基于多数据源的连铸软测量模型校正方法研究

Research of Model Calibration Method for Heat Transfer Soft Sensing in Continuous Casting Based on Multiple Data Sources

针对铸坯凝固传热软测量模型的准确性问题,提出了基于多数据源的模型校正方法。首先在铸坯不同位置测量其表面温度,然后根据测量数据和凝固传热数学模型的计算数据将参数辨识问题转化为参数优化问题,采用自适应粒子群优化算法进行模型参数辨识。最后利用凝固坯壳厚度的测量数据验证传热模型的准确性,结果表明:校正后的凝固传热模型可以满足工程应用要求,可用于实际连铸机的计算分析。

Aiming at the heat transfer model＇ s accuracy in casting blank solidification process, a soft-sensing model calibration method based on multiple data sources was proposed, at first, it has the casting blank＇ s sur- face temperature measured at different locations; then, has the measured data and numerical solution of the heat transfer model for the solidification based to transform the parameter identification into parameter optimiza- tion, and has adaptive particle swam optimization （APSO） introduced to identify the convection heat transfer co- efficient ; finally, it has the casting blank＇ s shell thickness measured to verify accuracy of the heat transfer model. The results show that the calibrated model can satisfy engineering requirements.