.png){kind=icon}
.png){kind=icon}
Return
Impact-Based Maintenance Efficiency Modeling for Multi-Component Systems With Condition-Based Component Selection
.png){kind=icon}
PRE
AI
DOI：10.1109/ACCESS.2026.3664001.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper develops an impact-based maintenance efficiency model that dynamically links maintenance effectiveness to the number and criticality of replacement components in multi-component repairable systems. Unlike existing approaches that assume static efficiency parameters, the model enables maintenance decisions that adapt to actual system degradation states. The efficiency formulation decomposes into a risk fraction capturing the proportion of weighted system degradation addressed, and a coverage factor with a diminishing-returns parameter reflecting the criticality distribution of maintained components. A condition-based selection rule achieves natural component rotation without explicit constraints, ensuring an equitable distribution of maintenance efforts while prioritizing high-criticality degraded components. Statistical validation using maximum likelihood estimation demonstrates parameter recovery within 2-8% of true values, while predictive performance evaluation shows 93% reduction in failure prediction error compared to approaches ignoring maintenance effects. A case study on a heterogeneous 20-component industrial system demonstrates that the proposed condition-based approach achieves a 30% cost reduction versus a no-maintenance baseline and a 5% improvement over fixed-selection strategies. Furthermore, the natural rotation mechanism maintains 65% of system components across maintenance epochs.
Keywords:
Maintenance
Maintenance engineering
Costs
Degradation
Reliability
Mathematical models
Arithmetic
Preventive maintenance
Optimization
Maximum likelihood estimation
Imperfect maintenance
virtual age model
condition-based maintenance
component selection
maintenance optimization
multi-component systems
Kijima model
reliability engineering
Journal
IF:
3.6
Papers:
9.7W
Citations:
29.4W
View more organizations