[
home]
[
Personal Program]
[
Help]
tag
11:40
20 mins
Probabilistic lifing of a second oversize hole modification
Guillaume Renaud, Éric Dionne, Min Liao
Session: Session 17: Probabilistic modelling and risk analysis
Session starts: Thursday 29 June, 11:20
Presentation starts: 11:40
Room: Theatre room: plenary
Guillaume Renaud ()
Éric Dionne ()
Min Liao ()
Abstract:
Enlarging a hole to its second oversize is a common airframe maintenance operation. Early in the life of the Royal Canadian Air Force (RCAF) CF-188, this structural modification was seen as a confidence cut, leading to a full life reset at the enlarged hole. However, after some accumulated usage, airframe cracks may become too large to be completely removed by the cut, yet too small to be reliably detected by typical Non-Destructive Inspection (NDI) methods. As such, guidelines have been established to assume that, after a certain level of expended airframe fatigue life, a residual crack would exist at all holes that underwent a second oversize modification. However, as the CF-188 fleet approaches retirement, it appears that this conservative guideline results in likely unnecessary costly inspections for several Life Limiting Items (LLI).
The National Research Council of Canada (NRC) has been tasked by the RCAF to evaluate and further develop a practical probabilistic analysis methodology proposed by the CF-188 maintenance, repair and overhaul contractor, L3Harris, to better evaluate the risk associated with second oversized holes. This probabilistic lifing method establishes a post-modification safe life by combining the probability of cracks exceeding the cut size at the modification incorporation time with the probability of crack detection associated with the selected NDI method. This updated safe life, corresponding to an acceptable cumulative probability of failure, is intended to be applied fleet-wide and was shown by NRC to be conservative for the analysis of an investigated location, LLI 855.
Enhancement to the proposed method were assessed and proposed by NRC. These enhancements assume the worst case of probability of detection, include modifications to apply the method tail-by-tail, and modify the method to make it compatible with the current CF-188 risk assessment methodology.
Although it needs to be tested on additional LLIs, the proposed probabilistic methodology has already been shown to be useful to reduce maintenance costs by saving a late inspection for the investigated LLI, while maintaining the risk level acceptable.