Интеллектуальные системы предупреждения осложнений для безопасного строительства скважин

Abstract

The results of many years of research on the prevention of complications in the construction of oil and gas wells using machine learning methods are presented in the article. The issues of creating prototypes of intelligent systems to prevent complications when drilling wells on land and offshore are considered. The purpose of the intelligent systems to prevent complications during well drilling is to warn the driller in advance about the possibility of a violation of the regular drilling regime. Intelligent systems for preventing complications during well construction help to increase the economic efficiency of drilling oil and gas wells. Large volumes of geodata from the stations of geological and technological measurements during drilling vary from units to hundreds of terabytes. Creation of the neural network modeling software components is aimed at revealing hidden and non-obvious patterns in the datasets, i.e. in the processed, labeled and structured information from the stations of geological and technological measurements in the tabular form. Hierarchical distributed data warehouse was formed containing real-time drilling data in WITSML format using a SQL server (Microsoft). The geodata preprocessing and loading module for the WITSML repository uses the Energistics Standards DevKit API and Energistic data objects to work with the geodata in the WITSML format. The accuracy of predicting drilling problems achieved with the help of the developed intelligent systems can significantly reduce unproductive time spent on eliminating stuck pipes, mud losses and gas, oil and water shows. Large-scale implementation of the intelligent systems to prevent complications in well drilling will ensure the achievement of a zero-carbon footprint in the environmentally friendly drilling of wells on the land and offshore.