矿井主要通风机性能曲线的最优多项式拟合

Optimal polynomial fitting for mine main fan characteristic curves

为确定主要通风机性能曲线多项式拟合函数的最优阶数,将回归系数的显著性作为最优阶数的判定依据。分析了基于最小二乘法的任意阶数主扇性能曲线拟合模型;指出了回归多项式的各项系数应与风压显著相关才是合理的,在合理基础之上阶数最大的即为最优;采用F检验法计算了回归多项式各系数的显著性。对于有m个样点的测定数据,通过m-1次拟合,可确定多项式的最优阶数。编制了相应的程序,实现了最佳阶数回归函数求解与曲线可视化,结合一包含13个测点的实例显示了这种方法的优越性。

For determinating the optimal power of polynomial fitting function of main fan＇s characteristic curve,the significance of regression coefficient was taken as the criterion of optimum power.Based on the least square method,the fitting model of main fan＇s characteristic curve with arbitrary power number was analyzed.The reasonable and optimal power of the regression functe were pointed out.F test was used to calculation significance of each coefficient.Where the count of sample data is n,fitting times of n-1 was needed to calculate the optimum power.Program was coded to calculate the regression function and to visualize the curve.Finally,one example including thirteen group of sample datas was employed to display the advantage of this method.