Data-Driven Operations Management and Production Optimization
Bill Tubbs, B. Tubbs & Associates Consulting; John Hedengren, Brigham Young University
In this talk I will present an overview of traditional operations management best practices that emphasise fact-based decision-making and a data-driven approach to optimizing large, complex, industrial production operations. These systematic approaches and best practices are nothing new yet they are as relevant today and arguably increasingly pertinent with the rapid increase in the amount of operational data and the availability of powerful new analytical tools and techniques (often described in abstract terms such as 'advanced process control', 'machine learning', 'digital twin', 'IoT', or 'Industry 4.0'). The purpose of the talk is to review fundamental operations management principles and traditional approaches to real-time control and dynamic process optimization in the context of recent developments such as machine learning approaches and present a clear, systematic approach to identifying and prioritizing new opportunities based on your current operational performance, your existing systems and practices, and the needs and capacity of your organization. I will draw on my experiences implementing production optimization best practices, production loss accounting systems and various process, cost, and energy optimization case studies. Based on these I conclude that there is significant opportunity for improvement in operational performance but that fundamental principles, basic, good management practices, and the continuing importance of the role of human operators, supervisors and engineers in the optimization process are as critical to success today as they were in the past.
Keywords:
Operational performance, production optimization, data-driven management, dynamic optimization, machine learning.
