Condition-based maintenance of CTCS on-board equipment using gamma-process degradation modeling
Abstract
The safe and punctual operation of high-speed trains increasingly relies on the reliability and availability of China Train Control System (CTCS) on-board equipment such as balise transmission modules (BTM) and on-board controllers. In current practice, maintenance of these devices is still dominated by fixed-interval inspections and time-based overhauls, which do not explicitly account for the actual degradation state and may lead to either excessive preventive actions or unexpected in-service failures. This paper proposes a gamma-process-based condition-based maintenance (CBM) framework for CTCS on-board equipment under periodic inspections. The monotone degradation of a representative on-board component is modelled as a homogeneous gamma process, and a CBM policy is defined by an inspection interval and a preventive replacement threshold. Maintenance downtime for both preventive and corrective actions is explicitly considered. For a given policy, the long-run maintenance cost rate and availability are evaluated by Monte Carlo simulation, and a scalar performance index called unit-cost availability, defined as the ratio between availability and cost rate, is adopted to compare alternative policies. A two-stage grid search is then used to identify near-optimal combinations of inspection interval and preventive threshold. Experiments based on a BTM-like component show that the proposed method achieves very high availability with a moderate cost rate, and that there exists a broad region of policy parameters that deliver almost the same unit-cost availability. These results indicate that the proposed framework provides an effective tool to support CBM decision making for CTCS on-board equipment.
Fields of Study & Subjects
Reliability and Maintenance Optimization
Transportation Systems and Safety
Power System Reliability and Maintenance
Publication details
- Journal / source
- International Conference on Big Data, Computational Intelligence, and Applications (BDCIA)
- Pages
- 138–138
- Publication date
- 24 February 2026
- Language
- EN