Safety evaluation of casing string based on BP artificial neural network

2018 International Joint Conference on Materials Science and Mechanical Engineering

Yan, Yifei^1 ; Yang, Xiujuan^2 ; Kong, Chao^1 ; Zhou, Xiaoqi^3 ; Shao, Bing^1

College of Mechanical and Electronic Engineering, China University of Petroleum, Qingdao

266580, China^1

College of Pipeline and Civil Engineering, China University of Petroleum, Qingdao

266580, China^2

Petroleum Engineering Technology Research Institute of Zhongyuan Oilfield, Sinopec

457001, China^3

关键词: BP artificial neural network;    BP neural networks;    Casing load;    Casing strings;    Complex well;    Oil and gas well;    Safety assessments;    Safety evaluations;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/383/1/012003/pdf
DOI  :  10.1088/1757-899X/383/1/012003

来源: IOP

【 摘 要 】

In the traditional safety assessment of casing string, some influencing factors are ignored for modelling convenience, which makes the casing string safety assessment effect of oil and gas well not very ideal. For complexity and randomness of casing load and its properties parameters in complex well conditions, BP artificial neural network is created in MATLAB based on the analysis of the influencing factors of casing string security. Casing string section whose safety assessment is more mature is taken as a sample to train the BP neural network. The trained network is applied to make case assessment. At the same time, GUI interface is applied to implement the visualization. The results show that safety evaluation of the casing string can be achieved by using BP neural network. The accuracy of the casing string network safety evaluation is high. It will realize the visualization of safety evaluation and provide more accurate and effective reference for the design of casing string.