﻿ 基于遗传算法的差速耦合式混合动力系统模糊控制策略优化 Optimization of Fuzzy Control Strategy for Differential Coupling Hybrid Power System Based on Genetic Algorithm

Dynamical Systems and Control
Vol. 08  No. 02 ( 2019 ), Article ID: 29463 , 13 pages
10.12677/DSC.2019.82010

Optimization of Fuzzy Control Strategy for Differential Coupling Hybrid Power System Based on Genetic Algorithm

Xin Wang1, Qing Zhang1, Yin Wang1, Feng Xiao2*, Jianjun Hu2

1Power Research Institute of Chang’an Automobile, Chongqing

2State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing

Received: Mar. 6th, 2019; accepted: Mar. 17th, 2019; published: Mar. 27th, 2019

ABSTRACT

In order to improve the economic performance and emission performance of hybrid vehicles, this paper is based on the study of the structure and working characteristics of the differential coupled power system. The ADVISOR simulation software is used to redevelop the model of the differential coupled power system. Two-input single-output fuzzy control strategy was designed for the input and the demand torque as input, and the engine was started and stopped. The 25 fuzzy rule variables were optimized using the genetic algorithm, and the vehicle performance simulation and logic were performed under UDDS conditions. The performance was compared under the threshold strategy. The results show that the optimized fuzzy control strategy can more reasonably control the start and stop of the engine of the car, and obviously improve the economy under the premise of ensuring the power of the car, and can effectively reduce the car emissions.

Keywords:ADVISOR Redevelopment, Differential Coupling Hybrid, Fuzzy Control Strategy, Genetic Algorithm

1重庆长安汽车股份有限公司动力研究院，重庆

2重庆大学机械传动国家重点实验室，重庆

1. 引言

2. 系统建模

2.1. 结构及工作特性

Figure 1. Differential coupling hybrid power system structure

Table 1. Vehicle performance index

Table 2. Vehicle system parameters

2.2. 数学模型

${w}_{2}=\frac{{w}_{e}}{{i}_{1}},{w}_{3}={w}_{m}$ (1)

${T}_{2}={T}_{e}\cdot {i}_{1},{T}_{3}={T}_{L}-{T}_{m}$ (2)

${T}_{L}={T}_{3}+{T}_{m}$ (3)

${T}_{g}={T}_{3}$ (4)

(5)

${T}_{e}\cdot {w}_{e}={T}_{2}\cdot {w}_{2}=\left({T}_{L}-{T}_{m}\right)\cdot {w}_{m}+{T}_{g}\cdot {w}_{g}$ (6)

$2\cdot {w}_{2}={w}_{g}+{w}_{3}={w}_{g}+{w}_{m}=\frac{2\cdot {w}_{e}}{{i}_{1}}$ (7)

${T}_{g}=\frac{{T}_{2}}{2}=\frac{{T}_{e}\cdot {i}_{1}}{2}$ (8)

${T}_{m}={T}_{L}-\frac{{T}_{2}}{2}={T}_{L}-\frac{{T}_{e}\cdot {i}_{1}}{2}$ (9)

2.3. 驱动模式与能量路线

Table 3. Energy flow paths in various modes

2.4. 差速耦合器与发动机开关控制

Figure 2. Engine switch control module

3. 模糊控制器的设计

1) 选择输入、输出变量

2) 输入量的模糊化

Figure 3. SOC subordinating degree function

Figure 4. Required torque membership function

3) 输出量的模糊化

Figure 5. The membership function of the output quantity

4) 模糊规则

Table 4. Initial fuzzy control rules

Figure 6. Initial input and output fuzzy 3D curve

4. 遗传算法优化模糊控制器

4.1. 遗传算法优化模糊控制器过程

1) 工作空间初始化

2) 遗传算法的基本参数确定。取初始种群50个体数目，最大遗传代数50代，优化变量个数25个，每个变量的二进制位数为3位，代沟取0.9，交叉概率0.97，变异概率0.001。

3) 创建初始种群并译码。使用函数crtbp来初始种群，函数bs2rv将二进制转换为十进制。最后得到一个包含50个个体，每个个体包含25个变量的初始种群。

4) 目标函数与适应度函数的建立。由于最大爬坡度与经济性和排放性的指标相反，爬坡度越大越好，经济性和排放性越小越好。考虑经济性、排放性和动力性等目标，本文利用爬坡度上限60%与最大爬坡度差值作为一个目标值，采用加权系数法来将多目标转换为单目标函数。同时以经济性为主，排放性和动力性为辅，取等效汽油消耗率的加权系数为0.7，动力性的加权系数为0.1，排放性各取0.05，采用排序法来进行适应度的分配。

Figure 7. The flow field diagram of fuzzy controller is optimized by genetic algorithm

5) 遗传操作。使用select函数、recombin函数、mut函数和reins函数作为遗传操作函数，分别进行选择、交叉、变异和重插入。

6) 终止判断。当遗传代数超过50代时，遗传算法终止。

4.2. 优化结果

5. 仿真分析

Figure 8. Genetic convergence process

Figure 9. The target value for each individual after the 50th iteration

Figure 10. The input and output fuzzy 3D curves are optimized

Figure 11. UDDS drive cycle

5.1. 发动机启停仿真

Figure 12. Engine startup time ratio

Figure 13. Engine start-stop accumulation area diagram

5.2. 动力性仿真

Figure 14. Simulation results of dynamic performance

5.3. 经济性仿真

Figure 15. Economic simulation results

5.4. 排放性仿真

Figure 16. Emission simulation results

6. 结束语

2) 将优化模糊控制策略前后的差速耦合式混合动力系统性能仿真结果对比，得到优化后的百公里综合油耗下降了13.8%，排放性中HC排放下降了35.14%、CO排放下降了34.15%、NOx排放下降了23.85%。动力性几乎没变。证明了遗传算法对于差速耦合混合动力系统的模糊控制优化的可行性。

3) 将遗传算法优化后的模糊控制策略的结果与逻辑门限策略控制发动机启停的性能仿真结果对比，得到优化模糊策略后的百公里综合油耗下降了23%，排放性中HC排放下降了27.71%、CO排放下降了26.46%、NOx排放下降了33.1%。动力性虽然轻微下降但仍远满足要求。

4) 基于遗传算法优化后的模糊控制策略能更合理地控制发动机的启停，极大地提高了汽车的经济性和排放性。

Optimization of Fuzzy Control Strategy for Differential Coupling Hybrid Power System Based on Genetic Algorithm[J]. 动力系统与控制, 2019, 08(02): 81-93. https://doi.org/10.12677/DSC.2019.82010

1. 1. 苏天晨. 差速耦合式混合动力汽车模式切换的协调控制[M]. 吉林: 吉林大学, 2011.

2. 2. 于永涛, 王庆年, 曾小华, 等. 传统差速器应用于HEV动力耦合的可行性[J]. 吉林大学学报, 2010, 40(2): 305-310.

3. 3. 杨章林. 差速器式HEV动力耦合装置设计分析[J]. 黑龙江工业学院学报, 2017, 17(9): 50-53.

4. 4. 曾小华, 于永涛, 宋大凤, 等. 基于差速耦合的混合动力汽车传递特性分析与控制[J]. 吉林大学学报, 2010, 40(1): 1-5.

5. 5. Zeng, X., Nie, L. and Wang, Q. (2011) Experimental Study on the Differential Hybrid System Hybrid Electric Vehicle. Procedia Engineering, 16, 708-715.

6. 6. 邹彦艳, 吴宇轩, 宋振宇, 等. 基于改进遗传算法的模糊控制器设计[J]. 自动化技术与应用, 2013, 32(11): 6-10.

7. 7. Hwang, H.S., Joo, Y.H., Kim, H.K., et al. (2014) Identification of Fuzzy Control Rules Utilizing Genetic Algorithms and Its Application to Mobile Robots. Algorithms and Architectures for Real-Time Control, 249-254.

8. 8. 葛安东. 基于遗传算法和模糊控制的某型纯电动汽车整车控制器的研究[D]: [硕士学位论文]. 合肥: 合肥工业大学, 2015.

9. 9. 中国国家标准化管理委员会. GB/T 19753-2013轻型混合动力电动汽车能力消耗量试验方法[S]. 北京: 中国标准出版社, 2013: 4-5.

10. 10. 吕力, 葛鹏, 柳邵辉. 中美插电式混合动力汽车能耗评价方法[J]. 汽车工程师, 2016(10): 55-58.

11. NOTES

*通讯作者。