Thesis project – Modeling Multiple Failures in a Multi-Indenture Inventory System
Modeling Multiple Failures in a Multi-Indenture Inventory System, thesis project by Systecon and KTH.
In this master thesis the problem studied is multiple failures in a multi-indenture inventory system. Multiple failures mean that at least two items breakdown at the same time, so that the customer,the system, demands the items simultaneously and they are replaced in parallel.
The spare parts optimization tool, OPUS10, which is evaluated in this report, assumes a serial exchange in the event of multiple faults. This results in overestimating the spare parts waiting time, because in reality the maximum waiting time governs serial exchanges. This conservative approach impacts spare parts optimization and leads to recommending unnecessarily large inventories to obtain the desired availability of the technical system.
One perspective that is being tested is based on the idea that clients arrive at the system and "order" one or more units. The probability distribution of these arrivals is then used to identify the expected waiting time, which is then used in Little's Law to find the average number of back orders. This reverse approach turns out to be very useful and effective in a multi-component system when using an order placement perspective.
The results show that OPUS10, as predicted, overestimates the contribution by the expected number of sub unit back orders, resulting in overestimating the number of main unit back orders. The best estimate of the contribution of the sub unit back orders is obtained using the order placement perspective by using the exact Song method, and the second-best result is obtained using the approximate Song method.
The full text of the thesis project can be downloaded here, the file name is "Modeling Multiple Failures in a Multi-Indenture Inventory System"
Title: Modeling Multiple Failures in a Multi-Indenture Inventory System
Thesis project S190 in Systems Engineering,
March 2010
Author: Kristina Abelin
Advisor: Thord Righard
Examiner: Per Enqvist