非参数投影寻踪回归在粉煤灰混凝土强度预测中的应用

Nonoarametric projection pursuit regression and its application of strength prediction about fly ash concrete

为了进一步优化粉煤灰混凝土配合比设计,改进混凝土施工工艺,提高混凝土质量。针对粉煤灰混凝土强度与其影响因素之间复杂的非线性关系,通过对影响粉煤灰混凝土强度的主要因素进行分析,提出了基于投影寻踪回归分析的粉煤灰混凝土强度预测模型,用单纯形优化算法确定最佳投影方向,并在Matlab环境下编写了相应的计算程序。利用某工程粉煤灰混凝土配合比的实测数据,进行粉煤灰混凝土强度的预测。实验结果表明,该方法具有较高的拟合度和良好的预测精度,满足工程实际要求。

In order to further optimize mix proportion design of fly ash concrete, improve the construction technology and the quality of concrete, according to the complex non-linear relationship between fly ash concrete strength and its influencing factors and analyzing the main factors of affect the strength of fly ash concrete, a prediction model based on the projection pursuit regression analysis for the strength of fly ash concrete is proposed, the best projection direction is found using the simplex optimization algorithm, and the corresponding calculation program is compiled under the Matlab.The strength of fly ash concrete will be predicted by the use of relevant data measured.It shows that this method has a high fit, good prediction accuracy, and applicable to practical engineering.