Room: Stars at Night Ballroom 2-3
Purpose: Risk assessment for the acceptance testing and commissioning (ATC) of a Halcyon linac was performed based on a conventional FMEA process in our previous studyÂ¹. However, the risk priority number (RPN) that was generated with equal weightings from the three factors, severity (S), occurrence (O), and detection (D), contains limitations. In this study, we re-evaluated the potential failure modes (FM) based on the improved RPN generated with a fuzzy logic inference system, and compared the final FM ranking to those obtained from the traditional FMEA.
Methods: Gaussian membership functions of 4 levels (low, medium, high, and very high) for each risk factor were generated considering their relative importance. A set of fuzzy if-then rules derived from clinical evaluators were used to relate S, O, D membership function to the fuzzy RPN membership function. A centroid defuzzification method was then chosen to obtain a fuzzy RPN value. Using the same O, S, and D values from our previous FMEA study as inputs, new fuzzy RPNs were obtained.
Results: The fuzzy RPN ranking were found to be different from those obtained from our previous study. The top ranking of â€œfailure in dose reproducibility with factory-calibrated MUâ€? for acceptance remained unchanged, however, third ranking of â€œerror in setup due to multiple options provided for reference dosimetry by the manufacturerâ€? changed to the top ranking for commission. Since a membership function of S was constructed with a broader range of â€œvery highâ€? to increase a weight, decision rules affected the FMs with high S value.
Conclusion: The proposed fuzzy inference RPN presents a more reasonable and effective method for assessing potential FMs by incorporating a relative importance among the O, S, and D parameters and factoring in the evaluation uncertainty for the risk assessment of the ATC. â‚?Â¹Teo et al, Med Phys. 2019â‚Ž