Diagnosis and prediction of failures in maintenance systems using fuzzy inference and Z-number method

Abstract

In this research, a timely diagnosis and prediction mechanism for drill failure are provided to improve the maintenance process in drilling through fuzzy inference systems. Failures and decisions are based on information and reliability as well, and that affects the quality of decision-making. We apply the potential of if-then rules and a new approach called Z-number that considers fuzzy constraints and reliability at the same time. Exerting Z-number in this research took maximum advantage of reducing uncertainty for predicting failures. Additionally, this research has a practical aspect in maintenance systems by using if-then rules that rely on Z-number. The proposed approach can cover the expert idea during drill operation time simultaneously. This approach also helps experts encounter ambiguous situations and formulate uncertainties. Experts or drill operators can consider key factors of drilling collapse along with the reliability of these factors. The proposed approach can be applied to a real-life situation of human inference with probability for the purpose of predicting failures during drilling. Hence, this method has excellent flexibility for implementation in various maintenance systems.