Preventive maintenance of a compressor station: A modeling framework for the assessment of performance
Résumé
Compressor stations are an essential part of natural gas networks, enabling the transportation of natural gas from producing wells to the final users. An efficient scheduling of maintenance tasks is key for ensuring high availability of the system and avoiding production losses and high costs from over expending. The demand of natural gas is highly dependent on weather and seasonal conditions, being higher during winter and lower during summer. Thus, the operation of the compressor stations alternates between periods of full capacity and periods of reduced capacity. The optimal planning of maintenance tasks needs to take into account such constraints from the production plans. Currently, preventive maintenance tasks are usually carried out periodically, following a calendar-based strategy. This maintenance strategy can be improved by moving to predictive maintenance in which the decisions are based on prognostics that consider both the deterioration of the system and the production profile. This work explores on the application of Piecewise-deterministic Markov Process (PDMP) as a framework for the optimization problem. The system level model captures the deterioration process of the units, the condition monitoring techniques, maintenance strategies and production profile in order to assess the performance of a maintenance strategy. Copyright