基于共轭梯度法辨识钢渣温度分布

Identification Temperature Distribution of Steel Slag Based on Conjugate Gradient Method

为提高钢渣回收效率,需要对渣罐内钢渣温度分布进行精确测量,传统的热电偶测温成本高且效率低下,而辐射测温无法满足精度要求。本文运用共轭梯度法(CGM),以局部测量壁温为输入值,对渣罐内钢渣的温度分布进行反演,并进行仿真实验,探讨初始值、测点数、测量偏差与测点位置等因素对反演结果的扰动效果。结果表明:初始猜测值越偏离于真实值,反演精度就越差,反演精度随着测点数目的增多而有所提高;测量标准差σ在相对的测量误差范围内对反演精度影响较小,但当误差进一步增大时,反演精度随之降低。当σ≤1.0℃时,共轭梯度法能获得符合实际的反演结果;测点位置对反演精度的干扰比较显著,反演结果因测点位置的不同而表现出较大差异。

In order to improve the recovery efficiency of steel slag,it is necessary to accurately measure the temperature distribution of steel slag in the slag tank.The traditional thermocouple temperature measurement has high cost and low efficiency,while the radiation temperature measurement cannot meet the accuracy requirements.In this paper,the conjugate gradient method(CGM)is used to reverse the temperature distribution of the steel slag in the slag tank and explore the perturbation effect of the initial value,measurement number,measurement deviation and the inversion position.The results show that the more the initial guess value deviates from the real value,the worse the inversion accuracy increases with the number of measurement points;The measurement standard deviation has less impact on the inversion accuracy within the relative measurement error,but when the error increases further,the inversion accuracy decreases.At that time,the conjugate gradient method could obtain the actual inversion results;The inversion accuracy and the inversion results showed the measurement position.