李晶，张捷

（华中科技大学 能源与动力工程学院，湖北 武汉  430074）

    摘要：对于甲醇裂解气掺烧发动机,甲醇替代率通过采用BP神经网络算法的控制系统进行控制,BP神经网络经过有限的甲醇替代率试验数据的训练,其预测和控制精度满足使用要求。利用Simulink基于模型设计工具建立了辅助控制系统模型,对所搭建的控制模型进行了模型在环(MIL)仿真,并对模型所生成的嵌入式代码与硬件相集成实现处理器在环(PIL)一致性验证,最终完成甲醇掺烧发动机辅助控制系统的功能开发,大幅降低开发时间。

    关键词：基于模型设计；甲醇裂解气；BP神经网络；自动代码生成

Abstract: For an engine which is power by dissociated methanol gas, the methanol substitution ratio was controlled by assistant control system that adopt the algorithm of BP neural network. The BP neural network was trained by limited experiment data, but the control precision can meet the requirement of the application. The model of assistant control system with BP neural network was established by Simulink. Then, the simulation of Model-In-The-Loop and the verification of Processor-In-The-Loop for the control model was carried out respectively. Finally, the function development of the assistant control system was achieved and the development time was also reduced greatly.

Key words: Model-Based Design（MBD）； Dissociated Methanol Gas；BP Neural Network；Automatic  Code  Generation