Using machine learning to enhance operator performance

Abstract

Machine learning is a powerful tool to analyse very large datasets. Although machine learning has been used for many years in other areas, such a social media, its value to process industry has only recently been realised. Operators interact with control systems in the same way people interact with social media and, as such, many of the algorithms that have been developed for modelling human interaction are applicable to industrial process operations. Recently, control systems companies have been developing analytical tools to leverage the vast amount of data collected in control systems over many years. These tools enable operations to understand the efficiency of their processes and procedures, identify gaps in their standard operating procedures and measure operator capability. This analysis assists in the improvement of procedures, highlights areas where further training is required and identifies opportunities for procedure automation. This has led to considerable improvements in operation performance, resulting in improved production and reduced downtime. This paper describes what machine learning is, how it can be applied to operation performance, the benefits this provides and possible applications in other areas of the industry.