A Comparative Study to Improve the Efficiency of Forecasting Shelf-Life Limited Maintenance Spare Parts with 4 Techniques of Time Series Forecasting, a case study of a Medium-Sized Airline Company

Abstract

The objectives of this research were 1. To study and analyze the problems of the loss of expired spare parts from inventory parts with a limited useful life 2. To increase the efficiency in purchasing maintenance spare parts with a limited useful life, and 3. To find the most appropriate model for forecasts used for an analysis of inventory needs with uncertain demand. There is qualitative research where the population is maintenance parts with a limited useful life. Therefore, the comparative study was employed to improve the efficiency of forecasts of maintenance spare parts inventory with a limited useful life using 4 time series forecasting techniques comprising the Moving Average Method, Simple Exponential Smoothing Method, and Holt-Winter Method by studying and collecting with the Purposive Sampling method by the data regarding spare parts purchased and expiring in 2020 and then selecting 7 items including maintenance parts with codes EC3524A-B, V04-018, PR1440B1-2, PS870B1-2, PR1784B1-2, PR1826B1-2 and P0LY-KIT. And analyzed the researchers by collecting data from in-depth interviews with experts in the organization, collecting the history issue data from the organization program, and using the Minitab 21 Statistical Software to do the forecasting and calculate both error methods. After that, 30 values of actual disbursement data from 1 January 2021 to June 2023 were brought to make forecasts with the 4 techniques by comparing the accuracy of forecasts with MAD and MAPE methods.

The results revealed that requirements for maintenance parts are diverse and change with the maintenance needs of each aircraft. According to the nature of damage caused by operations in addition, there is a shortage of products from OEMs in some periods, causing product lead- times to vary. Therefore, past procurement operations are difficult to predict or plan for. causing a shortage of inventory. This concern has led to attempts to stockpile products in larger quantities than necessary until the inventory expires. To increase efficiency in procurement, Therefore, all 4-time series forecasting techniques are used to study and compare the forecasts. This is because each technique is suitable for different time series data. To find the right and effective one. The forecast using a 6-period Moving Average Method provided the best forecast result with the error value from the MAPE method of maintenance parts code PR1440B1-2 with the lowest at 22.869, and the error value from the MAD method of maintenance parts code P0LY-KIT which was closest to 0 at 1. Hence, the forecast using the Moving Average method is suitable for the application of forecasts of maintenance parts with a limited useful life and other items.