

[0024] The three-axis angular velocity and acceleration values ​​collected by the vehicle-mounted inertial sensor are obtained, and the vehicle motion state is determined according to the pedal opening value fed back by the brake pedal displacement sensor and the wheel Hall speed sensor signal; the speed and torque signals fed back by the motor controller are obtained from the vehicle-mounted communication bus according to the vehicle motion state, and the driving resistance is calculated by the values ​​collected by the body posture sensor and the atmospheric pressure sensor; if the accelerator pedal opening is zero and the brake pedal opening is less than the brake pedal opening threshold, the coasting recovery power limit is determined by the voltage and current limit values ​​fed back by the battery management unit; if the brake pedal opening is greater than the brake pedal opening threshold, the proportional-integral controller is used to adjust the distribution ratio of the hydraulic braking and electric motor force according to the master cylinder pressure value collected by the brake pressure sensor and the remaining capacity state fed back by the battery management unit.

[0025] Exemplarily, the vehicle-mounted inertial sensor is used to collect the vehicle's three-axis angular velocity and acceleration values ​​in real time. By comparing the longitudinal acceleration threshold with the current acceleration value, combined with the brake pedal stroke collected by the brake pedal displacement sensor and the preset opening threshold comparison result, the wheel speed periodic signal fed back by the wheel Hall speed sensor is collected, and the speed frequency operation is performed to obtain the real-time vehicle speed value, and the current vehicle motion state is determined based on the above three sets of data. The motor speed signal and torque value fed back by the drive motor controller are obtained from the vehicle-mounted communication bus, and the instantaneous driving power is obtained by multiplying the speed and torque. Combined with the pitch angle data collected by the body posture sensor, the air density value measured by the atmospheric pressure sensor and the square of the vehicle speed are used to obtain the air resistance value, and the rolling resistance coefficient is obtained from the tire pressure sensor, and the driving resistance is calculated based on the above values. If the accelerator pedal opening is zero and the brake pedal opening is less than 5%, the motor controller outputs a braking torque of 0.5 times the real-time vehicle speed, and obtains the battery pack terminal voltage and the maximum allowable charging current limit from the battery management unit. The voltage and current threshold product is calculated to obtain the coasting recovery power limit, and the actual braking torque output is adjusted according to the temperature value collected by the battery pack temperature sensor. If the brake pedal opening is greater than 5%, the brake pressure sensor is used to collect the master cylinder pressure value, and the maximum allowable braking pressure is calculated by the linear regression predictor. Combined with the remaining capacity state fed back by the battery management unit, the proportional integral controller is used to adjust the hydraulic proportional valve opening to achieve dynamic distribution of hydraulic braking and electric machine power, and perform constant pressure charging control on the brake recovery electric energy. The vehicle driving state judgment involves multi-dimensional parameter fusion. The longitudinal acceleration collected by the inertial sensor in the range of -0.3 to 0.3 meters per square second indicates that the vehicle is in a uniform speed or slight acceleration and deceleration state. At the same time, when the brake pedal opening is less than 5%, it is determined to be a coasting state. The wheel speed sensor outputs 96 pulse signals per rotation. Under the condition of 2.1 meters of tire rolling circumference, the vehicle speed value is obtained by counting the pulse frequency. The pulse frequency of 80 kilometers per hour corresponds to 210 Hz. The drive motor outputs a peak torque of 280 Nm at a rated speed of 3000 rpm. The actual speed is detected by the speed sensor at 2800 rpm. When the torque sensor measures the output torque of 200 Nm, the drive power is calculated to be 58.6 kW. When the vehicle is driving on a 5-degree uphill road, the body pitch angle sensor measures an angle of 4.8 degrees with the horizontal plane. When the atmospheric pressure sensor measures an air pressure of 101.3 kPa, the air density is 1.29 kg per cubic meter, and the air resistance reaches 1200 Nm under the condition of 100 kilometers per hour. When the tire pressure sensor measures a tire pressure of 220 kPa, the rolling resistance coefficient is 0.012. Under highway coasting conditions, when the vehicle speed drops from 120 kilometers per hour to 100 kilometers per hour, the drive motor controller outputs 60 Newton-meters of braking torque. At this time, the battery management unit measures the battery pack terminal voltage of 360 volts, the maximum allowable charging current limit of 80 amperes, and the calculated coasting recovery power limit is 28.8 kilowatts.When the battery temperature sensor detects that the battery pack temperature rises to 45 degrees Celsius, the controller reduces the braking torque to 40 Nm to prevent the battery from overheating. During braking deceleration in urban conditions, the brake master cylinder pressure sensor measures a pressure value of 4 MPa, and the linear regression predictor calculates the maximum allowable braking pressure of 5 MPa based on historical data. When the battery management unit feedbacks that the remaining battery capacity is 80%, the proportional integral controller outputs a hydraulic braking ratio of 0.6, and the remaining 0.4 is supplemented by motor braking. Under the condition of 366 volts at the battery terminal, constant voltage charging control is performed, and the charging current is gradually reduced from 80 amps to 20 amps to achieve smooth recovery of braking energy. In mountain downhill conditions, frequent braking causes the hydraulic brake temperature to rise. At this time, the motor braking ratio is increased to 0.6, the hydraulic braking intensity is reduced, and the stable braking recovery power is maintained. In winter low temperature conditions, when the battery temperature is below 5 degrees Celsius, the charging current limit is reduced to 40 amps, and the controller correspondingly increases the hydraulic braking ratio to 0.8 to ensure braking safety.

[0026] S103. The vehicle speed and braking force are collected through sensors, and the energy recovery strategies in the gliding and braking states are subdivided according to the vehicle speed and braking force. In the gliding state, the recovery strategy is divided into light, medium or heavy recovery strategies according to the length of the gliding distance. In the braking state, the recovery strategy is divided into progressive, stable or sudden recovery strategies according to the braking force.

[0027] The pedal stroke value is obtained from the brake pedal displacement sensor, and the real-time vehicle speed value is obtained through the vehicle speed sensor. If the vehicle speed value is lower than the preset speed threshold and the braking torque is less than the preset torque threshold, it is determined to be in a gliding state; the gliding distance value is calculated according to the vehicle position coordinate sequence, and the gliding distance value is classified by a neural network in combination with the air resistance coefficient and the vehicle body mass, and the light, medium and heavy recovery torques are determined based on the remaining battery capacity and the gliding distance classification results; the pedal stroke value is subjected to wavelet decomposition to obtain the stroke change rate, and the progressive braking, stable braking or sharp braking type is determined according to the stroke change rate in combination with the wheel deceleration value and the braking torque value; a double-layer neural network is used to calculate the battery charging power according to the gear of the recovery torque and the braking type, and based on the motor speed and power conversion ratio, the motor braking torque of the corresponding gear is output through the regulator.

[0028] For example, a vehicle speed sensor is used to collect the real-time vehicle speed value, a brake pedal displacement sensor is used to collect the pedal stroke value, a global positioning satellite receiver is used to record the vehicle position coordinates every 0.1 second, a braking torque signal is obtained from the motor torque sensor, and the sliding state is determined when the vehicle speed is less than 5 meters per second and the braking torque is less than 10 Newton meters, and the braking state is determined when the braking torque is greater than 10 Newton meters. The sliding distance is calculated according to the vehicle position coordinate sequence, and the sliding distance length is subdivided by a long short-term memory neural network combined with the air resistance coefficient and the vehicle body mass. The light, medium and heavy gear recovery torques are determined based on the remaining capacity fed back by the battery management unit and the conditions that the sliding distance is less than 100 meters, 100 to 500 meters, and greater than 500 meters. The pedal stroke change rate is obtained from the brake pedal displacement sensor, and the stroke change rate curve is decomposed into 4 layers using multi-layer wavelet decomposition. Combined with the deceleration value collected by the wheel speed sensor and the braking torque value of the motor torque sensor, if the pedal stroke change rate is less than 5 mm per second, it is determined to be gradual braking, if the change rate is 5 to 20 mm per second, it is determined to be stable braking, and if it is greater than 20 mm per second, it is determined to be abrupt braking. According to the coasting gear and braking type, a double-layer neural network is used to calculate the battery's allowable charging power. Based on the speed value of each gear fed back by the motor speed sensor and the input-output power ratio measured by the power conversion device, the motor braking torque is output through the proportional-integral regulator, corresponding to 0.2 times the maximum torque for light recovery, 0.5 times the maximum torque for medium recovery, and 0.8 times the maximum torque for heavy recovery, and the hydraulic braking ratio is dynamically adjusted according to the three braking state judgment results. The speed threshold setting in the vehicle state judgment is closely related to the actual driving characteristics. Under urban road conditions, the brake pedal is in a released state when the vehicle speed is 5 meters per second, and the driving motor outputs a braking torque of 8 Nm, which belongs to a typical coasting condition. When the braking torque reaches 15 Nm, the vehicle is in an active braking state. At this time, the on-board sensor records the position coordinates every 0.1 seconds, and the deviation between the straight-line distance between the two coordinate points and the actual driving trajectory is less than 0.5 meters. In highway conditions, when the vehicle speed is 120 kilometers per hour, the accelerator pedal is released to enter the gliding state. The neural network is used to extract the characteristics of parameters such as the air resistance coefficient of 0.31 and the body mass of 1,800 kilograms, and it is calculated that the gliding distance of 800 meters is suitable for the heavy recovery strategy. When the remaining battery capacity is 85%, the maximum allowable charging power is 40 kilowatts, corresponding to 0.8 times the maximum braking torque. When gliding at low speed in urban conditions, a moderate recovery strategy is adopted for the gliding distance of 150 meters, and 0.5 times the maximum torque is output. The determination of the braking stroke change rate is achieved through wavelet decomposition to identify different working conditions. During progressive braking, the pedal stroke increases slowly at a rate of 3 mm per second. At this time, the wheel speed sensor measures a deceleration of 0.8 meters per square second, and the motor outputs a braking torque of 50 Nm. Under stable braking conditions, the pedal stroke increase rate is maintained at 12 mm per second, and the deceleration reaches 2.5 meters per square second.In emergency braking, the pedal stroke increases rapidly to 25 mm per second, triggering a sharp braking judgment. The energy recovery control strategy under different working conditions is dynamically adjusted through a two-layer neural network. Under the condition of a vehicle speed of 40 kilometers per hour, the power conversion efficiency at a motor speed of 2400 rpm is 0.85, and the light recovery mode outputs a braking torque of 60 Nm. The proportion of hydraulic braking is adjusted to 0.7 through a proportional-integral controller. When entering a sharp braking condition, the motor braking torque increases rapidly to 240 Nm, and the proportion of hydraulic braking is increased to 0.9 to ensure the timely braking response. In long downhill conditions, continuous heavy recovery causes the motor temperature to rise. At this time, the motor braking torque is reduced to 180 Nm and the hydraulic braking ratio is increased to 0.8. In cold weather conditions, when the battery temperature is below 5 degrees Celsius, the maximum allowable charging power is reduced to 25 kilowatts. The controller accordingly reduces the motor braking torque output and increases the hydraulic braking force compensation to achieve a smooth transition.

[0029] Obtain the preset braking force threshold range and the real-time braking force data of the vehicle to determine the current braking force type, which includes progressive, stable and sharp types. For different braking force types, obtain the corresponding recovery strategy parameters from the pre-established braking force-recovery strategy mapping table. According to the obtained recovery strategy parameters, determine the generator torque output and battery charging current limit under the current braking state.

[0030] A brake pedal displacement sensor is used to collect brake pedal displacement signals, and the braking force type is obtained based on the displacement signal and the brake master cylinder pressure signal through deep neural network processing; a pressure curve slope value is obtained based on the brake master cylinder pressure signal, and the braking condition type is determined by comparing the pressure curve slope value with a preset slope threshold value; corresponding recovery parameters are read from a preset force recovery mapping table for the braking condition type, and the motor output torque value and the charging current limit are determined based on the recovery parameters; the motor output torque is collected through a torque sensor, and the battery pack parameters are obtained from a battery manager, and a proportional-integral controller is used to perform closed-loop regulation of the motor torque and the charging current based on the motor output torque and the battery pack parameters.

[0031] Exemplarily, a brake pedal displacement sensor is used to collect the real-time travel value of the brake pedal at a frequency of 100 Hz, combined with the hydraulic brake pressure signal collected by the brake master cylinder pressure sensor, and the displacement and pressure data are processed by a three-layer deep neural network. The preset force threshold range is compared. When the pressure is less than 2 MPa, it is judged as light braking, the pressure is between 2 and 5 MPa, it is judged as moderate braking, and the pressure is greater than 5 MPa, it is judged as heavy braking, and the current braking force type is obtained. The slope value of the pressure curve is obtained from the brake master cylinder pressure sensor, and the pressure change rate is calculated in real time. The progressive braking condition is judged according to the slope of the curve being less than 0.5 MPa per second, the stable braking condition is judged according to the slope between 0.5 and 2 MPa per second, and the sharp braking condition is judged according to the slope greater than 2 MPa per second. The braking type determination result is used to index the preset force recovery mapping table. The corresponding recovery parameters are read from the preset force recovery mapping table. For progressive braking, the output is 0.3 times the rated torque and 0.4 times the rated charging current. For stable braking, the output is 0.6 times the rated torque and 0.7 times the rated charging current. For sharp braking, the output is 0.9 times the rated torque and 0.9 times the rated charging current. The motor output torque value and charging current limit are calculated based on the recovery parameters. The motor output torque is collected every 10 milliseconds through the torque sensor, and the battery pack shell temperature and remaining capacity values ​​are obtained from the battery manager. The proportional integral controller is used to perform closed-loop regulation of the motor torque and charging current. When the battery temperature is higher than 40 degrees Celsius, the charging rate is reduced to 0.5, and when the remaining capacity is higher than 90%, the torque output is reduced to 0.4 times, realizing coordinated control of torque output and charging current. The coordinated perception of brake pedal stroke and master cylinder pressure reflects the characteristics of the braking operation. When the driver lightly steps on the brake pedal, the displacement sensor collects pedal stroke data every 10 milliseconds. When the pedal moves from a static position to 20 mm, the master cylinder pressure gradually rises from 0 to 1.5 MPa. The input layer of the three-layer neural network receives two sets of displacement and pressure data, the hidden layer has 64 neurons, and the output layer determines it as a light braking type. The brake pressure change rate curve contains rich driving intention information. In urban conditions, the driver slowly steps on the brake pedal, and the slope of the pressure curve remains at 0.3 MPa per second, which is a typical progressive braking. Under the highway following condition, the slope of the pressure curve is stable at 1.2 MPa per second when maintaining a fixed distance. In the emergency avoidance condition, the slope of the pressure curve rises rapidly to 2.5 MPa per second, triggering a sharp braking judgment. The force recovery mapping table sets the recovery parameters according to different braking types. Under the conditions of 300 Nm of motor rated torque and 120 amperes of rated charging current, the progressive braking outputs 90 Nm of torque and 48 amperes of charging current, which is suitable for long-distance light braking conditions. In stable braking conditions, the output torque is 180 Nm and the charging current is 84 amps, which meets the general road braking requirements. In sudden braking, the output torque is 270 Nm and the charging current is 108 amps, which achieves a quick response.The coordinated control of torque and charging current takes into account the working state of the battery. The torque sensor collects torque data every 10 milliseconds and implements the standard charging strategy when the battery temperature is 35 degrees Celsius. When the battery temperature rises to 45 degrees Celsius, the charging rate automatically drops to 0.4, and the torque output is reduced to 0.5 times accordingly. In cold weather conditions, when the battery temperature is below 5 degrees Celsius, the charging rate is limited to less than 0.3, and the hydraulic braking force is increased to compensate for the insufficient braking force. When the motor temperature rises due to continuous braking on a downhill section, the controller automatically reduces the motor torque output to 0.3 times and increases the proportion of hydraulic braking. When the remaining battery capacity is 95%, the charging current limit is reduced to 24 amps and the torque output is reduced to 60 Nm to avoid overcharging the battery. When the vehicle stops and starts in urban conditions, frequent braking causes the motor temperature to rise. At this time, the controller dynamically adjusts the maximum torque output according to the temperature value fed back by the thermistor to prevent the motor from overheating while ensuring braking performance. During long downhill driving on mountainous roads, continuous high-intensity braking causes the battery temperature to gradually rise. When the temperature sensor detects that the battery pack temperature exceeds 50 degrees Celsius, the charging current automatically drops to 36 amperes and the hydraulic braking force is increased to compensate.

[0032] S104. According to the subdivision results of the energy recovery strategies under the coasting and braking states, different energy recovery strategies are combined and optimized, and vehicle speed, battery charging state, and energy recovery potential value parameters are input to obtain a target energy recovery strategy combination scheme.

[0033] Collect real-time data output by the vehicle speed sensor, acceleration sensor and battery manager, and obtain the recovery power limit and recovery torque limit through a support vector machine regressor based on the real-time data output; calculate the recovery potential value based on the recovery power limit and recovery torque limit using data collected by the vehicle positioning unit and the generator sensor; obtain the master cylinder pressure value output by the brake pressure sensor, and calculate the actual recovery power based on the recovery potential value through a proportional-integral controller; construct a state space based on the vehicle speed value, battery status value and recovery potential value, use deep reinforcement learning to optimize the actual recovery power, and output the motor braking torque command and the charging current limit command through a dual closed-loop controller.

[0034] Exemplarily, a vehicle speed sensor is used to collect the real-time vehicle speed value and an acceleration sensor is used to collect the vehicle motion state, the battery temperature and remaining capacity values ​​are obtained from the battery manager, a support vector machine regressor is input for the coasting condition and the braking condition, the recovery power limit and the recovery torque limit are calculated based on the Gaussian kernel function, and the recovery basic parameters are obtained by combining the charging power threshold fed back by the battery manager. According to the vehicle mass parameter and the rolling resistance coefficient, the road slope value is obtained from the vehicle positioning unit, the speed change rate within a 5-second time window is collected by the vehicle speed sensor, the generator speed sensor and the torque sensor are used to measure the speed and torque respectively, the real-time kinetic energy loss and the recovery power product are multiplied to obtain the recovery potential value, and the recovery power upper limit is calculated by the proportional integral regulator. The master cylinder pressure value is obtained from the brake pressure sensor, the hydraulic brake ratio is dynamically adjusted in the range of 0 to 0.8, the actual recovery power is calculated based on the remaining capacity and temperature parameters fed back by the battery manager, the corresponding torque signal is output through the motor controller, and the deviation between the recovery potential value and the actual recovery power is adjusted. The state space is constructed according to the vehicle speed value, battery state value and recovery potential value. Deep reinforcement learning is used to optimize the torque output and charging current. The maximum recovery power is set as the reward function. The motor braking torque command and the charging current limit command are output through the dual closed-loop controller to achieve the combined optimization of the recovery parameters. The integral term and the proportional term of the proportional integral controller are performed online parameter tuning. The vehicle speed and vehicle motion state reflect the basic conditions for energy recovery. When the vehicle speed is 80 kilometers per hour, the acceleration sensor measures a longitudinal deceleration of 0.5 meters per square second, indicating that the vehicle is in a sliding deceleration state. At this time, the battery manager feedback battery temperature is 35 degrees Celsius and the remaining capacity is 75%. The support vector machine regressor calculates the recovery power limit of 32 kilowatts and the recovery torque limit of 160 Newton meters based on these data. Under the conditions of a vehicle mass of 2000 kilograms and a rolling resistance coefficient of 0.012, when coasting on a 3-degree uphill section, the vehicle speed drops from 80 to 70 kilometers per hour within 5 seconds, the generator speed drops from 4800 to 4200 revolutions per minute, and the torque remains at 120 Newton meters. The calculation of kinetic energy loss shows that the process loses 15 kilojoules of mechanical energy, and the measured recovery power is 28 kilowatts. The proportional-integral regulator sets the recovery power upper limit to 30 kilowatts to avoid exceeding the battery charging capacity. The parameter configuration under braking conditions needs to balance the recovery efficiency and braking performance. When the master cylinder pressure is 4 MPa, the controller sets the hydraulic braking ratio to 0.6, and the motor brake supplements the remaining 0.4 braking force. When the battery temperature rises to 45 degrees Celsius and the remaining capacity reaches 90%, the actual recovery power automatically drops to 15 kilowatts, and the motor output torque is correspondingly reduced to 80 Nm. At this time, the hydraulic braking ratio is increased to 0.8 to ensure stable braking force. The deep reinforcement learning optimizer maps parameters such as vehicle speed, battery status, and recovery potential to a 12-dimensional state space, and calculates the optimal control strategy based on the principle of obtaining 1 unit reward for every kilowatt of recovery power.Under high-speed conditions, when the vehicle speed is 120 kilometers per hour, the battery temperature is 40 degrees Celsius, and the recovery potential value is 45 kWh, the dual closed-loop controller outputs a 200 Nm torque command and a 100 ampere charging current limit command. The controller proportional coefficient is set to 2.5, and the integral time constant is set to 0.8 seconds to achieve stable power tracking control. In downhill conditions in mountainous areas, continuous heavy recovery causes the battery temperature to rise. At this time, the optimizer automatically reduces the charging current to 60 amperes and increases the hydraulic braking ratio. In cold weather conditions, the charging power is limited when the battery temperature is 5 degrees Celsius. The optimizer accordingly reduces the motor torque output to 0.4 times the rated value, increases the hydraulic braking force compensation, and maintains the smoothness of the braking force through online parameter setting. In the stop-and-go scenario in urban conditions, frequent braking opportunities will prompt the optimizer to adopt a mild recovery strategy, limiting the torque to less than 120 Nm to avoid motor overheating.

[0035] S105. Determine the driver's intention of acceleration, deceleration or constant speed according to the operating characteristics of the accelerator and brake, where the operating characteristics refer to the strength of pressing the accelerator pedal and the brake pedal, and use the determined driver's intention as a basis for dynamically adjusting the energy recovery strategy.

[0036] The throttle and brake pedal displacement data are collected through the pedal displacement sensor, the pedal feedback force data is measured through the torque sensor, and the longitudinal acceleration data is obtained through the acceleration sensor; the acceleration intention is determined when the longitudinal acceleration is greater than the specified threshold and the throttle pedal displacement is greater than the set value, the deceleration intention is determined when the longitudinal acceleration is less than the negative threshold and the brake pedal displacement is greater than the set value, and the uniform speed intention is determined when the longitudinal acceleration is in the threshold range and the pedal displacement is less than the set value; the pedal force data and the vehicle speed data are mean filtered using a sliding time window to obtain the change rate of the force curve and the acceleration curve within the sampling period; according to the positive value of the force change rate under the acceleration intention and the negative value of the force change rate under the deceleration intention, a proportional-integral controller is used to adjust the recovery power output value.

[0037] Exemplarily, a pedal displacement sensor is used to collect the displacement values ​​of the accelerator and brake pedals every 10 milliseconds, the pedal feedback force is measured by a torque sensor, the longitudinal acceleration of the vehicle is obtained from the acceleration sensor, and a long short-term memory neural network is used to extract features from the three-dimensional input data of pedal displacement, force, and acceleration. The driver operation sequence is obtained based on the training of the labeled data collected from the actual vehicle, and the input data is subjected to a low-pass filter processing with a sampling frequency of 50 Hz. The intention is identified according to the longitudinal acceleration sequence, and the acceleration intention is determined when the acceleration is greater than 0.3 meters per square second and the accelerator pedal displacement is greater than 5 millimeters, the deceleration intention is determined when the acceleration is less than negative 0.3 meters per square second and the brake pedal displacement is greater than 5 millimeters, and the uniform speed intention is determined when the acceleration is between positive and negative 0.3 meters per square second and the pedal displacement is less than 5 millimeters. The operation sequence is marked with the intention determination result. A 2-second sliding time window is used to perform mean filtering on the pedal force data and vehicle speed data. Combined with the longitudinal acceleration signal, the change rate of the force curve and the acceleration curve in the same sampling period is calculated. The force change rate and acceleration change rate curves are subjected to noise suppression through a Butterworth filter, and the Pearson correlation coefficient is used to calculate the degree of correlation between the force feature and the acceleration feature. According to the positive value of the force change rate under the acceleration intention and the negative value of the force change rate under the deceleration intention, a proportional-integral controller is used to dynamically adjust the upper limit of the recovery power. The controller proportional coefficient is set to 2.5, and the integral time constant is set to 0.8 seconds. Based on the battery temperature and remaining capacity status fed back by the battery manager every 100 milliseconds, the recovery function is turned off when the acceleration intention is present, 0.8 times the maximum torque is output when the deceleration intention is present, and 0.3 times the maximum torque is output when the uniform speed intention is present. Driving intention recognition achieves accurate judgment through multi-dimensional sensor information fusion. When the vehicle speed is 80 kilometers per hour, the accelerator pedal moves quickly from the static position to the 15 mm position. The torque sensor measures the pedal feedback force of 2.5 Nm, and the acceleration sensor measures the longitudinal acceleration of 0.5 m / s2. This set of data is input into the neural network after low-pass filtering with a sampling frequency of 50 Hz, and is determined to be a typical acceleration intention feature. In urban road conditions, the driver lightly presses the brake pedal to slow down. The pedal displacement sensor measures the displacement value of 6 mm, and the torque sensor measures the feedback force of 1.8 Nm, resulting in a deceleration of negative 0.4 m / s2. In high-speed cruising conditions, both pedals are in the released state, the displacement values ​​are less than 2 mm, and the longitudinal acceleration is maintained within the range of plus or minus 0.2 m / s2, which meets the judgment criteria for uniform speed intention. The correlation analysis between operating force and acceleration reflects the characteristics of driving style. Within the 2-second sampling window, when the driver accelerates quickly, the pedal force change rate reaches 1.2 Nm per second, corresponding to an acceleration change rate of 0.3 m / s2 / s. After Butterworth filtering, a positive correlation coefficient of 0.85 is calculated. In the progressive braking condition, the pedal force increases at a rate of 0.8 Nm per second, resulting in a deceleration change rate of 0.2 m / s2 / s, with a correlation coefficient of 0.78.The energy recovery strategy under different driving intentions needs to be adaptively adjusted. When the acceleration intention is recognized, the energy recovery function is completely turned off to avoid affecting the vehicle's dynamic performance. When the driver steps on the brake pedal at a force change rate of 1.5 Nm per second, the controller outputs a 240 Nm recovery torque. At this time, the battery temperature is 38 degrees Celsius and the remaining capacity is 75%, which is in the best recovery condition. Maintain a basic recovery torque of 90 Nm in constant speed cruising conditions to achieve continuous energy recovery. When the battery temperature rises to 45 degrees Celsius, the proportional integral controller automatically reduces the recovery torque to 120 Nm, and further limits it to 72 Nm according to the state of 92% remaining capacity. Under frequent braking conditions on mountain roads, the controller dynamically adjusts the recovery force according to the force change characteristics to avoid excessive recovery torque affecting braking comfort. In cold weather conditions, when the battery temperature is below 5 degrees Celsius, the recovery torque is limited to 0.4 times the maximum value even under strong braking intentions to ensure battery safety.

[0038] S106. Dynamically adjust the intensity of the energy recovery strategy based on the real-time battery capacity status and the driver's intention. If the battery capacity status is higher than a preset threshold, reduce the recovery intensity. At the same time, determine the driver's intention based on the operating characteristics of the accelerator and brake. If the driver intends to accelerate, increase the recovery intensity.

[0039] A battery manager is used to obtain the voltage value and current value of the battery pack, and a real-time power value is obtained according to the product of the voltage value and the current value. The temperature value of the battery pack is obtained from the temperature sensor. If the voltage value is higher than a preset voltage threshold or the temperature value is higher than a preset temperature threshold, it is determined that the battery pack is in a high-capacity state. The accelerator pedal displacement value and the brake pedal displacement value are obtained through a pedal displacement sensor, and the accelerator pedal displacement value and the brake pedal displacement value are processed using a long short-term memory neural network, and the driving intention is determined according to the acceleration value. A support vector regression model is established according to the battery pack voltage value, the temperature value and the capacity state value, and a recovery torque reference value is output through the regression model. The speed value is obtained from the motor speed sensor, and a proportional-integral controller is used to process the speed value and the driving intention, and a recovery torque command value is output according to a preset proportional coefficient and an integral time constant.

[0040] Exemplarily, a battery manager is used to collect the voltage and current values ​​of the battery pack every 100 milliseconds, and the real-time power is obtained by multiplying the voltage and current. The temperature value of the battery pack is obtained from the temperature sensor. When the voltage is higher than 4.1 volts per cell or the temperature is higher than 45 degrees Celsius, it is determined to be in a high-capacity state. When the voltage is lower than 3.6 volts per cell or the temperature is lower than 5 degrees Celsius, it is determined to be in a low-capacity state. The recovery torque limit is corrected in combination with the temperature value. Based on the throttle and brake pedal data collected by the pedal displacement sensor every 10 milliseconds, the pedal displacement and operation force are feature extracted through a long short-term memory neural network. A driving intention identifier is obtained based on the labeled data training. An acceleration greater than 0.3 meters per square second is determined as an acceleration intention, an acceleration less than negative 0.3 meters per square second is determined as a deceleration intention, and an acceleration in the range of positive and negative 0.3 meters per square second is determined as a uniform speed intention. The support vector regressor is used to construct the regeneration torque mapping relationship for the battery capacity state. The three characteristic quantities of voltage value, temperature value and capacity state value are input. When the battery is in high capacity state, the regeneration torque is reduced to 0.4 times the maximum value, and when the battery is in low capacity state, the regeneration torque is increased to 0.8 times the maximum value. The regeneration torque is adjusted in percentage in combination with the battery temperature, and the regeneration torque is reduced by 0.1 times for every 5 degrees Celsius decrease. The current speed value is obtained from the motor speed sensor. Combined with the driving intention judgment result, the proportional integral controller is used to output the regeneration torque command. The controller proportional coefficient is set to 2.5, and the integral time constant is set to 0.8 seconds. When the acceleration intention is, the regeneration torque is increased to 0.9 times the maximum value, and when the deceleration intention is, it is reduced to 0.5 times the maximum value, and when the uniform speed intention is maintained at 0.3 times the maximum value. The torque and charging current are adjusted every 20 milliseconds through closed-loop control. The battery capacity state monitoring realizes accurate judgment through high-frequency sampling. When the voltage of any battery in the 96-cell battery pack reaches 4.15 volts, the high capacity judgment is triggered. At this time, the battery temperature sensor shows 42 degrees Celsius, and the real-time power calculation shows that the charging power reaches 35 kilowatts. In cold weather conditions, the battery pack temperature drops to 3 degrees Celsius, and the single cell voltage drops to 3.55 volts, entering a low-capacity state, and the charging power drops to 15 kilowatts. Driving intention recognition is based on real-time operation feature analysis. Under highway conditions, the driver quickly presses the accelerator pedal, the pedal position moves from 0 to 25 mm, the operation force reaches 3.5 Nm, and the vehicle generates an acceleration of 0.4 meters per square second. The intention recognizer obtained by the neural network training based on 1,000 hours of labeled data accurately captures this acceleration intention feature. When slowing down on urban roads, the brake pedal moves at a rate of 2 mm per second, generating a deceleration of negative 0.2 meters per square second, which meets the characteristics of a constant speed cruise condition. The regenerative torque mapping relationship is dynamically adjusted with the battery state. When the battery pack temperature is 40 degrees Celsius and the single cell voltage is 4.12 volts, the support vector regressor outputs a regenerative torque coefficient of 0.45, and the actual regenerative torque drops to 135 Nm.As the vehicle continues to run, when the battery temperature rises to 48 degrees Celsius, the regenerative torque is further reduced to 120 Nm. In winter conditions, the battery temperature drops to 0 degrees Celsius, causing the regenerative torque coefficient to drop to 0.3, and the actual regenerative torque is only 90 Nm. The coordinated control of driving intention and battery status is reflected in multiple scenarios. When the driver quickly steps on the brake pedal on a downhill section, generating a deceleration of 0.8 meters per square second, and the battery is in a low-capacity state, the controller increases the regenerative torque to 270 Nm and the charging current reaches 80 amps. In the acceleration condition with high battery capacity, even if the driver steps on the accelerator deeply, the regenerative torque is limited to less than 120 Nm. The proportional-integral controller performs closed-loop regulation every 20 milliseconds, and when the battery temperature is 35 degrees Celsius, the charging current is stably controlled at 65 amps. As the battery temperature rises, the charging current is gradually reduced to 45 amps. Under the condition of frequent speed changes on mountain roads, the controller quickly switches the regenerative torque output based on the driving intention to achieve maximum energy recovery while ensuring battery safety.

[0041] S107. Apply the dynamic adjustment method of the energy recovery strategy intensity to the target energy recovery strategy combination scheme, and monitor the battery's state of charge and temperature parameters at the same time to calculate the recovery power of the energy recovery strategy before and after the dynamic adjustment of the intensity, obtain the change in the recovery power, and switch the strategy in the target energy recovery strategy combination scheme in real time.

[0042] The three-dimensional sequence data of the voltage signal, current signal and temperature signal of the battery pack terminal are collected, and the initial recovery power value is obtained by processing the three-dimensional sequence data through a long short-term memory neural network; the temperature change rate and the state of charge change rate are calculated using a sliding time window according to the initial recovery power value and the state of charge threshold; the recovery coefficient is adjusted according to the temperature change rate and the state of charge change rate through a proportional-integral controller, and the generator output torque is obtained from the torque sensor; the gear range is divided according to the temperature change rate and the state of charge change rate, and if the recovery power change exceeds a preset threshold, the recovery strategy is switched according to the gear range, and the recovery torque is adjusted through a proportional-integral controller.

[0043] Exemplarily, a battery manager is used to collect the terminal voltage and current values ​​of the battery pack every 100 milliseconds, and the surface temperature value of the battery pack is obtained from the temperature sensor. The terminal voltage, current, and temperature three-dimensional sequence data are processed through a long short-term memory neural network. The number of input layer neurons is set to 32, and the number of hidden layer neurons is set to 64. A 50 Hz low-pass filter is performed on the sampled data. The initial recovery power is calculated based on the battery parameters when the recovery intensity is 0.5 times the nominal value, and the power value is stored in the data register. According to the current recovery power value and the state of charge threshold, a 10-second sliding time window is used to calculate the battery temperature change rate and the state of charge change rate. The change rate curve is processed by a polynomial fitter, and the corresponding recovery coefficient is read from the gear switching table in combination with the gear parameters in the recovery power combination scheme. If the temperature change rate exceeds 5 degrees Celsius per minute or the state of charge change rate exceeds 5% per minute, the dynamic adjustment of the recovery coefficient is triggered. A proportional-integral controller is used to adjust the recovery coefficient in real time. The controller proportional coefficient is set to 2.5, and the integral time constant is set to 0.8 seconds. The generator output torque is obtained through a torque sensor that samples once every 20 milliseconds, and the input power and output power are read from the power converter. The adjusted recovery power is dynamically calculated and compared with the initial recovery power value stored in the data register. According to the recovery power change curve and the battery charge state, a multiple comparator is used to grade the recovery power change rate. When the recovery power changes by more than 20%, the strategy switching is triggered. For the three temperature ranges of temperature less than 10 degrees Celsius, temperature between 10 and 40 degrees Celsius, and temperature greater than 40 degrees Celsius, combined with the three capacity ranges of charge state less than 20%, charge state between 20% and 80%, and charge state greater than 80%, the recovery strategy gear switching is executed, and the smooth transition of the recovery torque is achieved through the proportional-integral controller. Dynamic monitoring of battery parameters is precisely controlled through high-frequency sampling. When the battery pack terminal voltage is 3.8 volts and the charging current is 25 amperes, the initial recovery power is calculated to be 95 kilowatts, and the battery surface temperature is 35 degrees Celsius. The 32 input neurons of the neural network receive the historical data of voltage, current, and temperature respectively, and output the recovery power prediction value after processing by 64 hidden layer neurons. The 50 Hz low-pass filter removes the high-frequency noise in the sampled signal and retains the true change trend of the battery parameters. The dynamic adjustment of the recovery power is based on the battery state change rate. Within the 10-second time window, the battery temperature rises from 35 degrees Celsius to 36 degrees Celsius, and the temperature change rate is 6 degrees Celsius per minute, triggering the recovery factor adjustment mechanism. At the same time, the state of charge rises from 75% to 76.5%, with a change rate of 9% per minute, exceeding the preset threshold. The polynomial fitter smoothes these change rate curves and reads the corresponding recovery factor 0.8 from the recovery gear table.The closed-loop control of the recovered power is achieved through real-time feedback. The proportional-integral controller samples the generator torque every 20 milliseconds. When the torque is 250 Nm and the speed is 3000 rpm, the power converter input power is 78 kW. Considering the conversion efficiency of 95%, the output power is 74 kW. Compared with the initial recovered power, the power change rate reaches 22%, triggering the gear switching condition. The dynamic switching of the recovery strategy is based on multiple constraints. When the battery temperature drops to 8 degrees Celsius, the low-temperature interval control mode is entered, and the recovered power is reduced to 0.3 times the nominal value. Under the operating conditions of battery temperature of 38 degrees Celsius and state of charge of 85%, the recovered power is limited to 0.4 times the nominal value at the same time. Under the standard operating conditions of 35 degrees Celsius and state of charge of 60%, the 0.8 times nominal recovered power is maintained. When switching between different gears, the proportional-integral controller produces a smooth torque transition process to avoid sudden changes. In the uphill section of the electric vehicle, due to the reduction of energy recovery opportunities, the controller keeps the gear stable. On downhill sections, due to the sudden increase in recovery power, the controller quickly reduces the recovery coefficient to prevent battery overcharging. In stop-and-go urban conditions, frequent braking causes fluctuations in recovery power, and the controller achieves dynamic balance of power by adjusting torque output. In winter low temperature conditions, when the battery temperature is 5 degrees Celsius, even under strong braking requirements, the controller will limit the recovery power to the minimum gear to prioritize battery safety.

[0044] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations. In addition, the various different embodiments of the present invention can also be combined arbitrarily, as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A method for generating a control strategy for improving the performance of a lithium battery hybrid motorcycle, characterized in that: The method comprises: Vehicle dynamics analysis is performed based on the real-time acquired lithium battery capacity status, motor energy recovery efficiency, vehicle driving status and driver operation data to obtain instantaneous energy consumption, and combined with the actual speed of the vehicle, the energy recovery potential value of the current vehicle is calculated; The vehicle driving state includes a coasting state and a braking state. If the vehicle is currently in a coasting state, a preset coasting energy recovery mode is activated; if the vehicle is currently in a braking state, a preset braking energy recovery mode is activated; The vehicle speed and braking force are collected through sensors, and the energy recovery strategies in the gliding and braking states are subdivided according to the speed and braking force. In the gliding state, the recovery strategy is divided into light, medium or heavy recovery strategies according to the length of the gliding distance, and in the braking state, the recovery strategy is divided into gradual, stable or sharp recovery strategies according to the braking force; According to the subdivision results of energy recovery strategies under the coasting and braking states, different energy recovery strategies are combined and optimized, and the vehicle speed, battery charging state, and energy recovery potential value parameters are input to obtain the target energy recovery strategy combination scheme; Determine the driver's intention to accelerate, decelerate or maintain a constant speed based on the operating characteristics of the accelerator and brake pedals, where the operating characteristics refer to the strength of the accelerator and brake pedals, and use the determined driver's intention as the basis for dynamic adjustment of the energy recovery strategy; Dynamically adjust the intensity of the energy recovery strategy based on the real-time battery capacity status and driver intention. If the battery capacity status is higher than the preset threshold, the recovery intensity is reduced. At the same time, the driver's intention is determined by the operating characteristics of the accelerator and brake. If the driver intends to accelerate, the recovery intensity is increased. The dynamic adjustment method of the energy recovery strategy intensity is applied to the target energy recovery strategy combination scheme, and the battery's state of charge and temperature parameters are monitored at the same time to calculate the recovery power of the energy recovery strategy before and after the dynamic adjustment of the intensity, obtain the change in the recovery power, and switch the strategy in the target energy recovery strategy combination scheme in real time.

2. The method according to claim 1, characterized in that The vehicle dynamics analysis is performed based on the real-time acquired lithium battery capacity state, motor energy recovery efficiency, vehicle driving state and driver operation data to obtain instantaneous energy consumption, and combined with the acquired actual speed of the vehicle, the energy recovery potential value of the current vehicle is calculated, including: A first power value is obtained according to the product of the battery terminal voltage and the charging current, and a battery capacity state value is obtained by fitting the first power value using a least square method; A road slope value is calculated using global positioning satellite data, and a linear Kalman filter operation is performed on the road slope value and load mass data collected by a vehicle-mounted gravity sensor to obtain a second power value; If the motor recovery efficiency is higher than a preset threshold, the power recovery ratio value is obtained through discrete integration operation according to the battery capacity status value and the vehicle speed data in the vehicle controller bus; The power recovery ratio value is subjected to deviation standardization processing to obtain standardized data, a long short-term memory neural network is used to construct a time series prediction model for the standardized data to obtain an energy consumption prediction value, and the energy recovery potential value is calculated by comparing the instantaneous vehicle speed with the acceleration value.

3. The method according to claim 1, characterized in that The vehicle driving state includes a coasting state and a braking state. If the vehicle is currently in a coasting state, a preset coasting energy recovery mode is activated. If the vehicle is currently in a braking state, a preset braking energy recovery mode is activated, including: Obtain the three-axis angular velocity and acceleration values ​​collected by the vehicle-mounted inertial sensor, and determine the vehicle's motion state based on the pedal opening value fed back by the brake pedal displacement sensor and the wheel Hall speed sensor signal; According to the vehicle motion state, the speed and torque signals fed back by the motor controller are obtained from the vehicle communication bus, and the driving resistance is calculated through the values ​​collected by the vehicle posture sensor and the atmospheric pressure sensor; If the accelerator pedal opening is zero and the brake pedal opening is less than the brake pedal opening threshold, the coasting recovery power limit is determined using the voltage and current limits fed back by the battery management unit; If the brake pedal opening is greater than the brake pedal opening threshold, a proportional-integral controller is used to adjust the distribution ratio of hydraulic braking and electric braking force according to the master cylinder pressure value collected by the brake pressure sensor and the remaining capacity state fed back by the battery management unit.

4. The method according to claim 1, characterized in that: The vehicle speed and braking force are collected by the sensor, and the energy recovery strategies in the sliding and braking states are subdivided according to the vehicle speed and braking force. In the sliding state, the recovery strategies are divided into light, medium or heavy recovery strategies according to the sliding distance, and in the braking state, the recovery strategies are divided into gradual, stable or sharp recovery strategies according to the braking force, including: Obtaining a pedal travel value from a brake pedal displacement sensor and obtaining a real-time vehicle speed value through a vehicle speed sensor, and determining that the vehicle is in a coasting state if the vehicle speed value is lower than a preset speed threshold and the braking torque is lower than a preset torque threshold; A sliding distance value is calculated according to the vehicle position coordinate sequence, and the sliding distance value is classified by using a neural network combined with an air resistance coefficient and a vehicle body mass, and a light, medium and heavy recovery torque is determined based on the battery remaining capacity and the sliding distance classification result; The pedal stroke value is subjected to wavelet decomposition to obtain a stroke change rate, and the progressive braking, stable braking or sudden braking type is determined according to the stroke change rate in combination with the wheel deceleration value and the braking torque value; A double-layer neural network is used to calculate the battery charging power according to the gear position of the recovery torque and the braking type, and based on the motor speed and power conversion ratio, the motor braking torque corresponding to the gear position is output through a regulator; It also includes: obtaining a preset braking force threshold range and obtaining real-time braking force data of the vehicle, determining the current braking force type, which includes progressive, stable and sharp types, and obtaining corresponding recovery strategy parameters from a pre-established braking force-recovery strategy mapping table for different braking force types, and determining the generator torque output and battery charging current limit under the current braking state based on the obtained recovery strategy parameters.

5. The method according to claim 4, characterized in that The method of obtaining a preset braking force threshold range and real-time braking force data of the vehicle, determining the current braking force type, the braking force types including progressive, stable and sharp, obtaining corresponding recovery strategy parameters from a pre-established braking force-recovery strategy mapping table for different braking force types, and determining the generator torque output and battery charging current limit under the current braking state according to the obtained recovery strategy parameters, includes: A brake pedal displacement sensor is used to collect a brake pedal displacement signal, and a brake force type is obtained according to the displacement signal and a brake master cylinder pressure signal through deep neural network processing; Obtaining a pressure curve slope value according to a brake master cylinder pressure signal, and determining a braking condition type by comparing the pressure curve slope value with a preset slope threshold; Reading corresponding recovery parameters from a preset force recovery mapping table according to the braking condition type, and determining the motor output torque value and the charging current limit value according to the recovery parameters; The motor output torque is collected through a torque sensor, the battery pack parameters are obtained from a battery manager, and a proportional-integral controller is used to perform closed-loop regulation of the motor torque and the charging current according to the motor output torque and the battery pack parameters.

6. The method according to claim 1, characterized in that The method combines and optimizes different energy recovery strategies according to the subdivision results of the energy recovery strategies under the coasting and braking states, and inputs vehicle speed, battery charging state, and energy recovery potential value parameters to obtain a target energy recovery strategy combination scheme, including: Collecting real-time data output by a vehicle speed sensor, an acceleration sensor and a battery manager, and obtaining a recovery power limit and a recovery torque limit through a support vector machine regressor according to the real-time data output; According to the recovery power limit and the recovery torque limit, the recovery potential value is calculated using data collected by the vehicle positioning unit and the generator sensor; Obtaining a master cylinder pressure value output by a brake pressure sensor, and calculating actual recovery power according to the recovery potential value through a proportional-integral controller; A state space is constructed according to the vehicle speed value, battery state value and recovery potential value, deep reinforcement learning is used to optimize the actual recovery power, and the motor braking torque command and the charging current limit command are output through a dual closed-loop controller.

7. The method according to claim 1, characterized in that The method of judging the driver's intention of acceleration, deceleration or constant speed according to the operation characteristics of the accelerator and brake, wherein the operation characteristics refer to the strength of stepping on the accelerator pedal and the brake pedal, and using the judged driver's intention as the basis for dynamically adjusting the energy recovery strategy, includes: The accelerator and brake pedal displacement data are collected through the pedal displacement sensor, the pedal feedback force data is measured through the torque sensor, and the longitudinal acceleration data is obtained through the acceleration sensor; The acceleration intention is determined when the longitudinal acceleration is greater than the specified threshold and the accelerator pedal displacement is greater than the set value; the deceleration intention is determined when the longitudinal acceleration is less than the negative threshold and the brake pedal displacement is greater than the set value; the uniform speed intention is determined when the longitudinal acceleration is within the threshold range and the pedal displacement is less than the set value; The pedal force data and vehicle speed data are mean filtered using a sliding time window to obtain the change rate of the force curve and acceleration curve within the sampling period; A proportional-integral controller is used to adjust the recovery power output value according to the positive value of the force change rate under the acceleration intention and the negative value of the force change rate under the deceleration intention.

8. The method according to claim 1, characterized in that The energy recovery strategy intensity is dynamically adjusted according to the real-time battery capacity status and the driver's intention. If the battery capacity status is higher than a preset threshold, the recovery intensity is reduced. At the same time, the driver's intention is determined by combining the accelerator and brake operation characteristics. If the driver intends to accelerate, the recovery intensity is increased, including: A battery manager is used to obtain a voltage value and a current value of the battery pack, a real-time power value is obtained according to the product of the voltage value and the current value, and a temperature value of the battery pack is obtained from a temperature sensor. If the voltage value is higher than a preset voltage threshold or the temperature value is higher than a preset temperature threshold, it is determined that the battery pack is in a high capacity state; The accelerator pedal displacement value and the brake pedal displacement value are obtained by using a pedal displacement sensor, the accelerator pedal displacement value and the brake pedal displacement value are processed by using a long short-term memory neural network, and the driving intention is determined according to the acceleration value; Establishing a support vector regression model according to the battery pack voltage value, the temperature value and the capacity state value, and outputting a recovery torque reference value through the regression model; A speed value is obtained from a motor speed sensor, a proportional-integral controller is used to process the speed value and the driving intention, and a recovery torque command value is output according to a preset proportional coefficient and an integral time constant.

9. The method according to claim 1, characterized in that: The method of dynamically adjusting the intensity of the energy recovery strategy is applied to the target energy recovery strategy combination scheme, and the state of charge and temperature parameters of the battery are monitored at the same time to calculate the recovery power of the energy recovery strategy before and after the dynamic adjustment of the intensity, obtain the change of the recovery power, and perform strategy switching in the target energy recovery strategy combination scheme in real time, including: Collecting three-dimensional sequence data of voltage signals, current signals and temperature signals at the battery pack terminals, and processing the three-dimensional sequence data through a long short-term memory neural network to obtain an initial recovery power value; According to the initial regenerative power value and the state of charge threshold, a temperature change rate and a state of charge change rate are calculated using a sliding time window; The recovery coefficient is adjusted according to the temperature change rate and the charge state change rate by a proportional-integral controller, and the generator output torque is obtained from the torque sensor; The gear interval is divided according to the temperature change rate and the charge state change rate. If the recovery power change exceeds a preset threshold, the recovery strategy is switched according to the gear interval, and the recovery torque is adjusted by a proportional-integral controller.

Images