Range extender power generation system and control method

Abstract

The invention discloses a range extender power generation system and a control method, and relates to the technical field of power systems. The system comprises an energy storage module, a driving module, a range extender module and a control module, the control module combines model predictive control and reinforcement learning and optimizes energy distribution, so that the system can automatically select an optimal strategy according to different driving working conditions, the energy utilization efficiency is further improved, the total energy consumption is reduced, and the driving efficiency is improved. The efficient cooperative work of the internal combustion engine, the generator and the driving motor is realized; self-adaptive gear shifting control is carried out based on real-time vehicle working conditions, path prediction information and motor efficiency, and it is ensured that a driving motor always works in an efficient area. Besides, the system also has fault diagnosis and fault-tolerant control functions, can detect the performance deviation of the engine and the generator in real time, automatically enters a fault-tolerant mode when a fault occurs, and ensures that the vehicle can safely run to a preset place even under the fault condition by cutting off or adjusting the output of the engine.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and more particularly, to a range extender power generation system and a control method therefor. Background Art

[0002] With the development of new energy vehicles, extended-range electric vehicles (EREVs) have become an effective way to improve vehicle range due to their characteristics of pure electric drive and fuel power generation for energy replenishment. However, most traditional EREV power systems adopt a structure of a single motor + a single-stage reducer, having deficiencies such as the motor efficiency being difficult to cover the optimal value under all working conditions and the power performance being limited. In addition, the range extender (usually composed of an internal combustion engine) is completely decoupled from the wheels in the traditional series architecture. Although the control is convenient, the energy conversion loss through the fuel - power generation - electric link is relatively large; in the parallel architecture, the engine directly participates in driving, reducing the conversion loss, but the control is complex and may introduce shift shocks.

[0003] In the prior art, there is a lack of in-depth research and application of a multi-motor multi-gear extended-range power architecture, and the performance advantages of each component cannot be fully exerted. At the same time, in the actual operation of the vehicle, how to intelligently manage the energy distribution of the battery and the range extender according to complex and changeable working conditions to meet the power demand with the lowest fuel consumption is also a challenge faced by technicians; traditional rule-based energy management strategies are difficult to adapt to different driving environments, while model-based optimization strategies lack real-time performance. Another problem is that the reliability of the range extender is crucial for the whole vehicle. In particular, once the performance of the engine - generator system is abnormal, it will directly affect the vehicle range and safety. However, current fault diagnosis is mostly limited to the parameters of the engine itself, and the role of the motor signal in fault detection is not fully considered, and some potential fault signs cannot be detected in time.

[0004] In view of the above deficiencies, it is necessary to provide a new technical solution to introduce multi-motor drive and multi-gear transmission into the extended-range electric drive system, and integrate advanced energy management algorithms and fault diagnosis mechanisms, so as to improve the system efficiency, power performance and reliability, and meet the requirements of new energy vehicles for range and performance in multiple scenarios. Summary of the Invention

[0005] The purpose of the present invention is to provide a range extender power generation system and a control method therefor, which combine dual-motor drive, multi-gear transmission, intelligent energy management control and fault self-diagnosis technology to achieve more excellent energy management efficiency and smoother power output compared with the prior art, and significantly improve the performance and reliability of extended-range electric vehicles.

[0006] The present invention achieves the above purpose through the following technical solutions:

[0007] A range extender power generation system, the system comprising:

[0008] Energy storage module, including power battery and battery management system, for controlling power input and output, and providing state of charge (SOC) of the battery;

[0009] Drive module, including a first drive motor, a second drive motor and a multi-speed transmission. The first drive motor and the second drive motor are respectively connected to the odd-numbered gear shaft and the even-numbered gear shaft of the multi-speed transmission, for converting electric power into driving torque, which is transmitted to the vehicle drive wheels after being speeded up and torque-increased by the multi-speed transmission to generate driving force output;

[0010] Range extender module, including an internal combustion engine and a generator mechanically connected to the internal combustion engine; the internal combustion engine is mechanically coupled to the even-numbered gear shaft of the multi-speed transmission through a clutch, and the clutch is switched to the engaged state under the control of the control module to transmit the output torque of the internal combustion engine, or switched to the disengaged state to disconnect the mechanical connection between the internal combustion engine and the multi-speed transmission; the output terminal of the generator is electrically connected to the power battery, the first drive motor and the second drive motor through a power electronic device, for charging the energy storage module or directly supplying power to the first drive motor and the second drive motor when the engine is working;

[0011] Control module, electrically connected to the internal combustion engine, the generator, the first drive motor, the second drive motor and the multi-speed transmission, for comprehensively performing path condition prediction, energy distribution control, health state estimation and drive strategy scheduling.

[0012] As a preferred embodiment of the present invention, the control module specifically includes:

[0013] Path prediction unit, for predicting path condition information within a future set time domain based on road information of the navigation system, combined with historical driving data, through a machine learning algorithm;

[0014] Energy management unit, for constructing an energy optimization control model by combining model predictive control and reinforcement learning according to vehicle real-time state parameters, state of charge (SOC) of the battery and predicted path condition information, and generating an optimized control instruction, the optimized control instruction including the start-stop timing of the internal combustion engine, the power generation power of the generator, and the torque distribution of the first drive motor and the second drive motor, and the shift strategy of the multi-speed transmission, so as to achieve the comprehensive optimum of energy utilization efficiency and power response performance;

[0015] Health state estimation unit, for collecting the shaft power of the generator and system state signals, combining the optimized control instruction, constructing an extended Kalman filter model, taking the shaft power of the generator as the observation input in the augmented state space, and performing real-time estimation of the health parameters of the internal combustion engine. By comparing the true value and the estimated value of the health parameters, it is judged whether there is a fault in the internal combustion engine or the generator, and a fault flag is output when the judgment result is abnormal;

[0016] A drive strategy scheduling unit, which is used to select a vehicle operation mode according to the health status estimation result, including an extended range mode, a pure electric mode or a fault tolerance mode, and feed back the scheduling result to the energy management unit for updating the control optimization boundary conditions.

[0017] As a preferred solution of the present invention, the rolling optimization objective function J of the energy optimization control model is expressed as:

[0018]

[0019] In the formula, m is the current optimization starting step, p is the prediction time domain length of the model predictive control; FC(k) is the fuel consumption of the internal combustion engine at the k-th step; SOC(k) is the current state of charge of the battery, and ρ SOC is the SOC error penalty weight; S e (k), S e (k - 1) are the start-stop states of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, and ρ Se is the start-stop switching penalty weight of the internal combustion engine; T e (k), T e (k - 1) are the output torques of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, and ρ Te is the torque change penalty weight of the internal combustion engine; Sm1(k), Sm1(k - 1) are the working states of the first drive motor at the k-th step and the (k - 1)-th step respectively, and ρ Sm1 is the state switching penalty weight of the first drive motor; ΔT m1 is the torque change value of the first drive motor at the k-th step, and ρ Tm1 is the torque change penalty weight of the first drive motor; Sm2(k), Sm2(k - 1) are the working states of the second drive motor at the k-th step and the (k - 1)-th step respectively, and ρ Sm2 is the state switching penalty weight of the second drive motor; ΔT m2 is the torque change value of the second drive motor at the k-th step, and ρ Tm2 is the torque change penalty weight of the second drive motor; G(k - 1), G(k) are the current gears at the k-th step and the (k - 1)-th step respectively, and ρ G is the shift frequency penalty weight.

[0020] As a preferred solution of the present invention, the energy management unit calls a policy network trained based on reinforcement learning and inputs the following information: including vehicle real-time operation state parameters, including current vehicle speed, acceleration and drive power request; state of charge of the battery SOC; predicted path condition information, including road speed limit and slope information; drive motor speed and efficiency mapping information;

[0021] Generate the dynamic correction amount of each penalty weight according to the above input information

[0022] Δρ = [Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G T , where T represents the transpose operation, and combined with the basic penalty weight vector to obtain the dynamic penalty weights of each item, that is:

[0023]

[0024] The reward function R(k) of the reinforcement learning policy network is defined as:

[0025]

[0026] In the formula, Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G and are the dynamic correction amounts and initial values of the SOC error penalty weight, the start-stop switching penalty weight of the internal combustion engine, the torque change penalty weight of the internal combustion engine, the state switching penalty weight of the first drive motor, the torque change penalty weight of the first drive motor, the state switching penalty weight of the second drive motor, the torque change penalty weight of the second drive motor, and the shift frequency penalty weight respectively; W is the correction amplitude control factor; SOC ref is the target state of charge of the battery; s ref is the expected vehicle state on the predicted path; s achieved is the current actual driving state value; ε is the state deviation tolerance threshold; γ1, γ2, γ3, γ4 are the weight coefficients of fuel economy, SOC stability, path tracking and shift smoothness respectively.

[0027] As a preferred embodiment of the present invention, the expression of the extended Kalman filter model is defined as follows:

[0028]

[0029] In the formula, are the estimated values of the health parameters at the i-th sampling moment and the (i-1)-th sampling moment respectively, including the compressor efficiency η c , the turbine efficiency η​t , the combustion degradation factor f c and the friction degradation factor f t ; A aug is the state transition matrix; B aug is the control input matrix, and u(t i ) is the control input quantity; F aug is the shaft power observation coupling matrix, and P shaft (t i ) is the measured value of the generator shaft power; K aug is the Kalman gain matrix calculated based on the predicted covariance and the observation error covariance; y(t i ) is the actually observed output value; C aug is the observation matrix;

[0030] During the health parameter estimation process, the estimation covariance matrix P aug (t i ) is synchronously calculated to measure the uncertainty of the health parameter estimation value;

[0031] If any of the following conditions are met:

[0032] Any health parameter estimation value deviates from the preset nominal value by more than the tolerance threshold;

[0033] The main diagonal element of the covariance corresponding to any health parameter is greater than the preset threshold, then mark the current state as a potential fault, output the fault flag f fault = 1, and execute the hierarchical fault-tolerant control strategy according to the fault level.

[0034] As a preferred solution of the present invention, the drive strategy scheduling unit introduces a delay confirmation mechanism before mode switching, requiring that the abnormal health parameter continuously satisfies the trigger condition for more than the set duration, and completes the switching confirmation by combining the historical state fluctuation statistics;

[0035] Range extender mode: If the system confirms that the current health parameter is within the normal range, there is no abnormal fluctuation, and no fault flag signal is generated, then maintain the range extender mode;

[0036] Fault-tolerant mode: If any health parameter estimation value deviates from the tolerance interval, or the estimation covariance exceeds the limit, but the system has not triggered a fault flag, then the system enters the fault-tolerant mode;

[0037] In the fault-tolerant mode, the following control constraint conditions are fed back to the energy management unit to update the control optimization boundary conditions, specifically including:

[0038] Limit the maximum allowable output power of the engine not to exceed 70% of the rated power;

[0039] Increase the minimum start-stop interval time of the engine;

[0040] Limit the torque change rate of the drive motor;

[0041] Limit the gear shifting frequency of the transmission;

[0042] Pure electric mode: If the fault flag f fault = 1, the system directly switches to the pure electric mode, disconnects the mechanical connection between the internal combustion engine and the multi - speed transmission, and is only powered by the first and second drive motors for operation, and records the fault event for subsequent maintenance processing.

[0043] As a preferred solution of the present invention, the multi - speed transmission is an automatic transmission mechanism with at least three gears or more, and the transmission ratios of each gear are preset according to the highest efficiency working ranges of the first and second drive motors;

[0044] When the current gear of the drive motor is about to enter the gear - shifting state, the control module judges the power demand trend based on the path condition prediction, vehicle acceleration request and motor speed, and the other drive motor advances or maintains its output torque in advance for drive compensation, and performs gear selection control according to the gear - shifting map or the adaptive gear - shifting algorithm based on the motor efficiency region.

[0045] As a preferred solution of the present invention, the adaptive gear - shifting algorithm takes the current vehicle speed, motor speed, motor output torque and path condition prediction data as inputs, combines the efficiency characteristics of the first and second drive motors at different operating points, constructs the mapping relationship between motor efficiency and load, and predicts the motor efficiency performance at each gear in real time for dynamically judging the target gear;

[0046] When the current drive motor is about to get out of its high - efficiency range, and there is a continuous climbing or accelerating trend in the predicted path, the control module calculates the comprehensive efficiency improvement amplitude of the next gear or adjacent gear within the predicted time period, and considers the cooperative output efficiency of the other drive motor in this gear. If there is an efficiency overlap region or the efficiency improvement amount exceeds the set threshold and meets the minimum holding time requirement, an early gear shift is triggered, and the other drive motor undertakes the torque compensation to ensure the continuity and smoothness of power output;

[0047] The gear - shifting strategy introduces a sliding - mode judgment mechanism. When the efficiency improvement amount is not enough to exceed the set dead zone or the path prediction trend is unstable, the current gear is kept unchanged.

[0048] A control method for an extended - range generator system, the method includes:

[0049] Collect the real - time state parameters of the vehicle, including the current vehicle speed, power request, state of charge SOC of the battery, and the predicted path condition information;

[0050] Based on real-time information and preset control strategies, coordinated control of the internal combustion engine, generator, and first and second drive motors is carried out based on model predictive control and reinforcement learning to optimize energy distribution;

[0051] When shifting gears, it is judged whether the current gear is about to deviate from the high-efficiency working range. If it is judged that the current drive motor is about to enter the low-efficiency working range, then based on path condition prediction and vehicle acceleration request, it is switched to the gear with the optimal efficiency in advance, and the motor output torque is compensated through an adaptive shifting algorithm to maintain the smoothness of power output; during the control process, according to real-time feedback and prediction data, the shifting timing is dynamically adjusted to ensure that the system operates in the optimal energy efficiency range;

[0052] When it is detected that the internal combustion engine or generator fails or its performance degrades, the control system enters the fault-tolerant mode, including cutting off or reducing the fuel supply of the internal combustion engine, and only the battery and the first and second drive motors are used to maintain the vehicle's driving, and the output power or speed is restricted appropriately to ensure the vehicle can safely drive to the predetermined location.

[0053] The beneficial effects of the present invention are as follows: making full use of the advantages of multi-motor collaborative work, by real-time monitoring and predicting the efficiency changes of the two drive motors, dynamically adjusting the working state of the motors, optimizing their efficiency regions, the system can minimize the fluctuations of the motor efficiency and achieve the maximization of the overall system energy efficiency, effectively improving the vehicle's endurance and power performance; through the combination of model predictive control (MPC) and the motor efficiency prediction model, efficient collaborative work of the internal combustion engine, generator and drive motors is realized, which can not only adjust the energy distribution according to real-time working conditions, but also optimize the shifting timing based on the working efficiency of the motors to ensure that the system always maintains the best energy efficiency performance under dynamic working conditions; the adaptive optimization of the shifting strategy not only reduces the energy loss during the shifting process, but also greatly improves the driving smoothness of the vehicle; real-time detection of the performance deviation of the engine and generator, and automatically entering the fault-tolerant mode when a failure occurs; combining path condition prediction with the vehicle's real-time working conditions, obtaining road condition information in advance, and adjusting the energy management strategy according to the prediction results, can effectively predict and respond to the possible working condition changes during driving, avoiding the limitations of relying on static control parameters in the traditional range extender system, and improving the flexibility and economy of energy management. Description of the Drawings

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0055] Among them:

[0056] Figure 1 is a system structure block diagram of the present invention;

[0057] Figure 2 A method flow chart of a control module in an embodiment of the present invention;

[0058] Figure 3 4 is a flow chart of a method in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0060] like Figure 1 As shown, an embodiment of the present invention provides a range extender power generation system, including an energy storage module, a drive module, a range extender module and a control module.

[0061] The energy storage module includes a power battery and a battery management system, which is used to control the input and output of electric energy and provide the battery state of charge SOC.

[0062] The drive module includes a first drive motor, a second drive motor and a multi-speed gearbox. The first drive motor and the second drive motor are respectively connected to the odd-numbered gear shaft and the even-numbered gear shaft of the multi-speed gearbox, and are used to convert electric power into drive torque, which is transmitted to the vehicle drive wheels after speed change and torque increase by the multi-speed gearbox to generate drive force output.

[0063] The multi-speed gearbox is an automatic transmission mechanism with at least three gears or more, and the transmission ratio of each gear is pre-set according to the highest efficiency working range of the first drive motor and the second drive motor;

[0064] When the gear corresponding to the current drive motor is about to enter the gear shift state, the control module determines the power demand trend based on the path condition prediction, vehicle acceleration request and motor speed, and the other drive motor increases or maintains its output torque in advance for drive compensation, and performs gear selection control according to the gear shift map or the adaptive gear shift algorithm based on the motor efficiency area.

[0065] Specifically, the adaptive shifting algorithm takes the current vehicle speed, motor speed, motor output torque and path condition prediction data as input, combines the efficiency characteristics of the first and second drive motors at different operating points, and constructs a mapping relationship between motor efficiency and load to predict the motor efficiency performance under each gear in real time, and is used to dynamically determine the target gear;

[0066] When the current drive motor is about to exit its high-efficiency range and there is a continuous uphill or accelerating trend in the predicted path, the control module calculates the comprehensive efficiency improvement of the next gear or adjacent gear within the predicted period, and considers the cooperative output efficiency of the other drive motor in this gear. If there is an efficiency overlap area or the efficiency improvement amount exceeds the set threshold and meets the minimum holding time requirement, an early gear shift is triggered, and the other drive motor undertakes torque compensation to ensure the continuity and smoothness of power output;

[0067] The gear shift strategy introduces a sliding mode judgment mechanism. When the efficiency improvement amount is not enough to exceed the set dead zone or the path prediction trend is unstable, the current gear is maintained.

[0068] The range extender module includes an internal combustion engine and a generator mechanically connected to the internal combustion engine;

[0069] The internal combustion engine is mechanically coupled to the even-numbered gear shaft of the multi-speed transmission through a clutch. The clutch is switched to the engaged state under the control of the control module to transmit the output torque of the internal combustion engine, or switched to the disengaged state to disconnect the mechanical connection between the internal combustion engine and the multi-speed transmission, realizing the combined drive and separation control of the two motors;

[0070] The output end of the generator is electrically connected to the power battery, the first drive motor, and the second drive motor through a power electronic device, and is used to charge the energy storage module or directly supply power to the first drive motor and the second drive motor when the engine is working;

[0071] The control module is electrically connected to the internal combustion engine, the generator, the first drive motor, the second drive motor, and the multi-speed transmission, and receives signals from various vehicle sensors, including the throttle pedal position, the braking signal, the vehicle speed sensor, the engine speed, the generator voltage and current, the motor speed and current, the battery SOC, etc., and is used to comprehensively execute path condition prediction, energy distribution control, health state estimation, and drive strategy scheduling.

[0072] The software built into the control module implements the control method of the present invention, mainly including two major logics: energy management and fault diagnosis. In terms of energy management, the control module adopts a two-level control architecture: the upper layer is the strategy layer, and the lower layer is the execution layer. The strategy layer evaluates the degree of involvement of the range extender under the current working conditions through the policy network trained by reinforcement learning. For example, when long-distance high-speed cruising is expected, the engine output power ratio is increased, and vice versa, pure electric drive is preferred under the urban working conditions of frequent starts and stops. The execution layer specifically determines the start and stop times of the engine, power setting, motor torque distribution, and gear shifting operations through model predictive optimization. The execution layer performs rolling optimization at a frequency of several times per second to ensure meeting the power demand and constraint conditions. For example, when the vehicle has a high acceleration requirement and the battery power is insufficient, the strategy layer instructs the engine to intervene, and the execution layer then starts the engine at an appropriate time and gradually increases the power generation power for the motor to use after the engine warms up. At the same time, a suitable gear is selected to keep the motor speed within the efficient region. The entire energy management process aims to reduce the total fuel consumption and takes into account factors such as battery life (avoiding overcharging and discharging) and engine start-stop losses (avoiding overly frequent starts and stops), achieving a balance between economy and component life. In terms of fault diagnosis, the control module runs a real-time operating state monitoring algorithm. For the engine-generator set, a method combining model diagnosis and signal analysis is adopted: the actual shaft power calculated from the generator current and speed is compared with the theoretical output of the engine. If the difference between the two is within the normal range, it is regarded as healthy; if the difference continuously exceeds the threshold, the fault type is judged in combination with the difference pattern and other sensing data. For example, if it is detected that the engine fuel consumption increases while the shaft power output decreases, it may be an engine misfire or abnormal compression ratio, and the system will mark the engine as faulty; another example is that if the generator excitation current is abnormal, the generator fault is identified. Once the diagnosis module determines a fault, it immediately notifies the energy management module to take countermeasures and records the fault code.

[0073] As Figure 2 shown, in one of the embodiments, the control module specifically includes a path prediction unit, an energy management unit, a health state estimation unit, and a drive strategy scheduling unit.

[0074] The path prediction unit is used to predict the path working condition information (road conditions, slope changes, traffic flow, etc.) within a set future time domain based on the road information of the navigation system, combined with historical driving data, through machine learning algorithms.

[0075] The energy management unit is used to construct an energy optimization control model by combining model predictive control and reinforcement learning according to the real-time vehicle state parameters, the state of charge (SOC) of the battery, and the predicted path driving conditions, and generate optimized control instructions, including the start / stop timing of the internal combustion engine, the power generation power of the generator, the torque distribution of the first drive motor and the second drive motor, the shifting strategy of the multi-speed transmission, etc., so as to achieve the comprehensive optimum of energy utilization efficiency and power response performance; the energy management unit adjusts the working states of the internal combustion engine and the electric drive system according to the optimized power distribution scheme. Especially when climbing or accelerating, the system will appropriately increase the motor power to ensure the vehicle's power demand.

[0076] Furthermore, the rolling optimization objective function J of the energy optimization control model is expressed as:

[0077]

[0078] In the formula, m is the starting step of the current optimization, p is the prediction horizon length of the model predictive control; FC(k) is the fuel consumption of the internal combustion engine at the k-th step; SOC(k) is the current state of charge of the battery, ρ SOC is the SOC error penalty weight; S e (k), S e (k - 1) are the start / stop states of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, ρ Se is the start / stop switching penalty weight of the internal combustion engine; T e (k), T e (k - 1) are the output torques of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, ρ Te is the torque change penalty weight of the internal combustion engine; Sm1(k), Sm1(k - 1) are the working states of the first drive motor at the k-th step and the (k - 1)-th step respectively, ρ Sm1 is the state switching penalty weight of the first drive motor; ΔT m1 is the torque change value of the first drive motor at the k-th step, ρ Tm1 is the torque change penalty weight of the first drive motor; Sm2(k), Sm2(k - 1) are the working states of the second drive motor at the k-th step and the (k - 1)-th step respectively, ρ Sm2 is the state switching penalty weight of the second drive motor; ΔT m2 is the torque change value of the second drive motor at the k-th step, ρ Tm2 is the torque change penalty weight of the second drive motor; G(k - 1), G(k) are the current gear positions at the k-th step and the (k - 1)-th step respectively, ρ G is the shifting frequency penalty weight.

[0079] The energy management unit calls the policy network trained based on reinforcement learning and inputs the following information: including the real-time operating state parameters of the vehicle, including the current vehicle speed, acceleration, and drive power request; the state of charge (SOC) of the battery; the predicted path condition information, including the road speed limit and slope information; the drive motor speed and efficiency mapping information;

[0080] Generate the dynamic correction amounts of various penalty weights according to the above input information

[0081] Δρ = [Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G T , where T represents the transpose operation, and combined with the basic penalty weight vector to obtain various dynamic penalty weights, that is:

[0082]

[0083] The reward function R(k) of the reinforcement learning policy network is defined as:

[0084]

[0085] In the formula, Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G and are the dynamic correction amounts and initial values of the SOC error penalty weight, internal combustion engine start-stop switching penalty weight, internal combustion engine torque change penalty weight, first drive motor state switching penalty weight, first drive motor torque change penalty weight, second drive motor state switching penalty weight, second drive motor torque change penalty weight, and shift frequency penalty weight respectively; W is the correction amplitude control factor; SOC ref is the target state of charge of the battery; s ref is the expected vehicle state on the predicted path; s achieved is the current actual driving state value; ε is the state deviation tolerance threshold; γ1, γ2, γ3, γ4 are the weight coefficients of fuel economy, SOC stability, path tracking, and shift smoothness respectively.

[0086] ​The health status estimation unit is used to collect the shaft power of the generator and the system status signal, combine the optimized control instruction, construct an extended Kalman filter model, take the shaft power of the generator as the observation input in the augmented state space, and estimate the health parameters of the internal combustion engine in real time. By comparing the true value and the estimated value of the health parameters, it determines whether there is a fault in the internal combustion engine or the generator, and outputs a fault flag when the judgment result is abnormal. If it is found that the performance of the system components decreases or a fault occurs, the health status estimation unit will issue an alarm and trigger the corresponding fault handling mechanism (such as adjusting the energy distribution or entering the fault tolerance mode).

[0087] The expression of the extended Kalman filter model is defined as follows:

[0088]

[0089] In the formula, are the estimated values of the health parameters at the i-th sampling moment and the (i - 1)-th sampling moment respectively, including the compressor efficiency η c , the turbine efficiency η t , the combustion degradation factor f c and the friction degradation factor f t ; A aug is the state transition matrix; B aug is the control input matrix, u(t i ) is the control input quantity; F aug is the shaft power observation coupling matrix, P shaft (t i ) is the measured value of the generator shaft power; K aug is the Kalman gain matrix calculated based on the predicted covariance and the observation error covariance; y(t i ) is the actually observed output value; C aug is the observation matrix;

[0090] During the health parameter estimation process, the covariance matrix P aug (t i ) is synchronously calculated to measure the uncertainty of the health parameter estimation value;

[0091] If any of the following conditions is satisfied:

[0092] The estimated value of any health parameter deviates from the preset nominal value by more than the tolerance threshold;

[0093] The main diagonal element of the covariance corresponding to any health parameter is greater than the preset threshold, then mark the current state as a potential fault, output the fault flag f fault = 1, and execute the hierarchical fault tolerance control strategy according to the fault level.

[0094] The driving strategy scheduling unit is used to select the vehicle operation mode according to the health status estimation result, including the range - extender mode, pure - electric mode or fault - tolerant mode, and feedback the scheduling result to the energy management unit for updating the control optimization boundary conditions. The driving strategy scheduling unit not only optimizes the energy distribution, but also adjusts the power distribution of the drive motor according to the real - time demand to avoid low energy efficiency or unnecessary energy waste.

[0095] Furthermore, the driving strategy scheduling unit introduces a delay confirmation mechanism before mode switching, requiring that the abnormal health parameters continuously meet the trigger condition for more than the set duration, and completing the switching confirmation by combining the historical state fluctuation statistics.

[0096] Range - extender mode: If the system confirms that the current health parameters are within the normal range, there is no abnormal fluctuation, and no fault flag signal is generated, the range - extender mode is maintained.

[0097] Fault - tolerant mode: If the estimated value of any health parameter deviates from the tolerance interval, or the estimated covariance exceeds the limit, but the system has not triggered a fault flag, the system enters the fault - tolerant mode.

[0098] In the fault - tolerant mode, the following control constraint conditions are feedback to the energy management unit to update the control optimization boundary conditions, specifically including:

[0099] Limit the maximum allowable output power of the engine not to exceed 70% of the rated power.

[0100] Increase the minimum start - stop interval time of the engine.

[0101] Limit the torque change rate of the drive motor.

[0102] Limit the gear - shifting frequency of the transmission.

[0103] Pure - electric mode: If the fault flag f fault = 1, the system directly switches to the pure - electric mode, disconnects the mechanical connection between the internal combustion engine and the multi - speed transmission, and is powered only by the first and second drive motors, and records the fault event for subsequent maintenance processing.

[0104] Preferably, in the range - extender power generation system, the coordinated control of motor efficiency and the optimization of the shifting strategy are the keys to ensuring the efficient operation of the system. This control method optimizes the motor efficiency in real - time by dynamically adjusting the torque and speed of the motor, and coordinating with the road condition information and the state of charge of the battery, avoiding operation in the low - efficiency interval, and balancing the battery service life and the vehicle dynamic performance. The following is one of the implementation methods:

[0105] Monitoring of Motor Operating Status: The system monitors the operating status of two driving motors in real time, including motor speed, output torque, and motor efficiency. These data are communicated through sensors to the control unit and are used to calculate the motor efficiency. The motor efficiency is measured by the performance of the motor's output torque and speed relative to its maximum operating point.

[0106] The calculation formula for the motor efficiency η is:

[0107] Where P out is the motor output power, and P in is the motor input power. The efficiency represents the energy conversion efficiency of the motor;

[0108] Real-time Adjustment Strategy: When the vehicle starts to accelerate or climb a slope, the system dynamically adjusts the operating mode of the two motors according to the current state of charge (SOC) of the battery, the driver's acceleration request, vehicle speed demand, and road slope information. By coordinating the torque output of the two motors, they are made to operate within the optimal efficiency range. For example, if during a slope climb, the efficiency of the first driving motor starts to decline and is about to enter the low-efficiency range, the system will automatically increase the output torque of the second motor while decreasing the output of the first motor, thus keeping both motors operating within their highest efficiency range.

[0109] Cooperative Operation and Torque Distribution: The motor torque distribution is dynamically adjusted based on motor efficiency and load demand. By calculating the efficiency performance of each motor under the current load, the system will preferentially select the high-efficiency motor for main drive. If a motor's load is too large or its efficiency is below the set threshold, the system will switch the operating mode according to the real-time battery state, possibly by changing the motor load distribution or activating the range extender to supplement the battery energy.

[0110] Shift Timing Prediction: The shift strategy is based on the motor efficiency calculated in real time and the vehicle's dynamic requirements, and combines path condition prediction (such as road slope, traffic flow, etc.) to determine when to shift gears. For example, when the vehicle is about to enter an uphill section and the motor speed reaches the preset threshold, the system will predict in advance that the motor will enter the low-efficiency range, so it will adjust the motor output by shifting gears in advance to avoid the motor being in the low-efficiency state for a long time.

[0111] The optimization of the shift timing is achieved through the following strategies:

[0112] Real-time Torque and Speed Matching: By optimizing the motor torque and speed, the motor is always kept within the high-efficiency operating range. By combining road condition information, such as uphill or downhill sections, the system predicts the future workload and makes a shift decision in advance when the workload increases.

[0113] Calculation of improved shifting efficiency: When the system predicts that the efficiency of the current gear is lower than that of the next gear, by comparing the efficiency difference between the two gears and the load demand, the most suitable shifting timing is selected to maximize the energy utilization rate of the motor.

[0114] Adaptive shifting strategy: The adaptive shifting algorithm dynamically adjusts the shifting timing according to real-time road condition prediction, vehicle speed change, battery SOC, and driver demand to ensure a smooth and efficient shifting process. The system decides whether to shift gears earlier or later based on the efficiency change of the motor under the current load and the predicted load change.

[0115] Battery state of charge (SOC) and motor torque control: The system adjusts the torque output of the motor in real time according to the change of battery SOC to ensure that electric drive is preferentially used when the battery has low power, while reducing the use of the range extender. To extend the service life of the battery, when the battery SOC is lower than the set threshold, the system will automatically increase the participation of the range extender and reduce the dependence of electric drive on the battery. When the battery SOC is low, the system will reduce the power output of the motor to avoid over-discharging the battery. At the same time, by controlling the shifting logic, the motor torque is limited to a low-power range to extend the battery usage time.

[0116] Balancing of shifting strategy and battery performance: When the battery has low power, the shifting strategy will try to avoid frequent shifting operations to reduce the charge and discharge frequency of the battery. The system will preferentially select the working mode with the least energy consumption and balance the power demand of the vehicle to avoid excessive energy loss. For example: when the battery power is close to the set lower limit, the system will reduce the shifting frequency and extend the use time of the current gear until the system confirms that the battery power has recovered or reached a sufficient level, and then shift gears again.

[0117] Dynamic adjustment of motor efficiency and battery state: The system optimizes the torque distribution of the motor in advance according to the predicted path condition information (such as uphill or rapid acceleration, etc.), and dynamically selects the optimal shifting strategy according to the current battery SOC, which not only ensures the battery usage efficiency but also ensures that the motor output power meets the driving demand.

[0118] Through the dynamic coordinated control of the efficiency of two motors, intelligent shifting, and precise battery management, this embodiment effectively improves the energy efficiency of the range extender power generation system, reduces battery loss, and optimizes the overall system performance. The coordinated control of motor efficiency ensures that the motor operates within the optimal working range, while the adaptive shifting algorithm adjusts the shifting timing according to real-time data and predicted information to ensure the smoothness and efficiency of power output. Combined with the battery management strategy, this embodiment not only provides high efficiency but also ensures the long-term stability of the battery and the overall reliability of the system.

[0119] Such as Figure 3As shown, this is another embodiment of the present invention, which provides a control method for an extender power generation system, including:

[0120] Collect real-time vehicle state parameters, including the current vehicle speed, power request, state of charge (SOC) of the battery, and predicted path condition information;

[0121] According to the real-time information and a pre-set control strategy, coordinate the control of the internal combustion engine, generator, and first and second drive motors based on model predictive control and reinforcement learning to optimize energy distribution;

[0122] When shifting gears, determine whether the current gear is about to deviate from the high-efficiency working range. If it is determined that the current drive motor is about to enter the low-efficiency working range, then based on path condition prediction and vehicle acceleration request, switch to the gear with the optimal efficiency in advance, and compensate the motor output torque through an adaptive shifting algorithm to maintain the smoothness of power output; during the control process, dynamically adjust the shifting timing according to real-time feedback and predicted data to ensure that the system operates in the optimal energy efficiency range;

[0123] When it is detected that the internal combustion engine or generator fails or its performance deteriorates, the control system enters a fault tolerance mode, including cutting off or reducing the fuel supply of the internal combustion engine, and only the battery and the first and second drive motors are used to maintain vehicle driving, and the output power or speed is restricted in a timely manner to ensure the vehicle can safely drive to a predetermined location.

[0124] In summary, through innovative technical means such as energy management, shifting optimization, fault diagnosis, and fault tolerance control, the present invention significantly improves the overall performance and reliability of the extender power generation system. By combining model predictive control (MPC), reinforcement learning, and multi-motor coordinated control, the present invention can achieve efficient energy distribution and shifting strategies under various working conditions, optimize the utilization of battery power, and reduce energy consumption. At the same time, the introduction of the fault diagnosis and fault tolerance control mechanism enhances the robustness and safety of the system, ensuring the controllability and safety of the vehicle under various abnormal working conditions.

[0125] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of various changes or substitutions, and these should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. An extended-range generator power generation system, characterized in that, The system includes: An energy storage module, including a power battery and a battery management system, for controlling the input and output of electric energy and providing the state of charge (SOC) of the battery; A drive module, including a first drive motor, a second drive motor, and a multi-speed transmission. The first drive motor and the second drive motor are respectively connected to the odd-numbered gear shaft and the even-numbered gear shaft of the multi-speed transmission, for converting electric power into drive torque, which is transmitted to the vehicle drive wheels after being speeded up and torque-increased by the multi-speed transmission to generate a driving force output. An extender module, including an internal combustion engine and a generator mechanically connected to the internal combustion engine; The internal combustion engine is mechanically coupled to the even-numbered gear shaft of the multi-speed transmission through a clutch. The clutch is switched to the engaged state under the control of the control module to transmit the output torque of the internal combustion engine, or switched to the disengaged state to disconnect the mechanical connection between the internal combustion engine and the multi-speed transmission; The output end of the generator is electrically connected to the power battery, the first drive motor, and the second drive motor through a power electronic device, for charging the energy storage module or directly supplying power to the first drive motor and the second drive motor when the engine is working. A control module, electrically connected to the internal combustion engine, the generator, the first drive motor, the second drive motor, and the multi-speed transmission, for comprehensively performing path condition prediction, energy distribution control, health state estimation, and drive strategy scheduling.

2. The range extender power generation system according to claim 1, wherein The control module specifically includes: A path prediction unit, for predicting the path condition information within a future set time domain through a machine learning algorithm based on the road information of the navigation system and combined with historical driving data. An energy management unit, for constructing an energy optimization control model by combining model predictive control and reinforcement learning according to the vehicle's real-time state parameters, the state of charge (SOC) of the battery, and the predicted path condition information, and generating an optimization control instruction. The optimization control instruction includes the start-stop timing of the internal combustion engine, the power generation power of the generator, the torque distribution of the first drive motor and the second drive motor, and the shifting strategy of the multi-speed transmission, so as to achieve the comprehensive optimum of energy utilization efficiency and dynamic response performance. A health state estimation unit, for collecting the shaft power of the generator and the system state signal, combining the optimization control instruction, constructing an extended Kalman filter model, taking the shaft power of the generator as the observation input in the augmented state space, and performing real-time estimation of the health parameters of the internal combustion engine. By comparing the true value and the estimated value of the health parameters, it is judged whether there is a fault in the internal combustion engine or the generator, and a fault flag is output when the judgment result is abnormal. A drive strategy scheduling unit, for selecting the vehicle operation mode according to the health state estimation result, including the extended range mode, the pure electric mode, or the fault tolerance mode, and feeding back the scheduling result to the energy management unit for updating the control optimization boundary conditions.

3. The range extender power generation system according to claim 2, characterized in that, The rolling optimization objective function J of the energy optimization control model is expressed as: where m is the starting step of the current optimization, p is the prediction horizon length of the model predictive control; FC(k) is the fuel consumption of the internal combustion engine at the k-th step; SOC(k) is the current state of charge of the battery, and ρ SOC is the SOC error penalty weight; S e (k) and S e (k - 1) are the start-stop states of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, and ρ Se is the start-stop switching penalty weight of the internal combustion engine; T e (k) and T e (k - 1) are the output torques of the internal combustion engine at the k-th step and the (k - 1)-th step respectively, and ρ Te is the torque change penalty weight of the internal combustion engine; Sm1(k) and Sm1(k - 1) are the working states of the first drive motor at the k-th step and the (k - 1)-th step respectively, and ρ Sm1 is the state switching penalty weight of the first drive motor; ΔT m1 is the torque change value of the first drive motor at the k-th step, and ρ Tm1 is the torque change penalty weight of the first drive motor; Sm2(k) and Sm2(k - 1) are the working states of the second drive motor at the k-th step and the (k - 1)-th step respectively, and ρ Sm2 is the state switching penalty weight of the second drive motor; ΔT m2 is the torque change value of the second drive motor at the k-th step, and ρ Tm2 is the torque change penalty weight of the second drive motor; G(k - 1) and G(k) are the current gears at the k-th step and the (k - 1)-th step respectively, and ρ G is the shift frequency penalty weight.

4. An extended-range power generation system according to claim 3, characterized in that The energy management unit calls a policy network trained based on reinforcement learning and inputs the following information: including the vehicle's real-time operation state parameters, including the current vehicle speed, acceleration, and drive power request. Battery state of charge (SOC); predicted route condition information, including road speed limit and slope information; drive motor speed and efficiency mapping information; Generate dynamic corrections for each penalty weight based on the above input information Δρ = [Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G T , where T represents the transpose operation, and combined with the basic penalty weight vector to obtain the dynamic penalty weights for each item, that is:​ The reward function R(k) of the reinforcement learning policy network is defined as: where Δρ SOC , Δρ Se , Δρ Te , Δρ Sm1 , Δρ Tm1 , Δρ Sm2 , Δρ Tm2 , Δρ G and are the dynamic correction amounts and initial values of the SOC error penalty weight, the start-stop switching penalty weight of the internal combustion engine, the torque change penalty weight of the internal combustion engine, the state switching penalty weight of the first drive motor, the torque change penalty weight of the first drive motor, the state switching penalty weight of the second drive motor, the torque change penalty weight of the second drive motor, and the shift frequency penalty weight respectively; W is the correction amplitude control factor; SOC ref is the target state of charge of the battery; s ref is the desired vehicle state on the predicted path; s achieved is the current actual driving state value; ε is the state deviation tolerance threshold; γ1, γ2, γ3, and γ4 are the weight coefficients of fuel economy, SOC stability, path tracking, and shift smoothness respectively.

5. An extended-range generator power generation system according to claim 2, wherein, The expression of the extended Kalman filter model is defined as follows: Wherein, are the estimated values of the health parameters at the i-th sampling moment and the (i-1)-th sampling moment, respectively, including the compressor efficiency η c , the turbine efficiency η t , the combustion degradation factor f c and the friction degradation factor f t ; A aug is the state transition matrix; B aug is the control input matrix, and u(t i ) is the control input quantity; F aug is the shaft power observation coupling matrix, and P shaft (t i ) is the measured value of the generator shaft power; K aug is the Kalman gain matrix calculated based on the predicted covariance and the observation error covariance; y(t i ) is the actually observed output value; C aug is the observation matrix; Synchronously calculate the estimation covariance matrix P aug (t i ) during the health parameter estimation process to measure the uncertainty of the health parameter estimation value; if any of the following conditions is satisfied: Any health parameter estimate deviates from the preset nominal value by more than the tolerance threshold; If the covariance main diagonal element corresponding to any health parameter is greater than the preset threshold, then mark the current state as a potential fault, output the fault flag f fault = 1, and execute the hierarchical fault-tolerant control strategy according to the fault level.

6. The range extender power generation system according to claim 2, wherein The driving strategy scheduling unit introduces a delayed confirmation mechanism before switching the mode, requiring that the abnormal health parameters continue to meet the trigger conditions for more than a set time, and completes the switching confirmation in combination with historical state fluctuation statistics; Extended range mode: If the system confirms that the current health parameters are within the normal range, there is no abnormal fluctuation, and no fault flag signal is generated, the extended range mode is maintained; Fault-tolerant mode: If any health parameter estimate deviates from the tolerance interval, or the estimated covariance exceeds the limit, but the system has not yet triggered the fault flag, the system enters fault-tolerant mode; In the fault-tolerant mode, the following control constraints are fed back to the energy management unit to update the control optimization boundary conditions, including: Limit the maximum permissible engine output power to no more than 70% of the rated power; Increase the minimum engine start-stop interval; Limit the rate of change of the drive motor torque; Limit the frequency of transmission shifting; Pure electric mode: If the fault flag f fault = 1, the system directly switches to the pure electric mode, disconnects the mechanical connection between the internal combustion engine and the multi-speed transmission, and operates only powered by the first and second drive motors, and records the fault event for subsequent maintenance processing.

7. An extended-range generator power generation system according to claim 1, characterized in that The multi-speed gearbox is an automatic transmission mechanism with at least three gears or more, and the transmission ratio of each gear is pre-set according to the highest efficiency working range of the first drive motor and the second drive motor; When the gear corresponding to the current drive motor is about to enter the gear shift state, the control module determines the power demand trend based on the path condition prediction, vehicle acceleration request and motor speed, and the other drive motor increases or maintains its output torque in advance for drive compensation, and performs gear selection control according to the gear shift map or the adaptive gear shift algorithm based on the motor efficiency area.

8. An extended-range generator power generation system according to claim 7, characterized in that, The adaptive shifting algorithm takes the current vehicle speed, motor speed, motor output torque and path condition prediction data as input, combines the efficiency characteristics of the first and second drive motors at different operating points, and constructs a mapping relationship between motor efficiency and load to predict the motor efficiency performance under each gear in real time, and is used to dynamically determine the target gear; When the current drive motor is about to leave its high efficiency range and the predicted path has a continuous climbing or acceleration trend, the control module calculates the comprehensive efficiency improvement of the next gear or adjacent gear within the predicted period, and considers the coordinated output efficiency of the other drive motor in this gear. If there is an efficiency overlap area or the efficiency improvement exceeds the set threshold and meets the minimum holding time requirement, early gear shifting is triggered, and the other drive motor takes on torque compensation to ensure power output continuity and smoothness; The gear shifting strategy introduces a sliding mode judgment mechanism to keep the current gear unchanged when the efficiency improvement is not enough to exceed the set dead zone or the path prediction trend is unstable.

9. A control method for a range extender power generation system according to any one of claims 1-8, characterized in that The method comprises: Collect real-time vehicle status parameters, including current vehicle speed, power request, battery state of charge (SOC), and predicted route condition information; Based on real-time information and preset control strategies, coordinated control of the internal combustion engine, generator, and first and second drive motors is performed based on model predictive control and reinforcement learning to optimize energy distribution; When shifting gears, it is judged whether the current gear is about to deviate from the high-efficiency working range. If it is judged that the current drive motor is about to enter the low-efficiency working range, then based on path condition prediction and vehicle acceleration request, it is switched to the gear with the optimal efficiency in advance, and the motor output torque is compensated through an adaptive shifting algorithm to maintain the smoothness of power output; during the control process, according to real-time feedback and prediction data, the shifting timing is dynamically adjusted to ensure that the system operates in the optimal energy efficiency range; When it is detected that the internal combustion engine or generator fails or its performance deteriorates, the control system enters a fault-tolerant mode, including cutting off or reducing the fuel supply of the internal combustion engine, and only the battery and the first and second drive motors are used to maintain vehicle driving, and the output power or speed is limited in a timely manner to ensure that the vehicle can safely drive to a predetermined location.

Images