残差灰色预测模型在塑料老化行为预测中的应用  

Application of residual error grey prediction model in aging behavior prediction of plastic

作　　者：张天云[1] 陈奎[2] 胡玉霞[2] 郑小平[2] 

机构地区：[1]兰州城市学院信息网络中心,甘肃兰州730070 [2]兰州城市学院机械监测与故障诊断研究所,甘肃兰州730070

出　　处：《现代化工》2013年第5期119-121,123,共4页Modern Chemical Industry

基　　金：国家自然科学基金项目(11164012);教育部科学技术研究重点项目(210230);甘肃省自然科学基金项目(1010RJZA168);甘肃省教育厅基金项目(1111B-03)

摘　　要：针对数据的离散程度较大时,灰色GM(1,1)模型的预测精度较差这一问题,构建残差GM(1,1)预测模型,以湿热条件下聚丙烯力学性能变化的预测为例,研究所建模型在塑料老化行为预测中的适用性。结果表明,残差GM(1,1)预测模型得到的聚丙烯冲击强度与拉伸强度的模拟值与实际值的最大相对误差分别为0.43%和1.85%,明显优于灰色GM(1,1)模型得到的预测结果(最大相对误差分别为9.16%和4.25%),可用于塑料老化行为的预测。残差GM(1,1)模型所需实验数据少,预测精度高,为塑料老化行为的预测提供了一种简易而可靠的新途径。Due to the poor prediction accuracy of GM （ 1,1 ） model when the data are discrete, residual error GM （1,1） model is constructed according to Grey Theory. Taking the prediction of mechanical properties of PP under hygrothermal condition as an example, the applicability of constructed model in the prediction of plastic aging behavior is studied. Prediction results of residual error GM （ 1,1 ） model show that, the maximum relative error of simulation value and the actual value of impact strength and tensile strength of PP are O. 43% and 1.85% respectively,which are better than that （ the maximum relative error are 9. 16% and 4. 25% respectively） obtained from GM （ 1,1 ） model obviously. Residual error GM （ 1,1 ） model needs less original data and has a high prediction accuracy, which is a simple and reliable method for plastic aging behavior prediction.

关 键 词：残差 塑料 老化 预测 灰色理论 

分 类 号：O631.33[理学—高分子化学]