Comparing Artificial Neural Network and Failure Distribution Methods for Maintenance Scheduling: A Case Study of Wooden Door Industry

Maintenance plays a significant role to keep the machine in a good condition. Poor scheduling of machine maintenance can have unforeseen impacts and this will make the production process stopped. If this continues, losses due to stoppage of the machine can be a great concern to the company. This research focused on comparing Artificial Neural Network (ANN) and other failure rate distributions to determine the appropriate maintenance interval. The more reliable a machine the less likely it is to break down. The concept of reliability analysis uses statistical methods in calculating the probability of failure rate of a component over time. The data used are Time Between Failure (TBF) and Time To Repair (TTR) data from the machine's historical data of 2015-2017. The result shows that ANN gives the least error than other distribution models.