影响温度外推测试精度的数值因素

针对某些工况中温度不易测量的问题，提出了温度测试的外推方法，该方法以易测位置处温度测试为起点，采用传热规律的模型以及数值解决，反推出要测位置处的温度。在数值求解的过程中，引入了影响外推测试结果精度的数值因素－－阶数，并通过实例讨论了阶数对外推测试结果的影响程度以及获得满意精度时阶数的选取标准。分析结果表明，阶数决定着外推测试的结果以及结果的精度，阶数选取不当甚至会出现错误的外推测试结果，但采用反面求解求得的拟合曲线与实测得的原值具有最小的剩余误差平方和，可以有效地抑制其影响。对于表面特定温度为θ（0,t)的曲线，测得其0.0129cm处的温度变化规律反推，选取阶数为10时，剩余误差平方和为533％，外推结果错误，选取阶数为9时，剩余误差平方和为0．673％，外推结果正确。

Numerical Factor Affecting Measuring Precision of Temperature Extrapolation Method

To obtain temperature in some operating conditions, a temperature extrapolation method is presented. The method is based on easily measured temperature, and the required temperature is derived with the model of heat transfer and numerical solution. A numerical factor, the order number, is introduced in the process of numerical solution. The affecting degree of order number measuring precision is discussed through a case, and a method to decrease the error is given. It is concluded that the order number determines the precision of temperature extrapolation method, and it can be controlled when the quadratic sum of residual error between the opposite solving and the original measuring is minimum. For the specific surface temperature curve, it is derived with the temperature on the site 0 012 9 cm. Based on the order number 10, the quadratic sum of residual error is 533%, the extrapolating result is wrong. But with the order number 9, the quadratic sum of residual error is 0 673%, the result is right.