阅读说明：本技术 考虑动态响应能力的基于神经网络的汽车动力控制方法 (Automobile power control method based on neural network and considering dynamic response capability ) 是由 宋珂 丁钰航 王一旻 徐宏杰 于 2021-01-20 设计创作，主要内容包括：本发明涉及一种考虑动态响应能力的基于神经网络的汽车动力控制方法,包括：S1：实时获取汽车的整车工况特征速度、动力系统的需求功率、能量源的功率和蓄电池SOC；S2：根据蓄电池SOC判断燃料电池是否开启,若燃料电池开启则执行步骤S3；S3：将整车工况特征速度、动力系统的需求功率和能量源的功率载入神经网络中,获取能量源的当前最优功率分配参数；S4：根据燃料电池系统的动态响应能力曲线,对能量源的当前最优功率分配参数进行修正；S5：根据修正后的能量源的当前最优功率分配参数,对能量源的输出功率进行分配控制。与现有技术相比,本发明考虑了燃料电池的动态响应能力,而且具有燃油经济性好、结构简单、易于实车实现等优点。(The invention relates to an automobile power control method based on a neural network and considering dynamic response capability, which comprises the following steps: s1: acquiring the characteristic speed of the working condition of the whole automobile, the required power of a power system, the power of an energy source and the SOC of a storage battery of the automobile in real time; s2: judging whether the fuel cell is started or not according to the SOC of the storage battery, and executing step S3 if the fuel cell is started; s3: loading the characteristic speed of the whole vehicle working condition, the required power of a power system and the power of an energy source into a neural network, and acquiring the current optimal power distribution parameter of the energy source; s4: correcting the current optimal power distribution parameter of the energy source according to the dynamic response capacity curve of the fuel cell system; s5: and performing distribution control on the output power of the energy source according to the corrected current optimal power distribution parameter of the energy source. Compared with the prior art, the fuel cell system has the advantages of considering the dynamic response capability of the fuel cell, along with good fuel economy, simple structure, easy realization of real vehicles and the like.)

1. A neural network-based power control method for an automobile taking dynamic response capability into consideration, an energy source of a power system of the automobile including a fuel cell and a battery, the method comprising the steps of:

s1: acquiring energy state data of the automobile in real time, wherein the energy state data comprise the characteristic speed of the working condition of the whole automobile, the required power of a power system, the power of an energy source and the SOC of a storage battery;

s2: judging whether the fuel cell is started or not according to the SOC of the storage battery obtained in the step S1, and executing a step S3 if the fuel cell is started;

s3: loading the working condition characteristic speed of the whole vehicle, the required power of the power system and the power of the energy source, which are obtained in the step S1, into a pre-established and trained neural network, and obtaining the current optimal power distribution parameters of the energy source;

s4: correcting the current optimal power distribution parameter of the energy source according to a pre-acquired dynamic response capacity curve of the fuel cell system;

s5: according to the corrected current optimal power distribution parameter of the energy source obtained in the step S4, performing distribution control on the output power of the energy source;

the training process of the neural network comprises the steps of obtaining state data of an energy source of an automobile under each working condition, calculating a global optimization distribution result of the energy source under each known working condition according to the state data by adopting a dynamic planning method, constructing a training set according to the characteristic speed of the working condition of the whole automobile corresponding to each known working condition and the global optimization distribution result of the energy source, and training the neural network by adopting the training set.

2. The vehicle power control method based on the neural network considering the dynamic response capability of the claim 1, wherein the characteristic speed of the vehicle condition comprises an average speed, an average acceleration, a maximum acceleration, a minimum acceleration and an idle rate of the vehicle.

3. The method for controlling the vehicle power based on the neural network considering the dynamic response capability of the claim 1, wherein the expression of the cost function J of the dynamic programming method is as follows:

in the formula (I), the compound is shown in the specification,as the amount of hydrogen consumption of the fuel cell,is the equivalent hydrogen consumption of the battery, P_fcIs the fuel cell power, η_fcIn order to be efficient for the fuel cell,being of low calorific value, P, of hydrogen_batIs the battery power, η_disIn order to achieve the discharge efficiency of the secondary battery,η_chafor the charging efficiency of the accumulator, eta_cha,avgIs the average charging efficiency, η, of the battery_dis,avgThe average discharge efficiency of the storage battery; m is_fc,avgN is the number of steps, Δ t is the step size of a single step, P is the average instantaneous hydrogen consumption of the fuel cell_DCThe output power is DC/DC connected with the storage battery.

4. The method as claimed in claim 3, wherein the dynamic programming method comprises the following steps:

f_min(P_fc(t))≤P_fc(t+1)≤f_max(P_fc(t))

in the formula, P_fc(t +1) is the fuel cell power at time t +1, f_min(P_fc(t)) is the minimum value of the fuel cell power at time t, f_max(P_fc(t)) is the maximum value of the fuel cell power at time t.

5. The method for controlling vehicle power based on neural network according to claim 1, wherein in step S4, the current optimal power distribution parameters of the energy source include the required power of fuel cell and the required power of storage battery, and the modification of the current optimal power distribution parameters of the energy source is specifically:

calculating the power change rate of the fuel cell at the current moment according to the required power of the fuel cell; and based on the dynamic response capability curve of the fuel cell, if the power change rate of the fuel cell at the current moment is larger than the maximum change rate determined by the dynamic response capability of the fuel cell, outputting the required power of the fuel cell according to the maximum change rate.

6. The method for controlling vehicle power based on neural network considering dynamic response capability of claim 1, wherein in step S2, the step of determining whether the fuel cell is turned on is specifically as follows:

if the battery SOC acquired in step S1 is greater than 0.7, the fuel cell is not turned on, and if the battery SOC is less than 0.7, the fuel cell is turned on.

7. The method of claim 1, wherein the neural network is a NARX neural network.

8. The vehicle power control method considering dynamic response capability based on the neural network as claimed in claim 1, wherein the vehicle power system comprises a vehicle control unit, a CAN bus, an energy source, an energy controller and vehicle power accessories, and the vehicle power control method is executed by the vehicle control unit.

9. The neural network-based vehicle power control method considering dynamic response capability of claim 1, wherein in step S3, the required power of the powertrain includes an integral of the vehicle required power and the vehicle required power in the preset first time period before the current time.

10. The neural network-based vehicle power control method considering dynamic response capability of claim 1, wherein the power of the energy source includes the fuel cell power and the integral of the fuel cell power in the preset first time period before the current time at step S3.

Technical Field

The invention relates to the field of fuel cell automobile power control, in particular to an automobile power control method based on a neural network and considering dynamic response capability.

Background

Pure fuel cell vehicles have several disadvantages: the starting time is long, and the cold starting performance is poor; the dynamic response of the system is slow; when the output power is low and high, the efficiency of the fuel cell is low; energy cannot be recovered by regenerative braking. To overcome these disadvantages, fuel cells are typically used in conjunction with other energy storage devices, such as batteries and supercapacitors. Therefore, energy distribution among multiple energy sources, i.e., energy management strategies, is one of the key research issues in fuel cell vehicle design. The performance of the fuel cell vehicle is closely related to the energy management strategy, and the optimal energy management strategy can not only improve the economy of the whole vehicle, but also improve the service life of a power supply system.

Energy Management Strategies (EMS) of fuel cell vehicles can be largely classified into two types according to control methods: rule-based policies and optimization algorithm-based policies. Rule-based energy management strategies design certain rules between the fuel cell system and the battery to distribute the required power based on experimental results or research experience. The rule-based energy management strategy, while simple, is less economical.

Optimization-based EMS are generally divided into two categories: global optimization strategies and transient optimization strategies. The global optimization strategy requires that driving conditions are known in advance and therefore cannot be applied in real time. The transient optimization strategy solves the optimization problem by defining a transient cost function that is updated over time. The premise of realizing energy management by common transient optimization methods such as random dynamic programming and model predictive control is that driving condition prediction is adopted, and in actual conditions, behaviors and traffic conditions of drivers often have great randomness, so that the prediction precision is difficult to ensure.

The invention with publication number CN102951144B discloses a self-adjusting neural network energy management method based on a minimum power loss algorithm, which comprises the following steps: 1) the vehicle control unit acquires data required by energy management strategy calculation from the vehicle power accessory through the CAN bus, and simultaneously acquires the real-time efficiency value of the current energy source; 2) the vehicle control unit judges whether complete data are received; 3) the vehicle control unit judges whether an instruction of updating the neural network is received, and if so, the neural network is updated; 4) the vehicle control unit calculates required data according to the received energy management strategy, and calculates the current optimal power distribution through a neural network; 5) correcting the optimal power distribution calculated by the neural network by using the power gain coefficient; 6) and the vehicle control unit sends a power distribution result to the energy controller through the CAN bus.

In the training process of the neural network, the power distribution result is obtained through the minimum power loss algorithm, the working performance of the fuel cell and the working performance of the storage battery are not considered comprehensively, and the service life of the fuel cell is reduced to some extent when the scheme is adopted.

Disclosure of Invention

The invention aims to overcome the defects that the prediction precision is difficult to ensure and the service life of a fuel cell is reduced in the prior art, and provides an automobile power control method based on a neural network considering dynamic response capability.

The purpose of the invention can be realized by the following technical scheme:

a neural network-based vehicle power control method considering dynamic response capability, an energy source of a power system of the vehicle including a fuel cell and a battery, the method comprising the steps of:

s1: acquiring energy state data of the automobile in real time, wherein the energy state data comprise the characteristic speed of the working condition of the whole automobile, the required power of a power system, the power of an energy source and the SOC of a storage battery;

s2: judging whether the fuel cell is started or not according to the SOC of the storage battery obtained in the step S1, and executing a step S3 if the fuel cell is started;

s3: loading the working condition characteristic speed of the whole vehicle, the required power of the power system and the power of the energy source, which are obtained in the step S1, into a pre-established and trained neural network, and obtaining the current optimal power distribution parameters of the energy source;

s4: correcting the current optimal power distribution parameter of the energy source according to a pre-acquired dynamic response capacity curve of the fuel cell system;

s5: according to the corrected current optimal power distribution parameter of the energy source obtained in the step S4, performing distribution control on the output power of the energy source;

the training process of the neural network comprises the steps of obtaining state data of an energy source of an automobile under each working condition, calculating a global optimization distribution result of the energy source under each known working condition according to the state data by adopting a dynamic planning method, constructing a training set according to the characteristic speed of the working condition of the whole automobile corresponding to each known working condition and the global optimization distribution result of the energy source, and training the neural network by adopting the training set.

Further, the characteristic speed of the working condition of the whole vehicle comprises the average speed, the average acceleration, the maximum acceleration, the minimum acceleration and the idle speed of the whole vehicle.

Further, the expression of the cost function J of the dynamic programming method is:

in the formula (I), the compound is shown in the specification,as the amount of hydrogen consumption of the fuel cell,is the equivalent hydrogen consumption of the battery, P_fcIs the fuel cell power, η_fcIn order to be efficient for the fuel cell,being of low calorific value, P, of hydrogen_batIs the battery power, η_disIs the discharge efficiency of the accumulator eta_chaFor the charging efficiency of the accumulator, eta_cha,avgFor average charging of accumulatorsEfficiency, η_dis,avgThe average discharge efficiency of the storage battery; m is_fc,avgN is the number of steps, Δ t is the step size of a single step, P is the average instantaneous hydrogen consumption of the fuel cell_DCThe output power is DC/DC connected with the storage battery.

Further, in the planning process of the dynamic planning method, the expression of the optimizing range of the fuel cell power is as follows:

f_min(P_fc(t))≤P_fc(t+1)≤f_max(P_fc(t))

in the formula, P_fc(t +1) is the fuel cell power at time t +1, f_min(P_fc(t)) is the minimum value of the fuel cell power at time t, f_max(P_fc(t)) is the maximum value of the fuel cell power at time t.

Further, in step S4, the current optimal power distribution parameter of the energy source includes a required power of the fuel cell and a required power of the storage battery, and the modifying the current optimal power distribution parameter of the energy source specifically includes:

calculating the power change rate of the fuel cell at the current moment according to the required power of the fuel cell; and based on the dynamic response capability curve of the fuel cell, if the power change rate of the fuel cell at the current moment is larger than the maximum change rate determined by the dynamic response capability of the fuel cell, outputting the required power of the fuel cell according to the maximum change rate.

Further, in step S2, the specific determination as to whether the fuel cell is on is:

if the battery SOC acquired in step S1 is greater than 0.7, the fuel cell is not turned on, and if the battery SOC is less than 0.7, the fuel cell is turned on.

Further, the neural network is a NARX neural network.

Further, the power system of the automobile comprises a vehicle control unit, a CAN bus, an energy source, an energy controller and automobile power accessories, and the automobile power control method is executed through the vehicle control unit.

Further, in step S3, the required power of the powertrain includes the integral of the vehicle required power and the vehicle required power in the preset first time period before the current time.

Further, in step S3, the power of the energy source includes the fuel cell power and the integral of the fuel cell power in the preset first time period before the current time.

Compared with the prior art, the invention has the following advantages:

(1) the automobile power control method based on the neural network avoids the influence on energy management caused by inaccuracy of working condition prediction on one hand, and can realize energy management on various working conditions only by one neural network on the other hand, so the structure is simpler; because the neural network is trained by taking the result of dynamic planning as a data set, the established energy management strategy also has good overall economy;

in addition, the invention considers that the dynamic response process of the fuel cell system is slow, generally about 10s, and the larger the change rate of the required power is, the longer the response time of the system is. The required power of the whole vehicle usually changes dozens of or even dozens of kW within 1s, although the fuel cell is connected with the bus through the DC/DC, the response of the DC/DC is very quick (usually in millisecond level), so if the power change rate of the fuel cell is not limited in an energy management strategy, the DC/DC can pull current blindly, the fuel cell can not reach the required power, and fuel starvation can be caused, parts in a stack are damaged, and the service life of the fuel cell is reduced;

according to the invention, the current optimal power distribution parameter of the energy source output by the neural network is corrected according to the dynamic response capability curve of the fuel cell system, so that on one hand, the economy of the whole vehicle can be improved, and on the other hand, the fuel cell system can be prevented from being damaged due to the fact that the required power change rate of the fuel cell calculated by an energy management strategy is too large.

(2) The automobile power control method based on the neural network considering the dynamic response capability can be conveniently applied to hybrid power systems such as an internal combustion engine/a storage battery, a fuel cell/a super capacitor and the like, and has good expansibility.

Drawings

FIG. 1 is a schematic flow chart of a neural network-based vehicle power control method with dynamic response capability taken into account in an embodiment of the present invention;

fig. 2 is a schematic flow chart of a dynamic programming algorithm considering the dynamic response capability of the fuel cell in the embodiment of the invention.

Detailed Description

The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.

Example 1

As shown in fig. 1, the present embodiment provides a method for controlling power of an automobile based on a neural network considering dynamic response capability, a power system of the automobile includes a vehicle controller, a CAN bus, an energy source, an energy controller and power accessories of the automobile, the energy source includes a fuel cell and a battery, and the method includes the following steps:

s1: acquiring energy state data of the automobile in real time, wherein the energy state data comprise the characteristic speed of the working condition of the whole automobile, the required power of a power system, the power of an energy source and the SOC of a storage battery;

s2: judging whether the fuel cell is started or not according to the SOC of the storage battery obtained in the step S1, and executing a step S3 if the fuel cell is started;

s3: loading the working condition characteristic speed of the whole vehicle, the required power of the power system and the power of the energy source, which are obtained in the step S1, into a pre-established and trained neural network, and obtaining the current optimal power distribution parameters of the energy source;

s4: correcting the current optimal power distribution parameter of the energy source according to a pre-acquired dynamic response capacity curve of the fuel cell system;

s5: according to the corrected current optimal power distribution parameter of the energy source obtained in the step S4, performing distribution control on the output power of the energy source;

the training process of the neural network comprises the steps of obtaining state data of an energy source of an automobile under each working condition, calculating a global optimization distribution result of the energy source under each known working condition according to the state data by adopting a dynamic planning method, constructing a training set according to the characteristic speed of the working condition of the whole automobile corresponding to each known working condition and the global optimization distribution result of the energy source, and training the neural network by adopting the training set.

The expression of the cost function J of the dynamic programming method is:

in the formula (I), the compound is shown in the specification,as the amount of hydrogen consumption of the fuel cell,is the equivalent hydrogen consumption of the battery, P_fcIs the fuel cell power, η_fcIn order to be efficient for the fuel cell,being of low calorific value, P, of hydrogen_batIs the battery power, η_disIs the discharge efficiency of the accumulator eta_chaFor the charging efficiency of the accumulator, eta_cha,avgIs the average charging efficiency, η, of the battery_dis,avgThe average discharge efficiency of the storage battery; m is_fc,avgN is the average instantaneous hydrogen consumption of the fuel cell, n is the step number of the dynamic programming method, Δ t is the single step length of the dynamic programming method, P_DCThe output power is DC/DC connected with the storage battery.

In the planning process of the dynamic planning method, the expression of the optimizing range of the fuel cell power is as follows:

f_min(P_fc(t))≤P_fc(t+1)≤f_max(P_fc(t))

in the formula, P_fc(t +1) is the fuel cell power at time t +1, f_min(P_fc(t)) is the minimum value of the fuel cell power at time t, f_max(P_fc(t)) is the maximum value of the fuel cell power at time t.

As a preferred implementation mode, the characteristic speed of the working condition of the whole vehicle comprises the average speed, the average acceleration, the maximum acceleration, the minimum acceleration and the idle speed of the whole vehicle.

In a preferred embodiment, in step S3, the required power of the powertrain includes an integral of the vehicle required power and the vehicle required power in a preset first time period before the current time.

As a preferred embodiment, in step S3, the power of the energy source includes the fuel cell power and the integral of the fuel cell power in the preset first time period before the current time.

As a preferred embodiment, in step S4, the current optimal power allocation parameter of the energy source includes a required power of the fuel cell and a required power of the storage battery, and the modifying the current optimal power allocation parameter of the energy source specifically includes:

calculating the power change rate of the fuel cell at the current moment according to the required power of the fuel cell; and based on the dynamic response capability curve of the fuel cell, if the power change rate of the fuel cell at the current moment is larger than the maximum change rate determined by the dynamic response capability of the fuel cell, outputting the required power of the fuel cell according to the maximum change rate.

As a preferred embodiment, the step S2 of determining whether the fuel cell is on is specifically:

if the battery SOC acquired in step S1 is greater than 0.7, the fuel cell is not turned on, and if the battery SOC is less than 0.7, the fuel cell is turned on.

In a preferred embodiment, the neural network is a narx (nonlinear autoregegressive with external input) neural network.

As a preferred embodiment, the hybrid power system to which the neural network-based automobile power control method considering the dynamic response capability is applied includes a Vehicle Management System (VMS), a CAN bus, an energy source (fuel cell and battery), an energy controller, and automobile power accessories. The VMS controls all parts of the whole vehicle, and the energy management strategy is just one of main control software; the CAN bus is used for information communication among all parts of the whole vehicle; the energy source provides energy for the vehicle to run, the controller mainly controls the power output of the energy source, and the automobile power accessories comprise auxiliary systems such as a heat dissipation system and an air conditioning system.

The above preferred embodiments are combined to obtain an optimal embodiment, and a specific implementation process of the optimal embodiment is described below.

As shown in fig. 1, a neural network-based vehicle power control method considering dynamic response capability trains a non-linear autoregressive exogenous (NARX) neural network using a global optimization result calculated off-line by dynamic programming as a data set. Inputs to the NARX neural network include: the vehicle power demand, the integral of the power demand, the fuel cell power, the integral of the fuel cell power over a past time domain, and the average speed, the average acceleration, the maximum acceleration, the minimum acceleration, and the idle rate representing different operating conditions. In addition, the power change rate of the fuel cell is further constrained according to the dynamic response capability curve of the fuel cell.

The automobile power control method considering the dynamic response capability based on the neural network specifically comprises the following steps:

1) the vehicle control unit sends an access signal to the vehicle power accessory through the CAN bus, and obtains data required by energy management strategy calculation from the vehicle power accessory, wherein the data comprises vehicle speed, required power, storage battery SOC and the like. Then step 2) is executed;

2) judging whether the fuel cell is started or not according to the SOC of the storage battery, and if the SOC is larger than 0.7, not starting the fuel cell; if SOC is less than 0.7, otherwise, executing step 3);

3) and the vehicle control unit calculates required data according to the received energy management strategy and calculates the current optimal power distribution through a neural network. Then executing step 4)

The calculation of the current optimal power allocation through the neural network specifically comprises the following steps:

a) and (3) performing off-line calculation, namely calculating a global optimization distribution result under a known working condition by using a dynamic programming method, wherein a cost function J of the dynamic programming is as follows:

in the formula (I), the compound is shown in the specification,as the amount of hydrogen consumption of the fuel cell,is the equivalent hydrogen consumption of the battery, P_fcIs the fuel cell power, η_fcIn order to be efficient for the fuel cell,being of low calorific value, P, of hydrogen_batIs the battery power, η_disIs the discharge efficiency of the accumulator eta_chaFor the charging efficiency of the accumulator, eta_cha,avgIs the average charging efficiency, η, of the battery_dis,avgThe average discharge efficiency of the storage battery; m is_fc,avgIs the average instantaneous hydrogen consumption of the fuel cell.

It should be noted that, in the solution optimization process of the dynamic programming, the optimization range of the fuel cell power at the current moment depends on the fuel cell power at the previous moment, and the limitation can ensure that the required power calculated by the dynamic programming does not exceed the dynamic response capability of the fuel cell.

Specifically, the expression for the optimum range of fuel cell power is:

f_min(P_fc(t))≤P_fc(t+1)≤f_max(P_fc(t))

in the formula, P_fc(t +1) is the fuel cell power at time t +1, f_min(P_fc(t)) is the minimum value of the fuel cell power at time t, f_max(P_fc(t)) is the maximum value of the fuel cell power at time t.

b) Training a neural network according to the result of the dynamic programming off-line calculation, wherein the input of the neural network comprises characteristic parameters which can represent different working conditions: average speed, average acceleration, maximum acceleration, minimum acceleration, and idle rate;

c) calculating the current optimal power distribution through the trained neural network;

4) and limiting the power change rate of the fuel cell according to the dynamic response capability curve of the fuel cell system, and correcting the power distribution result. Then entering step 5);

5) the vehicle control unit sends power requirements to the fuel cell and the storage battery through the CAN bus to complete the distribution control of the vehicle control unit on the output power of each energy source in the hybrid power system.

Fig. 2 is a flow chart of a dynamic programming algorithm that takes into account the dynamic response capability of the fuel cell. The dynamic programming solving process is as follows: inputting a known working condition; setting the SOC of the storage battery, the power of the fuel cell and the power range of the storage battery; calculating the power optimizing range of the fuel cell at the current moment according to the power of the fuel cell at the previous moment; calculating the total hydrogen consumption of the whole vehicle under different fuel cell powers within the optimizing range to obtain the optimal fuel cell power; judging whether the solved solution is the optimal solution or not through a state transition equation; and repeating the steps until the calculation of the whole working condition is completed, and then finding out the optimal fuel cell output power curve and the SOC curve under the working condition.

The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.