Design Optimization Of Pervious Concrete Mix Proportion Based On BP Neural Network And M~2VA Method

As an important part of"sponge city"construction,pervious concrete can be widely used in parks,squares,sidewalks,urban road greenbelt,gymnasium,urban light traffic pavement and sewage purification.At present,there are many factors affecting the performance of pervious concrete,and the water retention and slow～release properties of ordinary pervious concrete are very poor,which can not effectively accumulate and make full use of natural precipitation,so its application and promotion are limited.In this paper,around the design of pervious concrete mix proportion,indoor test and numerical experiment are used to explore the mix optimization of pervious concrete based on BP neural network and M~2VA method.The optimal mix proportion of pervious concrete meeting the requirements of mechanical properties and permeability is studied.The water retention and water release properties of pervious concrete are preliminarily explored,so as to promote the popularization of pervious concrete And applications.(1)Based on the uniform test method,the experimental mix design of pervious concrete is carried out.Considering the five factors of aggregate size distribution,cement,water,VAE-vinyl acetate ethylene polymer emulsion,SAP-super absorbent resin,and 11 levels of variation,the compressive strength,flexural strength,porosity and permeability coefficient were determined.(2)According to the uniform test results,the L-M algorithm is used to train the test samples by MATLAB software,and the double hidden layer BP neural network of5-8-8-4 is established.Then the BP neural network simulation test is carried out to analyze the complex nonlinear mapping relationship between the performance index of pervious concrete and the influencing factors.Then,the optimum mix proportion of pervious concrete with good mechanical properties and permeability is determined as follows:water consumption is about 98.9l/m3,aggregate dosage is 1275kg/m3,aggregate gradation is ST1:ST2=3:1,cement dosage is 231.8kg/m3～259.6kg/m3,VAE content is about 9%～10%of cement dosage,sap is not mixed or less mixed.(3)Based on BP neural network analysis and optimization of pervious concrete mix proportion material dosage range,the performance test of improving mix proportion pervious concrete is carried out,and the optimization effect of pervious concrete mix design is verified.The results show that the optimization effect is good.Combined with the poor compressive strength performance of pervious concrete in the test,the visual analysis of the compressive strength of pervious concrete based on M~2VA method is further carried out to analyze the decisive factors and improvement ways of permeable concrete strength more intuitively.The order of the best mix proportion range to meet the requirements of road compressive strength and the factors influencing the compressive strength of pervious concrete from strong to weak are:water consumption>ST1>cement>VAE polymer>SAP polymer.(4)The effects of no polymer,single VAE,single SAP,double SAP and VAE polymer on water absorption,water retention and slow release of pervious concrete are preliminarily explored by means of simulated evaporation test.Based on the test results,relevant suggestions on the preparation of permeable concrete for"sponge city"construction are put forward.