发动机试车数据拟合算法的研究与实现

摘要

针对发动机虚拟测试系统的需要,对某型号发动机试车的14个测试参数采用最小二乘方法结合BP神经网络进行数据拟合.根据测试数据的特点进行分类研究.试验验证表明,所采用的方法结果准确,满足发动机虚拟测试系统的要求.%According to the needs of the virtual test system engine, 14 measuring parameters of an engine are fitted by using least square method combined with the BP neural network data fitting, considering of the characteristics of the measuring parameters classification. The experiment shows that these methods are accurate, and can meet the requirements of the engine virtual measurement system.