CAJ | 학술논문

机车柴油机台架实验成本高、调试周期长、不确定因素多,硬件在环仿真实验能够在实验室环境下完成控制系统基本功能和控制策略的验证,成为保证台架实验顺利进行的一种有效手段.结合神经网络模型和物理模型的优点,机车柴油机硬件在环仿真实验的主体部分采用神经网络训练实验数据建立模型,采用物理模型扩展系统的对外接口,对柴油机系统进行稳态工况和过渡工况仿真实验.结果表明,系统能够正确反映实际机车柴油机的运行情况.
궤차시유궤태가실험성본고、조시주기장、불학정인소다,경건재배방진실험능구재실험실배경하완성공제계통기본공능화공제책략적험증,성위보증태가실험순리진행적일충유효수단.결합신경망락모형화물리모형적우점,궤차시유궤경건재배방진실험적주체부분채용신경망락훈련실험수거건립모형,채용물리모형확전계통적대외접구,대시유궤계통진행은태공황화과도공황방진실험.결과표명,계통능구정학반영실제궤차시유궤적운행정황.
In order to solve the problems such as high costs,long debugging periods and a large number of uncertain factors in the bench-test experiments of the locomotive diesel engine,we use the hardware-in-loop simulation to verify the basic function and control strategy of the electronic control unit. The debug process of ECU can be accomplished in the laboratory. Combining the neural network and physical models,we research the dynamic and static characters on the hardware platform of virtual instruments. The simulation results prove the hardware-in-loop system can reflect real diesel running correctly. This system reduces the uncertain factors in the bench-test.