Prediction of Deep Geo-stress Based on Fusion Multiple Model
ZHANG Wendong
LV Shanshan
ZHANG Xingsen
ZHANG Weidong
Abstract：Aimed at the difficulty of deep geo-stress measurement due to the research methods and techniques,and the low accuracy of the existing prediction models,a prediction method based on fusion multiple model is proposed. This method con-structes multiple sub prediction models which are trained by randomly selected samples,then these sub models are integrated ac-cording to some optimization methods.Finally the voting decision results got by sub models were the final deep geo-stress prediction value. The measured data of a certain oil field is used to conduct experiments,and the proposed method is compared with the LS-SVM(least squares support vector machines),BP neural network and linear regression.Research results show that the accura-cy of the fusion multiple model method is better than the existing methods for the prediction of maximum and minimum horizontal stress,and the fitting result of the proposed method is better. So it can meet the needs of engineering and can be used to predict deep geo-stress in engineering.
Keywords：deep geo-stresssub modelfusion multiple model predictionmaximum(minimum)horizontal stresspredic-tion accuracy
Publication Date：2018-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 717-720,738 )