基于BP_RBF神经网络混凝土抗压强度预测_李红.pdfVIP

31 8 武 汉 理 工 大 学 学 报 Vol . 31 No. 8 2009 4 JOURNAL OF WUHAN UNIVERSITY OF TECHNOLOGY Apr. 2009 DOI: 10. 3963j . issn. 1671-4431. 2009. 08. 009 BPRBF 1 2 李 红 , 彭 涛 1. , 710061; 2. ; 450008) : 为了预测混凝土的抗压强度, 在分析 BPRBF 神经网络原理的基础上, 提出用BPRBF 神经网络模拟混凝土抗 压强度与搅 机各主要影响参数间关系的方法根据搅 机的实际工作状况, 分别建立了4 维输入向量1 维输出向量 的BPRBF 神经网络模型, 通过 19 组试验, 验证了2 种模型的可靠性结果表明, 实测结果与预测结果相接近, 该2 种神 经网络模型能较准确地快速预测混凝土抗压强度 : BP 神经网络; RBF 神经网络; 预测; 混凝土; 抗压强度 : TP 31 TU 528. 1 : A : 1671- 4431 2009) 08-0033- 04 Prediction of Concrete Compression Stren th Based on BP and PBF Neural Network Theories 1 2 LI Hong , PEN G Tao 1. School of Architectural Engineering, Changan University, Xian 71006 1, China; 2. Henan Sunrise Architectural Design Institnte, Zhengzhou 450008, China) Abstract: For prediction of concrete compression strength , the method for simulating the relation betw een concrete compression strength and major influencing factors of stirring mill using BP and RBF neural network is presented. According to the practical working situation of stirring mill, four- dimensional output vector BP and one- dimensional output vector RBF neural netw ork are established. The reliability of the tw o models is proved by 19 tests . The result indicates that the test data is closed to the prediction , these tw o neural net- w ork models is a preferable method for predicting concrete compression strength precisely and quickly. Key words: BP neural network; RBF neural network; predi