论文部分内容阅读
温度是带钢热连轧过程中几个最重要的工艺参数之一,由于温度将直接影响到热轧轧制力,因此精确预报各道次,特别是精轧机组各机架的轧制温度,是保证厚度、板形及宽度数学模型命中率的关键。而精轧温度预报技术是热连轧的核心技术。由于传统的模型技术已经不能进一步提高精轧温度的预报精度,针对带钢热连轧精轧温度传统模型的固有缺陷,根据CMAC神经网络具有很强的泛化能力,并且误差收敛速度快的特点,提出了基于CMAC神经网络的热连轧精轧温度预报模型。运用实际生产数据对该网络进行了训练和测试。结果表明,该CMAC模型能准确、实时地预报精轧温度,实现了通过提高精轧温度预报精度来达到提高带钢终轧温度命中精度的目的。通过CMAC预报精轧温度方法与传统的经验模型预报相比,CMAC系统误差的响应速度快、稳定性好,此模型具有良好的在线应用前景。
Temperature is one of the most important process parameters in the hot strip rolling process. Since the temperature will directly affect the hot rolling force, the accuracy of each pass, especially the rolling temperature , Is to ensure that the thickness, shape and width of the mathematical model hit rate key. The finishing rolling temperature forecasting technology is the core technology of hot rolling. Because the traditional model technology can no longer further improve the prediction accuracy of finishing temperature, aiming at the inherent defects of the traditional model of finish rolling temperature of hot strip mill, according to CMAC neural network has strong generalization ability, and the error convergence speed is fast , Proposed a hot strip mill finishing temperature forecasting model based on CMAC neural network. The network was trained and tested using actual production data. The results show that the CMAC model can predict the finishing temperature accurately and in real time and achieve the goal of improving the precision of final rolling temperature by improving the precision of finishing temperature. Compared with the traditional empirical model prediction, the CMAC system error response speed is fast and the stability is good. The model has a good online application prospect.