Date of Award


Publication Type


Degree Name



Industrial and Manufacturing Systems Engineering


Fault detection diagnosis and recovery, FDDR, Resilience assessment


Hoda ElMaraghy



Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


Failures are unwanted but also inevitable part of any manufacturing system. If not handled properly, they will lead to machine breakdown and line stoppage, which can cause a loss of production time and increased operating costs. One of the causes of failures is unattended faults. Faults are any unwanted deviations that can occur in any component of the system, mechanical parts, software, sensors, actuators, etc. To prevent that, manufactures have taken fault detection and diagnosis measures, from simpler knowledge and model-based techniques to more complicated traditional machine learning since the 2010s and to present-day deep learning theories. Another concept that helps manufacturers to continue manufacturing efficiently at a reasonable cost and production time is resilience, which is a system’s capability to withstand and accommodate disruptions and recover back to an acceptable state as soon as possible. Even though resilience is a widely studied concept against sudden high-impact disruptions of the supply chain, it is yet to be studied against fault and lighter disruptions. A disruptive event, such as machine or station failure, will lead to full or partial loss of production in the manufacturing system. Therefore, gaining understanding and evaluation of disruptive events and their impacts will have significant impact on the economic sustainability of the manufacturing system. This research will focus on linking the concept of fault detection and diagnosis with fault recovery and resilience of manufacturing systems. To obtain the results, different types of faults are developed in a manufacturing system, and the time taken to handle (detect, diagnose, and recover) each fault is presumed. This means that the author utilized both empirical research findings from literature and case studies, as well as some conjectural data due to a scarcity of available information. Next, with the help of simulation software, the manufacturing system is modeled, and the impact of each type of fault on the system and its resiliency is studied. Finally, with quantitative measures, the system’s resilience is calculated.

Available for download on Saturday, January 25, 2025