Research On Strength Prediction Of Solid Waste Cementitious Material Concrete Based On CGA-LSO-BP Neural Network

While the booming industry enhance economic development,the pollution of industrial solid waste discharge also follows.The preparation of solid waste cementitious material concrete based on industrial solid wastes can alleviate the environmental and resource problems and benefit the sustainable development of China’s industrial system.Considering that the compressive strength is the key character of concrete,the hydration reaction of the new type solid waste cementitious material is not quite clear,and its strength is depends on various factors.Therefore,in this thesis,the optimized neural network is used to forecast the compressive strength of concrete instead of the traditional method.The influence of a series factors on the strength of solid waste cementitious material concrete at different ages is obtained through the analysis test of influencing factors.The input variables of the forecasting model are the amount of cement,fly ash,water,etc.The study sample of forecasting model is formed by using solid waste binding material to partially replace cement to prepare concrete and carrying out compressive strength test.The total number of samples is 240,including 192 for training model and 48 for testing model.For the purpose of improving the prediction accuracy and generalization ability of the bp neural network(BPNN),the chaotic map is used to enhance the ergodicity of the genetic algorithm in the search of solution space,and the lion swarm optimization keep further refining the initial weight and threshold on the basis of the chaotic genetic algorithm,so as to avoid the algorithm getting stuck in local optimum.Subsequently,the BPNN prediction model optimized by the hybrid algorithm is applied to the forecasting task of the solid waste cementitious material concrete strength to improve the prediction accuracy of the solid waste cementitious material concrete compressive strength.The prediction result of the improved prediction model and the ordinary BPNN model shows that the prediction data of the improved model is in accordance with the real value,and also with better prediction performance.The above results provide new ideas for the prediction of the compressive strength of solid waste cementitious material concrete,provide theoretical guidance for actual production,decrease the resources consumed in the early trial and late test of the production of solid waste cementitious material concrete,and decrease the resource consumption caused by non-repeatable testing.