A Health Assessment Method for Satellite Components Based on a Two-Dimensional Baseline Model

Satellite component health parameters vary significantly across different types and environmental conditions, making it challenging to establish a reliable baseline model. In addition, dynamically adjusting parameter weights increases the complexity of the assessment process. To address these challenges, this article proposes a novel health assessment method for satellite components based on a 2-D baseline model. The baseline is initially established using the components' initial health data, forming a regression trend model and a probability density distribution model (PDDM) via a random forest regression algorithm and probability density distribution. This facilitates robust quantification of deviations in the component's operational state from the baseline state. To accurately capture performance variations under varying conditions, an adaptive weight adjustment mechanism that combines the criteria importance through intercriteria correlation method and the entropy weight method is introduced. This allows dynamic adjustment of parameter weights according to diverse environments and operating conditions. In addition, a weighted k-means clustering algorithm is utilized for a comprehensive analysis of health indicators, enabling the categorization of component health status and establishing a health assessment model. Experimental results demonstrate that this method effectively characterizes changes in component performance under diverse orbital environments, unknown anomalies, and performance degradation, highlighting its potential application value in engineering practice.