A New Health State Assessment Method for Complex Systems Based on Approximate Belief Rule Base With Attribute Reliability

Assessing the health of complex systems is crucial for reducing failure risks and ensuring long-term stability and reliability. The Belief Rule Base (BRB) method effectively evaluates systems by combining observational data with expert knowledge. However, it faces challenges such as rule explosion, unreliable attributes, and uncertainty in expert knowledge. This paper proposes the Approximate Belief Rule Base with Attribute Reliability (ABRB-r) to address these issues. The ABRB-r model simplifies rule combinations by using a linear structure and considers attribute correlations to minimize interdependence effects. It incorporates attribute reliability into the inference process to enhance rule matching and dynamically adjusts activation weights to reflect expert knowledge uncertainty. A case study on diesel engines demonstrates the method’s effectiveness, significantly improving the accuracy and reliability of health state assessments.