Analytic tool for identifying bottlenecks using Turning Point method

Abstract

Manufacturing industry has come a long way ahead since the first industrial revolution. The fourth industrial revolution, Industry 4.0 gives companies an opportunity to make better informed fact-based decisions. Identifying system bottleneck and improving them provides companies the opportunity to utilize their resources more efficiently and be more productive. Productivity and system throughput are two major Key Performance Indicators. It is thus crucial to have an efficient production system that delivers quality products and the desired throughput. In order to achieve this and meet the customer demands in a short lead time, SKF wants to identify the bottleneck machine in a FMC that causes a mismatch between the observed and the desired throughput. The purpose of the thesis is to improve the productivity/throughput of the production line by facilitating real-time decision making. The aim of the thesis is to demonstrate an approach towards bottleneck detection based on real-time data for a flexible manufacturing system and the Turning point bottleneck detection method was selected. It involves developing a template in Microsoft Excel by identifying important parameters from the data collected to identify the bottleneck machine. CRISP-DM methodology has been used for data mining purpose. Machine states (working, idle, breakdown, set-up) are a prerequisite for most of the real-time bottleneck detection methods. In this thesis work, an algorithm is developed to calculate machines states from the PLC signals (communication signal between the robot and the machines) which in turn is used to identify bottleneck using Turning point method. The data was cleaned, prepared and analyzed to successfully identify the bottleneck machine. The machine states calculated can be used with different bottleneck detection methods. Analysis is carried out on historical data but the algorithm can be modeled in real-time to identify shifting bottleneck in a FMC.