Demand estimation of aerospace spare parts based on a Weibull model with stochastic hazard rate.
Date of Issue2011
School of Mechanical and Aerospace Engineering
Competition on Maintenance, Repair, and Overhaul (MRO) services in the aerospace industry is intense. It is important to keep operations as lean and efficient as possible, in order for companies to survive. However, the airline industry store billions worth of spare parts in their inventory due to the difficulty in estimating the demand and balancing the risks/costs. A good failure model can help in estimating the demand more accurately, however current methods in modelling do not consider stochastic hazard rate, and instead assume deterministic/constant hazard rate. An example of IDG situation is used to illustrate the current situation where failure models with constant hazard rates are ineffective and inefficient in representing the failure data of the IDG. A proposed failure model: mixed weibull (with stochastic hazard rate unknown scale parameter and gamma distributed shape parameter) is used to model the failure data, to achieve a more accurate representation of the actual operating environment. A simulation assisted Maximum Likelihood Estimation (MLE) method (with the help of statistical sampling and differential evolution algorithm) is used for parameters estimation to overcome the computational challenges. Simulated failure data is used for this experiment, and results have demonstrated that in situations where the operating environment is stochastic, the weibull model with stochastic hazard rate will perform better and is more accurate than the weibull model with constant hazard rate in estimating the demand for MRO spare parts.
Final Year Project (FYP)
Nanyang Technological University