ANALYTICAL AND EMPIRICAL MODELS FOR PROCESS TRANSFORMATION TOWARDS SMARTER MAINTENANCE
Open Access
- Author:
- Shin, Hyunjong
- Graduate Program:
- Industrial Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 26, 2021
- Committee Members:
- Vittaldas V Prabhu, Dissertation Advisor/Co-Advisor
Vittaldas V Prabhu, Committee Chair/Co-Chair
Ling Rothrock, Committee Member
Paul Marshal Griffin, Committee Member
Suvrat Saurabh Dhanorkar, Outside Member
Felisa Preciado Higgins, Special Member
Steven James Landry, Program Head/Chair - Keywords:
- Smart Maintenance
Artificial Intelligence
Machine Learning
System Interaction
Business Process
Business Process Transformation
Human Centered System - Abstract:
- It is widely believed that the world’s economies are on the cusp of a mega-trend in automation in terms of physical and cognitive tasks. This dissertation investigates analytical and empirical models for process transformation especially focused on maintenance. We have elected to focus on maintenance because associated processes represent significant opportunities in terms of productivity improvement. Furthermore, maintenance could be among the most challenging processes for automation because it relies heavily on human cognition and decision-making. This dissertation has the following three major parts: 1. An empirical study in a controlled laboratory setting to compare diagnosis tasks in maintenance using traditional fault-tree and a new artificial intelligence supported system. 2. Modeling maintenance time using negative hypergeometric distribution considering the experience effect of technicians and evaluation of the variability in the number of diagnosis attempts depending on diagnosis support systems 3. Modeling stochastic dynamics in business processes using Markov Chains with a focus on tacit processes, making it an attractive approach for investigating and reengineering maintenance processes In the first part of the dissertation, the effect of diagnosis support systems on the performance of a technician during maintenance activity is investigated by conducting a set of controlled lab experiments. These experiments reveal that human subjects took different amounts of time to complete a task depending on whether a fault-tree diagnosis support system or an artificial intelligence diagnosis support system was provided. However, there was no statistical difference in the workload perceived.