Developing a fuzzy logic-based risk assessment for groundwater contamination from well integrity failure during hydraulic fracturing

Recent natural gas development by means of hydraulic fracturing requires a detailed risk analysis to eliminate or mitigate damage to the natural environment. Such geo-energy related subsurface activities involve complex engineering processes and uncertain data, making comprehensive, quantitative risk assessments a challenge to develop. This research seeks to develop a risk framework utilising data for quantitative numerical analysis and expert knowledge for qualitative analysis in the form of fuzzy logic, focusing on hydraulically fractured wells during the well stimulation stage applied to scenarios in the UK and Canada. New fault trees are developed for assessing cement failure in the vertical and horizontal directions, resulting in probabilities of failure of 3.42% and 0.84%, respectively. An overall probability of migration to groundwater during the well injection stage was determined as 0.0006%, compared with a Canadian case study which considered 0.13% of wells failed during any stage of the wells life cycle. It incorporates various data types to represent the complexity of hydraulic fracturing, encouraging a more complete and accurate analysis of risk failures which engineers can directly apply to old and new hydraulic fracturing sites without the necessity for extensive historic and probabilistic data This framework can be extended to assess risk across all stages of well development, which would lead to a gap in the modelled and actual probabilities narrowing. The framework developed has relevance to other geo-energy related subsurface activities such as CO2 sequestration, geothermal, and waste fluid injection disposal. (C) 2021 The Authors. Published by Elsevier B.V.

Automated summary

The research aims to develop a risk framework using data and expert knowledge for quantitative and qualitative analysis of risk for hydraulically fractured wells in the UK and Canada. New fault trees were developed to determine probabilities of cement failure in vertical and horizontal directions, as well as the probability of groundwater migration during well injection stage. The framework can be extended to assess risk across all stages of well development, narrowing the gap between modeled and actual probabilities.