【摘要】 同时利用稳态和动态校准数据,辨识了基于Hammerstein模型的热膜式空气质量流量(MAF)传感器的模型结构。在用多项式逼近静态非线性特性的基础上,动态线性环节分别选取带外生输入的自回归(ARX)模型、输出误差(OE)模型和Box-Jenkins(BJ)模型结构,采用交叉准则法进行参数估计和阶次选择,通过残差分析和仿真比较对模型进行检验。结果表明,用估计数据选择阶次时,最终预报误差(FPE)准则与最终输出误差(FOE)准则具有良好的一致性,基于预测误差法的2阶OE模型和BJ模型均可用于热膜式空气质量流量传感器Hammerstein模型动态线性环节的建模。更多还原

【Abstract】 The model structure of a hot film type mass air flow (MAF) sensor based on Hammerstein model was identified using steady state and dynamic calibration data of the sensor. On the basis of polynomial approaching of static non-linear characteristics of the sensor, the linear dynamic components of the model were chosen to be auto-regression model with exogenous variable (ARX), output error (OE) model and Box-Jenkins (BJ) model, respectively. The cross-criterion method was used in parameter estimation and order selection, and these models were verified using residual analysis and simulation comparison. Results show that final prediction error (FPE) criterion and final output error (FOE) criterion are in good agreement for selecting order with estimation data, both OE model and BJ model with two orders based on prediction error method are adequate for linear dynamic modeling of Hammerstein model for hot film type MAF sensor.更多还原