Design of an Infological Model for Prediction of Problems during Drilling Using Machine Learning and Systems Analysis

Abstract

Abstract The article deals with the actual problem of identification of probabilistic processes as a result of the operation of drilling rigs in the oil industry. The world experience of finding ways to solve optimal forecasting tools using machine learning is summarized. Mnemonic rule for the implementation of classification and ranking systems in the detection of feedbacks as probable indicators of complications of ongoing technological processes is implemented through the description of the formal model of the drilling process in form of a hidden Markov model. The results of evaluation of the developed mathematical apparatus in the form of predictive analytics and a cut of basic complications in the drilling process are presented. An infological diagram of the developed architectural solution of the analysis project is proposed. The results of the control algorithms formalization are given in conclusion. These results allow to ensure the effective procees modes of equipment operation and make it possible to save electricity and water.