Adaptive Mamdani fuzzy model for condition-based maintenance

Authors:
Ranganath Kothamasu;Samuel H. Huang
Affiliations:
Intelligent Systems Laboratory, Department of Mechanical, Industrial, and Nuclear Engineering, University of Cincinnati, Cincinnati, OH 45221, USA;Intelligent Systems Laboratory, Department of Mechanical, Industrial, and Nuclear Engineering, University of Cincinnati, Cincinnati, OH 45221, USA
Venue:
Fuzzy Sets and Systems
Year:
2007

Citing 4
Cited 2

Cognitive factors in user/expert-system interaction

Human Factors
HyFIS: adaptive neuro-fuzzy inference systems and their application to nonlinear dynamical systems

Neural Networks
Learning from Data: Concepts, Theory, and Methods

Learning from Data: Concepts, Theory, and Methods
A novel fuzzy classification solution for fault diagnosis

Journal of Intelligent & Fuzzy Systems: Applications in Engineering and Technology - Engineering applications of Computational Intelligence

A fuzzy neural network with fuzzy impact grades

Neurocomputing
A fuzzy inference system for pump failure diagnosis to improve maintenance process: The case of a petrochemical industry

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.21

Visualization

Abstract

Proper maintenance of equipment to prevent failures has become increasingly important. For manufacturing companies, it enables uninterrupted production to support lean manufacturing. For commercial carriers, it ensures the safety of passengers and crew members. Maintenance technology has progressed from time-based to condition-based. The idea of condition-based maintenance (CBM) is to monitor equipment using various sensors to enable real-time diagnosis of impending failures and prognosis of equipment health. The success of CBM hinges on the ability to develop accurate diagnosis/prognosis models. These models must be cognitive friendly for them to gain user acceptance, especially in safety critical applications. This paper presents a neuro-fuzzy modeling approach for CBM. The emphasis is on model comprehensibility so it can effectively serve as a decision-aid for domain experts. The comprehensibility of a neuro-fuzzy system usually deteriorates once rules are tuned. To solve this problem, Kullback-Leibler mean information is used to evaluate and refine tuned rules so they remain easily interpretable. The effectiveness of this modeling approach is demonstrated via a couple of real-world applications.