基于遗传算法优化支持向量机工况识别的燃料电池混合动力汽车能量管理策略

为了提高氢燃料电池混合动力汽车的燃料经济性,延长蓄电池寿命,选取中国重型商用车行驶工况-货车工况中3种典型工况代表"市区""市郊"和"高速公路",分别制定相应的最优能量管理策略;运用遗传算法优化支持向量机(gentic algorithm-support vector machine,GA-SVM)算法识别车辆运行工况,动态选择相应的能量管理策略,使其对选定的几种代表性工况具有自适应性,从而降低氢耗量,延长蓄电池寿命。仿真结果表明,与无工况识别的能量管理策略和采用传统算法优化的支持向量机(support vector machine, SVM)工况识别能量管理策略相比,使用GA-SVM工况识别的能量管理策略的等效氢耗量分别降低了7.78%和1.31%,蓄电池电池荷电状态(battery state of charge, SOC)变化量减小,变化相对平稳,有利于延长电池寿命。

Energy Management Strategy of Fuel Cell Hybrid Electric Vehicle Based on GA-SVM Condition Recognition

In order to improve fuel economy and prolong battery life of hydrogen fuel cell hybrid electric vehicle , three typical working conditions in "China heavy-duty commercial vehicle test cycle-truck" were selected to represent "urban", "suburban" and "expressway", respectively, and corresponding optimal energy management strategies were formulated; support vector machine optimized by genetic algorithm was used. GA-SVM algorithm identifies vehicle operating conditions and dynamically chooses corresponding energy management strategies to make them adaptive to selected representative conditions, thus reducing hydrogen consumption and prolonging battery life. The simulation results show that the equivalent hydrogen consumption of the energy management strategy based on GA-SVM was reduced by 7.78% and 1.31% respectively, compared with the energy management strategy based on condition-free identification and the energy management strategy based on SVM optimized by traditional algorithm, and the SOC change of storage battery was reduced, and the change was relatively stable, which is beneficial to prolonging battery life.