结构混凝土抗压强度预测RBF-ANN模型  

Study on Structure Concrete Compression Strength Prediction RBF-ANN Model

作　　者：高峰[1] 李津[1] GAO Feng LI Jin(School of Coal Engineering, Shanxi Datong University, Datong Shanxi, 03700)

机构地区：[1]山西大同大学煤炭工程学院,山西大同037003

出　　处：《山西大同大学学报（自然科学版）》2016年第4期57-59,共3页Journal of Shanxi Datong University(Natural Science Edition)

基　　金：山西省大学生创新创业训练项目[SDC2012274];大同市基础研究计划项目[201369];山西省软科学研究计划项目[2014041068-4]

摘　　要：为了达到对C20~C40范围内的结构混凝土抗压强度快速准确预测的目的,基于RBF-ANN模型基本原理及应用特点,在钻芯法、回弹法、超声波法及回弹-超声波综合法等大量室内和现场无损检测试验基础上,进一步应用Matlab2012b神经网络工具箱,建立了混凝土抗压强度RBF-ANN预测模型。该模型经充分训练后,应用于山西省重点建设工程项目的结构混凝土质量控制。工程实践表明,构建的RBF-ANN模型预测精度为4.4%,满足工程实际需要,具有较好的便捷性、经济性和准确性。To realize the prediction of the structure concrete compression strength rapidly and accurately within the scope of C20 ~C40, the basic principle of radial base function artificial neural network（RBF-ANN） and its application characteristics are described firstly. Based on a large number of nondestructive testing such as the core drilling method, rebound method, ultrasonic method and rebound^ultrasonic synthesis method, the RBF-ANN model has been built to predict the structure concrete compression strength which has the topological structure of 3-1-1, by use of the neural network toolbox（NNT） of common software Matlab2012 b. And the RBFANN model after fully training is used to control the quality of concrete structure in the Shanxi provincial key construction project. The result shows that the error of concrete compression strength forecasting value using the RBF-ANN model is 4.4%,meeting the needs of practical engineering. So, the RBF-ANN method has convenience, economy and accuracy.

关 键 词：混凝土抗压强度 RBF-ANN模型 预测 MATLAB神经网络工具箱8.0 

分 类 号：TU528[建筑科学—建筑技术科学]