基于BP网络的自密实混凝土配合比设计

原材料性能的波动对自密实混凝土性能影响很大,把水泥强度、砂石含泥量、砂石细度模数、石子的最大粒径、石子针片状含量、粉煤灰细度和粉煤灰烧失量这些原材料性能参数作为BP网络的输入,以对应的优化配合比作为网络的输出,用网络结构描述它们之间的非线性关系。利用正交试验数据样本完成了网络的训练并进行检验,计算结果表明,利用正交试验数据样本训练的BP网络可以预测不同情况下的配合比,预测精度高,完全可以代替繁重的实验室配合比设计。应用该技术可实现自密实混凝土配合比的实时优化,对控制自密混凝土的生产质量具有重要意义。

MIX PROPORTION DESIGN OF SELF-COMPACTING CONCRETE BASED ON BP NETWORK

Taking the main parameters as the input of BP neural network,suchas raw materials including strength of cement,mud content and modulus of fineness of fine aggregate,nominal maximum size of aggregate,faller gills content of coarse aggregate,loss on ignition and fineness of fly ash,which directly affect the properties of self-compacting concrete,and the corresponding optimized mix proportion as the output of it,the nonlinear relation between them is expressed by the test data.The calculation results of the actual example indicate that the concrete mix proportion prediction through the pre-trained BP neural network used orthogonal test data can substitute for some waste-time and heavy laboratory tests.On the basis of this technology,the real time optimization of self-compacting concrete could come to realization,which has great effect on the quality control of manufacturing self-compacting concrete.