Availability-Based Design Optimization Using Markov Chain and a Fuzzy Goal Programming Approach

Abstract

Designing system configurations is essential for ensuring the safety and efficiency, particularly in strategic sectors such as aerospace, and energy. Optimizing system configurations involves maximizing system availability while minimizing overall costs. This study focuses on the development of a feed water system (FWS) for heat-recovery steam-generator boilers used in combined cycle power plants in Iran. The process includes sourcing components from various suppliers, each offering different part characteristics and pricing, followed by installation at the power plant and subsequent utilization. The project is constrained by a deadline for installation and construction, with penalties incurred for project completion beyond this timeframe.This paper introduces a novel binary nonlinear fuzzy goal programming (FGP) model designed for the selection of part suppliers for multistate parallel-series industrial systems, considering both availability and manufacturing costs. The model incorporates factors such as quantity-based discounts, component purchase costs, and penalties for delays in system construction. Additionally, a fuzzy target programming approach is used to minimize deviations from the target expense goals. A system reliability block diagram is employed to represent the system's operational states—whether it operates at full capacity, partial capacity, or is non-operational. The Markov chain model is utilized to depict the sequence of possible states, where each state's probability is dependent on the reliability of the system components. The model also accounts for the impact of ordering multiple parts from the same supplier, which influences both economies of scale and delivery lead times.Results demonstrate the practical applicability of this approach in optimizing system reliability.