基于模拟退火粒子群算法的混合动力车参数优化

Parameters Optimization for HEV Based on Simulated Annealing Particle Swarm Algorithm

分别采用多层次参数扫描(MLPS)算法和模拟退火粒子群优化(SAPSO)算法对并联式混合动力车逻辑门限控制策略的参数进行优化。将优化后的车辆以TEST-CITY-HWY测试循环进行仿真,并将结果与优化前的车辆的仿真结果进行对比。结果表明,经MLPS算法优化后,燃油消耗和HC与NOx排放分别下降了11.98%、6.01%和4.03%,但CO排放增加了25.18%;经SAPSO算法优化后,燃油消耗和HC、CO与NOx排放分别下降了13.61%、9.57%、27.78%和18.53%,且电池荷电状态(SOC)比MLPS优化略高。说明SAPSO算法在混合动力车控制参数优化效果上明显优于MLPS算法。

Multi-layer parameter scanning （MLPS） algorithm and simulated annealing particle swarm opti- mization （SAPSO） algorithm are used respectively to optimize the parameters of logic threshold control strategy for parallel hybrid electric vehicle （PHEV）. The PHEV with optimized parameters is simulated with TEST-CITY-HWY test procedure, and the results are compared with that before optimization. The results indicate that after optimiza- tion with MLPS algorithm, the fuel consumption and the emissions of HC and NOx reduce by 11.98%, 6.01% and 4. 03% respectively, but with an increase of 25.18% in CO emission; while the optimization with SAPSO algorithm leads to an all-round reduction in fuel consumption and the emissions of HC, CO and NOx respectively of 13.61%, 9.57%, 27.78% and 18.53%. In addition, SAPSO optimization also results in a slight increase of battery SOC compared with MLPS, showing the superiority of SAPSO over MLPS in respect of the effects of control parameter op- timization for PHEV.