Research On Valuation Of Prefabricated Concrete Buildings Based On BP Neural Network

Prefabricated concrete buildings have the advantages of energy saving and environmental protection,reducing on-site labor,shortening the construction period,improving the quality of the project,and promoting the reform of the supply side of construction products.Therefore,they have been vigorously promoted and developed in my country in recent years.Engineering cost estimation is an important basis for construction project feasibility studies,scheme comparison and selection,bid quotation,project financing,etc.However,the difficulty of controlling the cost of prefabricated concrete construction restricts its development,and it is beneficial to quickly evaluate the project Cost management can further provide guarantee for the development of prefabricated concrete buildings.At present,there are relatively few studies on the cost estimation models and methods of prefabricated concrete buildings in China.Because BP neural network has great advantages in forecasting,it can bring greater convenience to cost estimation,so BP neural network is used in The evaluation of prefabricated concrete buildings has certain practical significance.The thesis takes prefabricated concrete construction cost as the object,and uses BP neural network to establish prefabricated concrete construction evaluation model to realize the rapid evaluation of its cost.From the perspective of different construction methods of prefabricated concrete buildings and traditional cast-in-place buildings,the cost difference of the two is analyzed as the basis of the paper.Secondly,BP neural network is selected as the estimation tool,and the rationality of its application is described.Then use the analytic hierarchy process to establish the recursive model,select the most important engineering characteristic index,and further simplify the input vector of the valuation model.By setting the basic structure of BP neural network,initializing relevant parameters,analyzing program algorithm flow,etc.,the programming of the assembled concrete building valuation model is realized.And input the normalized sample data into the network for training,and finally use example simulation analysis to verify its accuracy.Through the above research,the following conclusions are obtained:(1)Based on the analytic hierarchy process,the selection of engineering characteristic indexes is realized.Eleven types of important engineering feature indexes of fabricated concrete buildings were selected as input vectors of BP neuralnetwork,which achieved the effect of simplifying the valuation model.(2)An evaluation model of prefabricated concrete buildings based on BP neural network was established.Using MATLAB to achieve program programming: set the network structure and related parameters.And after the sample training,the network can have better convergence effect.(3)The accuracy of the appraisal model is verified,and the cost of prefabricated concrete construction is evaluated efficiently and quickly.The example analysis shows that the average error of the unilateral cost of assembled concrete buildings can be controlled within 3%,which meets the requirements of my country’s investment estimation and achieves the ideal estimation effect.