论文部分内容阅读
针对模压时效炉锻件温度难以直接测量的问题,建立了基于混合核偏最小二乘(KPLS)算法的模压时效炉锻件温度软测量模型,通过采集较易获得的模压时效炉工作室炉壁温度估计锻件的实际温度.并采用局部加权算法确定训练样本权值,以提高软测量模型的精确度.实验结果表明,所建局部加权混合核偏最小二乘的软测量(LWKPLS)模型具有较好的数据适应性且能够满足实际温度预测的精度要求,解决了模压时效炉锻件温度难以直接测量导致铝合金产品欠温或过烧而带来的质量问题.
Aiming at the problem that it is difficult to directly measure the temperature of forging aging furnace, a temperature soft measurement model of the mold forging aging based on Kernel Partial Least Squares (KPLS) algorithm is established. By measuring the temperature of the furnace wall, The actual temperature of the forging and the local weighted algorithm to determine the weight of the training samples to improve the accuracy of the soft measurement model.The experimental results show that the LWKPLS model with partially weighted mixed kernel partial least squares Data adaptability and can meet the actual temperature prediction accuracy requirements to solve the mold aging furnace forging temperature is difficult to directly measure lead to aluminum alloy products caused by the temperature or over-burning of the quality problems.