Study on New Sampling Plans and Optimal Integration with Proactive Maintenance in Production Systems
Date of Graduation
Doctor of Philosophy in Engineering (PhD)
Second Committee Member
Third Committee Member
Maintenance, Production systems, Quality, Reliability, Sampling plans
Sampling plans are statistical process control (SPC) tools used mainly in production processes. They are employed to control processes by monitoring the quality of produced products and alerting for necessary adjustments or maintenance. Sampling is used when an undesirable change (shift) in a process is unobservable and needs time to discover. Basically, the shift occurs when an assignable cause affects the process. Wrong setups, defective raw materials, degraded components are examples of assignable causes. The assignable cause causes a variable (or attribute) quality characteristic to shift from the desired state to an undesired state.
The main concern of sampling is to observe a process shift quickly by signaling a true alarm, at which, maintenance is performed to restore the process to its normal operating conditions. While responsive maintenance is performed if a shift is detected, proactive maintenance such as age-replacement is integrated with the design of sampling. A sampling plan is designed economically or economically-statistically. An economical design does not assess the system performance, whereas the economic-statistical design includes constraints on system performance such as the average outgoing quality and the effective production rate.
The objective of this dissertation is to study sampling plans by attributes. Two studies are conducted in this dissertation. In the first study, a sampling model is developed for attribute inspection in a multistage system with multiple assignable causes that could propagate downstream. In the second study, an integrated model of sampling and maintenance with maintenance at the time of the false alarm is proposed.
Most of the sampling plans are designed based on the occurrence of one assignable cause. Therefore, a sampling plan that allows two assignable causes to occur is developed in the first study. A multistage serial system of two unreliable machines with one assignable cause that could occur on each machine is assumed where the joint occurrence of assignable causes propagates the process's shift to a higher value. As a result, the system state at any time is described by one in-control and three out-of-control states where the evolution from a state to another depends on the competencies between shifts. A stochastic methodology to model all competing scenarios is developed. This methodology forms a base that could be used if the number of machines and/or states increase.
In the second study, an integrated model of sampling and scheduled maintenance is proposed. In addition to the two opportunities for maintenance at the true alarm and scheduled maintenance, an additional opportunity for preventive maintenance at the time of a false alarm is suggested. Since a false alarm could occur at any sampling time, preventive maintenance is assumed to increase with time. The effectiveness of the proposed model is compared to the effectiveness of separate models of scheduled maintenance and sampling.
Inspired by the conducted studies, different topics of sampling and maintenance are proposed for future research. Two topics are suggested for integrating sampling with selective maintenance. The third topic is an extension of the first study where more than two shifts can occur simultaneously.
Obaidat, S. (2020). Study on New Sampling Plans and Optimal Integration with Proactive Maintenance in Production Systems. Graduate Theses and Dissertations Retrieved from https://scholarworks.uark.edu/etd/3736
Industrial Engineering Commons, Operational Research Commons, Technology and Innovation Commons