| Summary: | 碩士 === 國立成功大學 === 交通管理學系碩博士班 === 94 === Airlines’ reliability control program is aiming at discovering root causes of poor aircraft system or component reliability through monitoring their performances and execute appropriate corrective actions involving changes of aircraft maintenance program(AMP), modification of aircraft system or component, and relevant training requirements to ensure the reliability performance of aircraft system or component are within the predetermined and acceptable level.
As for aircraft system reliability, it is the monthly technical delay and cancellation rate that is used to monitor each aircraft systems’ performance respectively. For those aircraft systems whose three-month rolling average delay rate exceed the predetermined upper control limit(UCL), which will generate “Alert”, engineers should conduct investigations to find out root causes and propose corrective actions to restore the system performance to an acceptable level.
Therefore, the goal of this research is to analyze real technical delay and cancellation data through different approaches including analysis of variance(ANOVA) and logistic regression modeling to discuss the possible deficiencies of setting up UCL. At the same time, this research also proposes three different methods of setting up the UCL, and further applies the methodology of system simulation to conduct the experimental test of those proposed settings.
The result of this research recommends that in considering the successful alert ability, false alert conditions, and flight delay and investigation cost, it is better to conduct investigations when continued alert signal appeared, and use the upper limit of 90% confidence interval of previous year’s average 3-month rate to set up next year’s UCL and compare it with the 3-month rate of the next year. On the other hand, if the cost of each investigation is 40% higher than each flight delay cost, it is suggested that setting up the UCL with the upper limit of 95% confidence interval of the previous year’s average monthly rate, and compare it to next year’s 3-month rolling average rate will be more suitable.
The main contribution of this research is to provide a guideline of adjusting UCL. However, exactly which method to apply is deeply depending on each airlines policy and the difference of its flight delay and investigation cost.
|
|---|
