Multi-state system reliability analysis combined PH distribution with universal generating function

• Published in: Chinese Control and Decision Conference pp. 2497 - 2501
• Main Authors: Peng Di, Tong Chen, Luan Wan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2017
• Subjects: Boilers; Continuous Phase-type Distribution; Exponential distribution; Markov processes; Multi-state System; Reliability; System performance; Time measurement; Turbines; Universal Generating Function
• ISSN: 1948-9447
• DOI: 10.1109/CCDC.2017.7978934

To make multi-state system reliability model reveal the randomness of component or system state changes more authentically, reduce the number of system states effectively and simplify the solution of model, this paper employs continuous phase type (PH) distribution to improve universal generating function. It is firstly assumed that a multi-state component's residence time in different states is subject to general distribution, which is presented in the PH distribution. A procedure for solving Instantaneous probability of each state is provided to construct the improved universal generating function for multi-state component. Subsequently, this paper describes how to construct the improved universal generating function for serial, parallel and complex systems, and obtains the reliability measures, such as the system instantaneous availability, average performance and mean performance deficiency measures. Finally, an example is used to verify the improvements of the model descriptive and state dimension control capability, and the algorithm effectiveness is illustrated demonstrated simultaneously.