Research On Intelligent Optimization Of Concrete Mixture Driven By Big Data

In this paper,we present a model based on support vector machine(SVM)and artificial bee colony algorithm(ABC)to optimization of concrete mix proportion,which enables us to save engineering cost and reduce the workload of mix proportion test.We present a prediction model of concrete performance based on SVM.Combined with this,the optimization model of mix proportion is established with ABC.Then,experiments on the tests of concrete strength and slump were made to verify the prediction results of SVM-ABC model.1.Establish the database of concrete mix proportion testTaking the amount of various raw materials and their important performance indexes of concrete as input variables and concrete performance indexes(concrete strength,slump,etc.)as output variables,the mix proportion database is constructed.According to the design mix proportion target,a database of mix ratios is constructed for further discussion and research.2.Study the application of support vector machine(SVM)in concrete performance predictionSVM is used to build a concrete performance model,and the 7-day strength,28 day strength and initial slump of concrete are fitted and predicted by given raw material dosage and main properties of materials.This method can simplify the high-dimensional complex calculation between raw materials and concrete performance,quickly establish the corresponding functional relationship,and reduce the design cycle and workload of concrete.3.Study the application of artificial bee colony algorithm in concrete mix proportion optimizationBee colony can complete complex overall tasks through the cooperation between individuals.The mix proportion optimization model is based on the performance model combined with ABC algorithm.The model can combine the constraints of concrete materials,design the target value,and optimize the more economic concrete mix proportion.4.Concrete block test is carried out to verify the applicability and reliability of the modelThe prediction results of different mix proportion optimization models are compared and verified by experiments.The rationality and applicability of the optimization model are explored by comparing the data obtained from strength and slump tests of the test block.