Author: Mehdi Dadfarnia
Date/Time: October 30TH, 2025 at 3:00 PM EST
Location: 1117 Clark Memorial Conference Room (limited seating) | Zoom(Code: 541012)
Committee Members:
- Dr. Jeffrey W. Herrmann, Chair
- Dr. Katrina M. Groth
- Dr. Mohammad Modarres
- Dr. Mark Fuge
- Dr. Bilal M. Ayyub, Dean’s Representative
Title of Dissertation: Approaches to Evaluate Condition Monitoring-Based Technologies for Manufacturing Maintenance and Risk Management
Abstract:
Emerging advancements in artificial intelligence (AI) and the internet of things (IoT) create new opportunities for improving manufacturing system performance. Maintenance plays a critical role in keeping manufacturing systems productive and operational, and advancements in AI and IoT have increased opportunities to deploy condition-based maintenance policies and condition monitoring systems (CMSs). However, factors that hinder the adoption of such tools and technologies in existing maintenance practices include the obfuscated internal logic of AI-embedded CMSs and the inadequate methods available to evaluate the manufacturing benefits and financial implications of CMS-enabled maintenance policies.
This dissertation addresses this challenge by exploring various approaches that can be used for evaluating condition monitoring-based technologies in manufacturing maintenance. This dissertation systematically identifies key items that would comprise any comprehensive evaluation method, then pinpoints the opportunities in computational approaches that could be used as part of an evaluation strategy. Computational approaches can be used to test CMSs and obtain performance measures relevant to CMS detection and diagnosis, maintenance work effectiveness, and manufacturing productivity and quality. Traceability between these different performance measures can reveal true CMS impacts to an asset and inform decisions on CMS investments.
This dissertation systematically surveys evaluation methods of monitoring-related technologies used in industrial assets. The iterative review process helps to construct a conceptual model of key items that should be considered in an evaluation process. It also yields insights into shortcomings of the existing evaluation methods and opportunities for incorporating analytical frameworks, discrete-event simulation, performance measures, and uncertainty analysis in both CMS-related performance and investment analysis.
This dissertation then investigates simulation-based approaches and approximate-analytical models to evaluate the impacts that CMSs have on manufacturing systems when integrated with maintenance workflows. These investigations characterize the strengths and limitations of each of these simulation-based approaches and approximate-analytical models in their ability to evaluate manufacturing performance measures for manufacturing systems that use CMS-enabled maintenance policies. This dissertation operationalizes the simulation-based approach by developing a proof-of-concept simulator and applying the simulator in a simulation study comparing maintenance policies and manufacturing configurations. These simulations demonstrate their ability to capture the interacting behaviors between machine health degradation, CMS detection, part quality, and maintenance repair work across various manufacturing and maintenance performance metrics. This dissertation also proposes a novel evaluation process as a framework that leverages computational tools and techniques, such as those proposed in this dissertation, which lead to conducting a comprehensive assessment of CMS impacts from manufacturing performance metrics to investment analysis.