CN115071673A - PHEV real-time energy management strategy based on model-free adaptive control - Google Patents

Abstract

The invention discloses a plug-in hybrid electric vehicle real-time energy management strategy based on model-free adaptive control, wherein the energy management strategy is divided into an upper layer of the energy management strategy and a lower layer of the energy management strategy by adopting a layered optimization control framework; the strategy upper layer is based on a feedback control principle, and can output a control variable lambda in real time for the control of the strategy lower layer through the feedback of the target state and the actual state of the vehicle; the key control parameter of the strategy lower layer is a covariate lambda. The online optimization of lambda under unknown working conditions can be realized through the online control adjustment of the model-free adaptive controller, so that the adaptive capacity of the lower-layer PMP strategy to the unknown working conditions can be ensured, the algorithm performance is improved, and the vehicle economy is ensured.

Description

Translated fromChinese

基于无模型自适应控制的PHEV实时能量管理策略Real-time energy management strategy for PHEV based on model-free adaptive control

技术领域technical field

本发明涉及新能源汽车控制技术，确切地说是一种基于无模型自适应控制的PHEV实时能量管理策略。The invention relates to a new energy vehicle control technology, specifically a PHEV real-time energy management strategy based on model-free adaptive control.

背景技术Background technique

插电式混合动力汽车(PHEV)是缓减全球能源与环境危机，实现汽车产业跨越式发展的重要产品，其多能量源耦合驱动的构型特点为车辆节能减排潜力的深度挖掘提供了基础。如何实现PHEV多动力源之间能量的实时最优分配，是混合动力汽车能量管理的技术难点。Plug-in hybrid electric vehicle (PHEV) is an important product to alleviate the global energy and environmental crisis and realize the leap-forward development of the automobile industry. . How to realize the real-time optimal distribution of energy among the multiple power sources of PHEV is the technical difficulty of energy management of HEV.

目前，已实现工程应用的混合动力系统能量管理策略多以规则型控制为主，其逻辑结构简单，实时性好，但对不同工况的适应性较差，容易出现能耗较高的情况。因此，科学研究主要以基于优化的控制策略为主。尽管该类方法对车辆经济性的提升具有显著作用，但是，基于全局优化的算法仅适用于全局工况已知的情况，无法适应行驶工况的随机变化，在线应用困难；基于实时优化的算法多需通过复杂的计算和控制结构来实现，对实时控制系统运算能力的要求较高，尚不具备在线应用的条件。At present, the energy management strategies of hybrid power systems that have been implemented in engineering are mostly rule-based control, which has a simple logic structure and good real-time performance, but has poor adaptability to different working conditions and is prone to high energy consumption. Therefore, scientific research is mainly based on optimization-based control strategies. Although this type of method has a significant effect on the improvement of vehicle economy, the algorithm based on global optimization is only suitable for the situation where the global operating conditions are known, and cannot adapt to the random changes of driving conditions, so it is difficult to apply online; the algorithm based on real-time optimization is difficult to apply. Most of them need to be realized through complex calculation and control structure, and have higher requirements on the computing power of real-time control system, and do not have the conditions for online application.

专利CN107284441A公开了一种实时工况自适应的插电式混合动力汽车能量优化管理方法，其存在以下缺点：1)该策略的实施需要通过云端服务器实时获取道路工况信息，并通过能耗模型与车速规划模型对混合动力汽车的经济车速进行在线规划，对应用场景要求较高；2)车载控制器端，需对目标经济车速、预测经济车速以及实际车速三者进行在线分析，从而获取车辆行驶的最佳需求转矩序列，然后通过等效燃油最小消耗与动态规划对车辆进行实时控制，其控制算法较为复杂，对控制器的在线算力要求较高。Patent CN107284441A discloses a real-time working condition self-adaptive plug-in hybrid electric vehicle energy optimization management method, which has the following disadvantages: 1) The implementation of this strategy requires real-time acquisition of road working condition information through a cloud server, and energy consumption model On-line planning of the economic speed of hybrid vehicles with the speed planning model requires high application scenarios; 2) On the vehicle controller side, online analysis of the target economic speed, predicted economic speed and actual speed is required to obtain the vehicle The optimal demand torque sequence for driving, and then the vehicle is controlled in real time through the equivalent minimum fuel consumption and dynamic programming. The control algorithm is relatively complex, and the online computing power of the controller is required to be high.

发明内容SUMMARY OF THE INVENTION

本发明要解决的技术问题是提供一种基于无模型自适应控制的PHEV实时能量管理策略，该能量管理策略能够实现优化算法在混合动力系统能量管理过程中的实时应用，同时保证能量管理策略对随机工况的良好适应能力，达到节能减排的目标。The technical problem to be solved by the present invention is to provide a PHEV real-time energy management strategy based on model-free adaptive control, which can realize the real-time application of the optimization algorithm in the energy management process of the hybrid power system, and at the same time ensure the correct Good adaptability to random working conditions to achieve the goal of energy saving and emission reduction.

为解决上述技术问题，本发明采用如下技术手段：In order to solve the above-mentioned technical problems, the present invention adopts the following technical means:

基于无模型自适应控制的PHEV实时能量管理策略，包括以下步骤：所述能量管理策略采用分层优化控制架构分为能量管理策略的上层、能量管理策略的下层；所述能量管理策略的上层为基于MFAC算法的控制层，基于反馈控制设计，其控制输入为动力电池参考SOC与电池实际SOC，其控制输出为下层PMP算法的协态变量λ，能量管理策略的上层根据行驶工况变化自适应输出控制参数；所述能量管理策略的下层为基于PMP算法的能量管理优化层，对混合动力系统多动力源间的能量输出进行实时优化；其中，PMP算法以发动机能耗最小为目标函数，分别以动力电池输出功率P_batt与动力电池SOC为控制变量与状态变量，其关键控制参数为协态变量λ；能量管理策略的下层根据能量管理策略上层输出的控制参数完成车辆动力系统不同动力源之间能量的优化分配。The PHEV real-time energy management strategy based on model-free adaptive control includes the following steps: the energy management strategy is divided into an upper layer of the energy management strategy and a lower layer of the energy management strategy by adopting a hierarchical optimization control architecture; the upper layer of the energy management strategy is The control layer based on the MFAC algorithm, based on the feedback control design, the control input is the reference SOC of the power battery and the actual SOC of the battery, and the control output is the co-state variable λ of the lower PMP algorithm, and the upper layer of the energy management strategy is adaptive according to the driving conditions. Output control parameters; the lower layer of the energy management strategy is an energy management optimization layer based on the PMP algorithm, which optimizes the energy output between multiple power sources of the hybrid power system in real time; wherein, the PMP algorithm takes the minimum engine energy consumption as the objective function, respectively The power battery output power P_batt and the power battery SOC are used as control variables and state variables, and the key control parameter is the co-state variable λ; optimal distribution of energy between.

其中，无模型自适应控制算法，即MFAC算法；庞特里亚金极小值原理算法，即PMP算法。Among them, the model-free adaptive control algorithm is the MFAC algorithm; the Pontryagin minimum principle algorithm is the PMP algorithm.

本发明提出的策略采用简单的反馈控制架构，并利用分层控制的思想对模型进行了简化，其中，上层根据下层反馈的电池SOC以及目标车速对其输出参数进行实时调整，下层能量管理层根据上层输入的控制量实现车辆能耗的在线优化。该策略逻辑结构简单，实时性好，对控制器硬件要求较低，且对未来行驶工况没有任何需求，对应用场景要求较低，仅需知道目标行驶里程与电池SOC即可，具有较高的普适性。The strategy proposed by the present invention adopts a simple feedback control structure and simplifies the model by using the idea of layered control. The control quantity input by the upper layer realizes the online optimization of vehicle energy consumption. The strategy has a simple logical structure, good real-time performance, low requirements for controller hardware, and no requirements for future driving conditions, and low requirements for application scenarios. It only needs to know the target driving range and battery SOC, which has high universality.

进一步的优选技术方案如下：Further preferred technical solutions are as follows:

所述的基于PMP算法的能量管理优化层，对混合动力系统多动力源间的能量输出进行实时优化包括以下步骤：The energy management optimization layer based on the PMP algorithm, the real-time optimization of the energy output between the multiple power sources of the hybrid power system includes the following steps:

(1)考虑PHEV可通过电网获取充足电能，在一定行驶条件下，为提高车辆经济性，电池能量基本耗尽，因此，PMP能量管理策略的优化目标以发动机能耗为主，优化目标表示为：(1) Considering that PHEVs can obtain sufficient electric energy through the power grid, under certain driving conditions, in order to improve the vehicle economy, the battery energy is basically exhausted. Therefore, the optimization objective of the PMP energy management strategy is mainly based on engine energy consumption, and the optimization objective is expressed as :

式中，t表示时间，x(t)和u(t)分别表示状态变量与控制变量，

为发动机瞬时能耗。where t represents time, x(t) and u(t) represent state variables and control variables, respectively,

is the instantaneous energy consumption of the engine.

(2)分别采用电池SOC与动力电池输出功率P_batt作为状态变量与控制变量，控制系统的哈密尔顿函数可表示为：(2) Using battery SOC and power battery output power P_batt as state variables and control variables respectively, the Hamiltonian function of the control system can be expressed as:

式中，λ为协态变量；In the formula, λ is a covariate variable;

(3)由上式可知，PMP能量管理策略的性能取决于关键参数λ的选取，即通过式(2)可将下层的PMP能量管理优化问题转化为协态变量λ的优化问题，其中，λ与电池SOC的动态方程可表示为：(3) It can be seen from the above formula that the performance of the PMP energy management strategy depends on the selection of the key parameter λ, that is, the lower-layer PMP energy management optimization problem can be transformed into the optimization problem of the co-state variable λ by formula (2), where λ The dynamic equation with battery SOC can be expressed as:

(4)在PMP算法求解过程中，考虑动力电池的充放电性能，分别将初始与终止时刻的电池SOC设置为SOC(t₀)＝0.8，SOC(t_f)＝0.3；此外，考虑车辆动力系统物理特性的限制，算法求解过程还需满足以下物理限制：(4) In the process of solving the PMP algorithm, considering the charging and discharging performance of the power battery, the battery SOC at the initial and termination time is set to SOC(t₀ )=0.8, SOC(t_f )=0.3; in addition, considering the vehicle power Due to the limitations of the physical characteristics of the system, the algorithm solution process also needs to meet the following physical limitations:

式中，n_e(t)与n_m(t)表示发动机与电机的转速，n_{e_min}(t),n_{e_max}(t),n_{m_min}(t),n_{m_max}(t)分别为最低与最高转速限制；P_e(t)与P_m(t)为发动机与电机功率，P_{e_min}(t),P_{e_max}(t),P_{m_min}(t),P_{m_max}(t)分别表示最小与最大功率限制。In the formula, n_e (t) and n_m (t) represent the speed of the engine and motor, n_{e_min} (t), n_{e_max} (t), n_{m_min} (t), n_{m_max} (t) are the minimum and maximum speeds, respectively Limits; P_e (t) and P_m (t) are the engine and motor power, and P_{e_min} (t), P_{e_max} (t), P_{m_min} (t), and P_{m_max} (t) represent the minimum and maximum power limits, respectively.

所述的基于MFAC算法的控制策略上层根据行驶工况变化自适应输出控制参数，包括以下步骤：采用伪偏导数PPD估计算法估计每一时刻系统的伪偏导数

采用MFAC算法根据电池参考SOC与实际SOC的差，上一时刻系统的输出λ(k-1)以及通过估计算法得到伪偏导数的估计值

计算得到下一时刻系统的输出。The upper layer of the control strategy based on the MFAC algorithm adaptively outputs the control parameters according to the driving conditions, including the following steps: using the pseudo partial derivative PPD estimation algorithm to estimate the pseudo partial derivative of the system at each moment

The MFAC algorithm is used to obtain the estimated value of the pseudo partial derivative according to the difference between the battery reference SOC and the actual SOC, the output λ(k-1) of the system at the previous moment and the estimation algorithm.

Calculate the output of the system at the next moment.

所述的采用MFAC算法建立系统控制模型为：The described adoption of MFAC algorithm to establish a system control model is:

式中，SOC(k)与λ(k)为系统在k时刻的状态变量与输出变量，分别表示电池当前SOC与协态变量λ，

为系统的伪偏导数；其控制规律可表示为：In the formula, SOC(k) and λ(k) are the state variables and output variables of the system at time k, which represent the current SOC of the battery and the co-state variable λ, respectively,

is the pseudo-partial derivative of the system; its control law can be expressed as:

式中，ρ_k为步长因子，λ_k为权重因子，二者均为小于1的正数；SOC^*(k+1)表示k+1时刻的电池参考SOC，

为

的估计值，即伪偏导数的估计值，可通过下面的式子得到：In the formula, ρ_k is the step size factor, and λ_k is the weight factor, both of which are positive numbers less than 1; SOC^* (k+1) represents the battery reference SOC at time k+1,

for

The estimated value of , that is, the estimated value of the pseudo partial derivative, can be obtained by the following formula:

式中，η_k与μ_k分别为步长因子与权重因子，二者均为小于1的正数，ε表示一个极小的正数，

为

的初值；在PHEV行驶过程中，MFAC控制器根据预先设定的动力电池参考SOC_ref与车辆实际反馈的电池SOC可实时调整并输出协态变量λ用于下层能量管理策略的在线优化；为了提高本发明的适用性，动力电池SOC参考轨迹的设置可根据车辆目标行驶里程进行规划，电池参考SOC可表示为：In the formula, η_k and μ_k are the step size factor and the weight factor respectively, both of which are positive numbers less than 1, ε represents a very small positive number,

for

During the driving process of the PHEV, the MFAC controller can adjust in real time according to the preset power battery reference SOC_ref and the actual feedback battery SOC of the vehicle and output the co-state variable λ for the online optimization of the underlying energy management strategy; in order to To improve the applicability of the present invention, the setting of the power battery SOC reference trajectory can be planned according to the vehicle target mileage, and the battery reference SOC can be expressed as:

式中，SOC_ref,SOC₀ and SOC_f分别为电池参考SOC，SOC初值与SOC终值；D_dist与D_total表示当前行驶里程与目标行驶里程。In the formula, SOC_ref , SOC₀ and SOC_f are the battery reference SOC, the initial SOC value and the final SOC value, respectively; D_dist and D_total represent the current mileage and target mileage.

附图说明Description of drawings

图1为基于MFAC的PHEV实时能量管理策略架构图。Figure 1 is an architecture diagram of a real-time energy management strategy for PHEV based on MFAC.

图2为基于MFAC的实时能量管理策略工作原理图。Fig. 2 is the working principle diagram of the real-time energy management strategy based on MFAC.

图3为基于MFAC的实时能量管理策略应用效果图。Figure 3 is an application effect diagram of a real-time energy management strategy based on MFAC.

具体实施方式Detailed ways

下面结合附图对本发明做进一步详细描述。The present invention will be further described in detail below with reference to the accompanying drawings.

结合图1至图2，本发明所述的一种基于无模型自适应控制的PHEV实时能量管理策略，所述能量管理策略采用分层优化控制架构分为能量管理策略的上层、能量管理策略的下层；1 to 2 , a PHEV real-time energy management strategy based on model-free adaptive control according to the present invention, the energy management strategy adopts a layered optimization control architecture and is divided into the upper layer of the energy management strategy and the upper layer of the energy management strategy. lower layer;

所述的能量管理策略的上层为基于MFAC算法的控制层，基于反馈控制设计，其控制输入为动力电池参考SOC与电池实际SOC，其控制输出为下层PMP算法的协态变量λ，能量管理策略的上层根据行驶工况变化自适应输出控制参数；The upper layer of the energy management strategy is the control layer based on the MFAC algorithm, based on the feedback control design, the control input is the reference SOC of the power battery and the actual SOC of the battery, and the control output is the co-state variable λ of the lower PMP algorithm, the energy management strategy. The upper layer of the adaptive output control parameters according to the change of driving conditions;

所述能量管理策略的下层为基于PMP算法的能量管理优化层，对混合动力系统多动力源间的能量输出进行实时优化；其中，PMP算法以发动机能耗最小为目标函数，分别以动力电池输出功率P_batt与动力电池SOC为控制变量与状态变量，其关键控制参数为协态变量λ；能量管理策略的下层根据能量管理策略的上层输出的控制参数完成车辆动力系统不同动力源之间能量的优化分配。The lower layer of the energy management strategy is an energy management optimization layer based on the PMP algorithm, which optimizes the energy output between multiple power sources of the hybrid power system in real time; wherein, the PMP algorithm takes the minimum engine energy consumption as the objective function, and the power battery output Power P_batt and power battery SOC are control variables and state variables, and the key control parameter is the co-state variable λ; the lower layer of the energy management strategy completes the energy transfer between different power sources of the vehicle power system according to the control parameters output by the upper layer of the energy management strategy. Optimize allocation.

具体实施例如下：Specific examples are as follows:

(1)能量管理策略的下层，过程如下：(1) The lower layer of the energy management strategy, the process is as follows:

考虑PHEV可通过电网获取充足电能，在一定行驶条件下，为提高车辆经济性电池能量基本耗尽，因此，PMP能量管理策略的优化目标以发动机能耗为主，优化目标表示为：Considering that PHEVs can obtain sufficient power through the power grid, under certain driving conditions, the battery energy is basically exhausted to improve vehicle economy. Therefore, the optimization objective of the PMP energy management strategy is mainly based on engine energy consumption, and the optimization objective is expressed as:

式中，t表示时间，x(t)和u(t)分别表示状态变量与控制变量，

为发动机瞬时能耗。本发明中，分别采用电池SOC与动力电池输出功率P_batt作为状态变量与控制变量，控制系统的哈密尔顿函数可表示为：where t represents time, x(t) and u(t) represent state variables and control variables, respectively,

is the instantaneous energy consumption of the engine. In the present invention, the battery SOC and the power battery output power P_batt are respectively used as state variables and control variables, and the Hamiltonian function of the control system can be expressed as:

式中，λ为协态变量。由上式可知，PMP能量管理策略的性能取决于关键参数λ的选取，即通过式(2)可将下层的PMP能量管理优化问题转化为协态变量λ的优化问题，其中，λ与电池SOC的动态方程可表示为：where λ is a covariate variable. It can be seen from the above formula that the performance of the PMP energy management strategy depends on the selection of the key parameter λ, that is, the lower-layer PMP energy management optimization problem can be transformed into the optimization problem of the co-state variable λ by formula (2), where λ is related to the battery SOC. The dynamic equation of can be expressed as:

在PMP算法求解过程中，考虑动力电池的充放电性能，分别将初始与终止时刻的电池SOC设置为SOC(t₀)＝0.8，SOC(t_f)＝0.3。此外，考虑车辆动力系统物理特性的限制，算法求解过程还需满足以下物理限制：In the process of solving the PMP algorithm, considering the charging and discharging performance of the power battery, the battery SOC at the initial and termination time is set to SOC(t₀ )=0.8 and SOC(t_f )=0.3 respectively. In addition, considering the limitations of the physical characteristics of the vehicle power system, the algorithm solution process also needs to meet the following physical limitations:

式中，n_e(t)与n_m(t)表示发动机与电机的转速，n_{e_min}(t),n_{e_max}(t),n_{m_min}(t),n_{m_max}(t)分别为最低与最高转速限制；P_e(t)与P_m(t)为发动机与电机功率，P_{e_min}(t),P_{e_max}(t),P_{m_min}(t),P_{m_max}(t)分别表示最小与最大功率限制。In the formula, n_e (t) and n_m (t) represent the speed of the engine and motor, n_{e_min} (t), n_{e_max} (t), n_{m_min} (t), n_{m_max} (t) are the minimum and maximum speeds, respectively Limits; P_e (t) and P_m (t) are the engine and motor power, and P_{e_min} (t), P_{e_max} (t), P_{m_min} (t), and P_{m_max} (t) represent the minimum and maximum power limits, respectively.

(2)能量管理策略的上层，其过程如下：(2) The upper layer of the energy management strategy, the process is as follows:

本实施例通过下层基于PMP的能量管理策略的实施，将混合动力系统能量优化分配问题转化为协态变量λ的优化问题，为提高其对不同工况的适应能力，基于MFAC控制算法建立系统控制模型为：In this embodiment, through the implementation of the PMP-based energy management strategy at the lower level, the energy optimization distribution problem of the hybrid power system is transformed into the optimization problem of the co-state variable λ. In order to improve its adaptability to different working conditions, a system control system is established based on the MFAC control algorithm The model is:

式中，SOC(k)与λ(k)为系统在k时刻的状态变量与输出变量，分别表示电池当前SOC与协态变量λ，

为系统的伪偏导数。其控制规律可表示为：In the formula, SOC(k) and λ(k) are the state variables and output variables of the system at time k, which represent the current SOC of the battery and the co-state variable λ, respectively,

is the pseudo-partial derivative of the system. Its control law can be expressed as:

式中，ρ_k为步长因子，λ_k为权重因子，二者均为小于1的正数；SOC^*(k+1)表示k+1时刻的电池参考SOC，

为

的估计值，可通过下面的式子得到：In the formula, ρ_k is the step size factor, and λ_k is the weight factor, both of which are positive numbers less than 1; SOC^* (k+1) represents the battery reference SOC attime k+1,

for

The estimated value of , can be obtained by the following formula:

式中，η_k与μ_k分别为步长因子与权重因子，二者均为小于1的正数，ε表示一个极小的正数，

为

的初值。In the formula, η_k and μ_k are the step size factor and the weight factor respectively, both of which are positive numbers less than 1, ε represents a very small positive number,

for

the initial value of .

通过上述方案的实施，在PHEV行驶过程中，MFAC控制算法根据预先设定的动力电池参考SOC与车辆实际反馈的电池SOC可实时调整并输出协态变量λ用于下层能量管理策略的在线优化。Through the implementation of the above solutions, during the driving process of the PHEV, the MFAC control algorithm can adjust in real time according to the preset reference SOC of the power battery and the battery SOC actually fed back by the vehicle and output the co-state variable λ for the online optimization of the underlying energy management strategy.

为了提高本发明的适用性，动力电池SOC参考轨迹的设置可根据车辆目标行驶里程进行规划，电池参考SOC可表示为：In order to improve the applicability of the present invention, the setting of the power battery SOC reference trajectory can be planned according to the vehicle target mileage, and the battery reference SOC can be expressed as:

式中，SOC_ref,SOC₀与SOC_f分别为电池参考SOC，SOC初值与SOC终值；D_dist与D_total表示当前行驶里程与目标行驶里程。In the formula, SOC_ref , SOC₀ and SOC_f are the battery reference SOC, the initial SOC value and the final SOC value, respectively; D_dist and D_total represent the current mileage and target mileage.

本实施例的工作原理如下：The working principle of this embodiment is as follows:

参考图2，当已知PHEV目标行驶里程D_total以及动力电池SOC的初值SOC₀与终值SOC_f时，可得到动力电池在未来行驶过程中的放电参考轨迹(即SOC_ref)。车辆开始行驶时，其功率需求与电池实际SOC均可通过CAN信号获取，此时，上层的MFAC控制算法根据设定的电池参考SOC与车辆反馈的真实SOC开始计算得到其输出变量λ，并将该值传递至下层的PMP能量管理策略中，下层能量管理策略则根据λ的变化实时调整每一时刻发动机与动力电池的功率分配以满足每一时刻车辆行驶的总功率需求。Referring to FIG. 2 , when the target driving range D_total of the PHEV and the initial value SOC₀ and final value SOC_f of the power battery SOC are known, the discharge reference trajectory (ie SOC_ref ) of the power battery in the future driving process can be obtained. When the vehicle starts to drive, its power demand and the actual SOC of the battery can be obtained through the CAN signal. At this time, the upper-layer MFAC control algorithm starts to calculate the output variable λ according to the set reference SOC of the battery and the real SOC fed back by the vehicle, and calculates the output variable λ. This value is passed to the lower-layer PMP energy management strategy, which adjusts the power distribution of the engine and the power battery at each moment in real time according to the change of λ to meet the total power demand of the vehicle at each moment.

结合图1与图3可以看出本实施例的优点在于：1 and 3, it can be seen that the advantages of this embodiment are:

(1)通过所述发明能够实现PMP算法在PHEV能量管理过程中的实时应用，大大简化了策略的复杂程度，使策略具有较好的实时性；(1) The invention can realize the real-time application of the PMP algorithm in the PHEV energy management process, greatly simplify the complexity of the strategy, and make the strategy have better real-time performance;

(2)通过无模型自适应控制算法的引入，车辆能够在未知工况行驶过程中自适应调整PMP算法的协态变量，使其较好地收敛于全局工况已知条件下得到的恒定协态变量，与传统PID自适应控制相比，可以实现主动收敛，控制性能更稳定；(2) Through the introduction of the model-free adaptive control algorithm, the vehicle can adaptively adjust the co-state variables of the PMP algorithm in the process of unknown operating conditions, so that it can better converge to the constant co-variable obtained under the known conditions of the global operating conditions. Compared with the traditional PID adaptive control, it can achieve active convergence and the control performance is more stable;

(3)从车辆仿真得到的电池SOC来看，能够满足电池SOC初值和终值的设定，在车辆运行过程中，电池SOC出现一定的波动，能够较好地利用电能驱动车辆，并且能够合理利用制动能量回收给电池充电，从而降低发动机能耗；(3) Judging from the battery SOC obtained by vehicle simulation, it can meet the setting of the initial value and final value of the battery SOC. During the operation of the vehicle, the battery SOC fluctuates to a certain extent, which can make better use of electric energy to drive the vehicle, and can Reasonable use of braking energy recovery to charge the battery, thereby reducing engine energy consumption;

(4)从发动机能耗来看，采用本发明所述策略后，发动机能耗远低于传统规则型策略，其能耗介于全局最优与PID自适应控制之间，该策略在提升车辆经济性方面具有显著作用。(4) From the perspective of engine energy consumption, after the strategy of the present invention is adopted, the engine energy consumption is much lower than that of the traditional rule-based strategy, and its energy consumption is between the global optimum and PID adaptive control. Economical aspect has a significant effect.

以上所述仅为本发明较佳可行的实施例而已，并非因此局限本发明的权利范围，凡运用本发明说明书及附图内容所作的等效结构变化，均包含于本发明的权利范围之内。The above descriptions are only preferred feasible embodiments of the present invention, and are not intended to limit the scope of rights of the present invention. Any equivalent structural changes made by using the contents of the description and accompanying drawings of the present invention are included in the scope of rights of the present invention. .

Claims (4)

1. The PHEV real-time energy management strategy based on model-free adaptive control is characterized by comprising the following steps of:

the energy management strategy is divided into an upper layer of the energy management strategy and a lower layer of the energy management strategy by adopting a layered optimization control architecture;

the upper layer of the energy management strategy is a control layer based on an MFAC algorithm, the control input of the energy management strategy is a power battery reference SOC and a battery actual SOC based on a feedback control design, the control output of the energy management strategy is a covariate lambda of a lower layer PMP algorithm, and the upper layer of the energy management strategy outputs control parameters according to self-adaption to running condition changes;

the lower layer of the energy management strategy is an energy management optimization layer based on a PMP algorithm, and the energy output among multiple power sources of the hybrid power system is optimized in real time; the PMP algorithm takes the minimum energy consumption of an engine as an objective function and respectively takes the output power P of a power battery_batt The SOC and the power battery are control variables and state variables, and key control parameters are co-modal variables lambda; and the lower layer of the energy management strategy completes the optimal distribution of energy among different power sources of the vehicle power system according to the control parameters output by the upper layer of the energy management strategy.

2. The model-free adaptive control-based PHEV real-time energy management strategy of claim 1, wherein: the energy management optimization layer based on the PMP algorithm is used for optimizing the energy output among multiple power sources of the hybrid power system in real time and comprises the following steps:

(1) considering that the PHEV can obtain sufficient electric energy through the power grid, under certain driving conditions, in order to improve vehicle economy, the battery energy is basically exhausted, therefore, the optimization objective of the PMP energy management strategy is mainly engine energy consumption, and the optimization objective is expressed as:

wherein t represents time, x (t) and u (t) represent a state variable and a control variable, respectively,

instantaneous energy consumption of the engine;

(2) respectively adopting the battery SOC and the power battery output power P_batt As state variables and control variables, the hamiltonian of the control system can be expressed as:

in the formula, lambda is a covariate;

(3) as can be seen from the above equation, the performance of the PMP energy management strategy depends on the selection of the key parameter λ, that is, the PMP energy management optimization problem of the lower layer can be converted into the optimization problem of the covariate λ by equation (2), where λ and the dynamic equation of the battery SOC can be expressed as:

(4) in the PMP algorithm solving process, the charging and discharging performance of the power battery is considered, and the SOC of the battery at the initial time and the terminal time is respectively set as the SOC (t)₀ )＝0.8，SOC(t_f ) 0.3; in addition, considering the limitation of the physical characteristics of the vehicle power system, the algorithm solving process also needs to meet the following physical limitations:

in the formula, n_e (t) and n_m (t) indicates the rotational speeds of the engine and the motor, n_{e_min} (t),n_{e_max} (t),n_{m_min} (t),n_{m_max} (t) minimum and maximum speed limits, respectively; p_e (t) and P_m (t) is the engine and motor power, P_{e_min} (t),P_{e_max} (t),P_{m_min} (t),P_{m_max} (t) represents the minimum and maximum power limits, respectively.

3. The model-free adaptive control-based PHEV real-time energy management strategy of claim 1, wherein: the MFAC algorithm-based upper layer outputs control parameters for the lower layer according to the self-adaption to the change of the driving condition, and the MFAC algorithm-based upper layer comprises the following steps: pseudo partial derivative of PPD (direct digital display) -based estimation algorithm to system at each moment

Carrying out online estimation; obtaining a pseudo partial derivative estimated value according to the difference between the battery reference SOC and the actual SOC, the system output lambda (k-1) at the last moment and an estimation algorithm based on an MFAC algorithm

And calculating to obtain the output of the system at the next moment.

4. The model-free adaptive control-based PHEV real-time energy management strategy of claim 3, wherein: the system model established by adopting the MFAC algorithm is as follows:

wherein SOC (k) and λ (k) are the state variable and output variable of the system at time k, respectively representing the current SOC and the co-state variable λ of the battery,

is the pseudo partial derivative of the system; the control law can be expressed as:

in the formula, ρ_k Is a step size factor, λ_k Are weighting factors, both positive numbers less than 1; SOC^* (k +1) represents the battery reference SOC at the time k +1,

is composed of

The estimated value of (c), i.e. the estimated value of the pseudo-partial derivative, can be obtained by the following equation:

in the formula eta_k And mu_k Step-size factors and weight factors, respectively, both being positive numbers smaller than 1, epsilon representing a very small positive number,

is composed of

An initial value of (d); in the PHEV driving process, the MFAC controller can adjust and output a co-modal variable lambda for online optimization of a lower-layer energy management strategy in real time according to a preset power battery reference SOC and a battery SOC actually fed back by a vehicle; in order to improve the applicability of the invention, the setting of the power battery SOC reference track can be planned according to the target driving mileage of the vehicle, and the battery reference SOC can be expressed as:

in the formula, SOC₀ ，SOC_f Respectively an initial SOC value and a final SOC value; d_dist And D_total And representing the current driving mileage and the target driving mileage.

Images