基于改进鲁棒最小二乘法的发动机试验数据处理

对发动机试验数据进行拟合分析过程中,自动去除试验数据中测试误差较大或者是错误数据点的计算方法进行了研究.通过对最小二乘法,局部权重最小二乘法,鲁棒性最小二乘法,MATLAB提供的拟合方法进行比较,提出了改进的鲁棒性最小二乘拟合方法.该方法用于发动机配气机构试验数据拟合后,能够自动去除试验数据中的离异点,而且得到的拟合曲线具有连续的一阶导数和二阶导数,为发动机的试验结果分析提供了一种简便可行的方法.

Application of Improved Robust Least-Square Method to Experimental Data Processing of Engine

The method for removing the serious-error and wrong data automatically was studied for the experimental data fitting of engine.Comparing the least-squares method(LSM),the partial weighted LSM,the robust LSM with the MATLAB smooth function,we proposed an improved robust least-square method.Being used in the data processing of the engine,the method can remove the outlier point automatically.The curve obtained has the continual first and second derivative,and provides a feasible method for the engine test resu...