Research On Application Of Deformation Monitoring In High-rise Buildings Based On GRNN Optimization Model

With rapid development of economy,a variety of high-rise buildings are constructed frequently in our country.The buildings are convenient to people’s life,but there are also bring some potential hidden dangers.The buildings are prone to deformation specially when constructed and used under the influence of some factors,such as temperature,stability of soil,hydrogeology,structure and weight of itself.When the limit of the building deformation is beyond,it may be exert influence on the safe of the buildings even the people’s life.So,it is very important to carry out the necessary deformation monitoring on the buildings and analyze and forecast the deformation monitoring data with reasonable analysis and forecasting model.It is difficult to reflect the deformation law of the buildings comprehensively of the the traditional single deformation analysis and prediction model on account of the complexity of the high-rise buildings and the much factors that cause the deformation.This paper is analyzed and compared the prediction effect of the model through the deformation combine model of the buildings with an example based on GM(1,1),Markov,Generalized Regression Neural Network(GRNN)and Fruit Fly Optimization Algorithm(FOA).The main research content of this paper are as follows:(1).The theory and the advantage and disadvantage of GM(1,1)and Markov are discussed.Then GM-Markov model is established to forecast analyze the deformation data that an engineering example by using the Markov model to improve optimization of GM(1,1)combine with the advantages of the two model.The combinatorial optimization model has a certain feasibility,and the prediction accuracy is obviously higher than the ordinary single model through the analysis and comparison of the forecast results.(2).The paper expound the basic theory of Generalized Regression Neural Network.And then establish GM-Markov-GRNN through discussed the method on the combination of Neural Network and Gray System Model by Generalized Regression Neural Network as an example.The model prediction accuracy is higher than the contrast significantly when using the deformation data of the same engineering example to forecast analysis because of the GM-Markov-GRNN included of three models advantages that GM(1,1),Markov,GRNN.(3).Using the improved algorithm to optimize the GM-Markov-GRNN model base on discussed the basic theory of Fruit Fly Optimization Algorithm(FOA)and its improved algorithm.After the study found that the smoothing factor has a effect on the prediction accuracy of the network in the Generalized Regression Neural Network significantly,but the Fruit Fly Optimization Algorithm has some weakness.So the common Fruit Fly Optimization Algorithm algorithm is improved by referring to the velocity variable of the Particle Swarm Optimization Algorithm(PSO).Then established VFOA optimized GM-Markov-GRNN model by using improved Fruit Fly Optimization Algorithm to optimize the GM-Markov-GRNN.At the same time,researched the feasibility and prediction accuracy of the combined model based on the analysis and prediction of deformation of high-rise buildings.The prediction accuracy of the VFOA optimized GM-Markov-GRNN model is higher than other models obviously,and the prediction effect is improved obviously.