No abstract

No abstract

:The FMEA is a widely used technique to pre-evaluate and avoid risks due to potential failures for developing an improved design. The conventional FMEA does not consider the possible time gap between occurrence and detection of failure cause. When a failure cause is detected and corrected before the failure itself occurs, there will be no other effect except the correction cost. But, if its cause is detected after the failure actually occurs, its effects will become more severe depending on the duration of the uncorrected failure. Taking this situation into account, a risk metric is developed as an alternative to the RPN of the conventional FMEA. The severity of a failure effect is first modeled as linear and quadratic severity functions of undetected failure time duration. Assuming exponential probability distribution for occurrence and detection time of failures and causes, the expected severity is derived for each failure cause. A new risk metric REM is defined as the product of a failure cause occurrence rate and the expected severity of its corresponding failure. A numerical example and some discussions are provided for illustration.

Taguchi method is a technique widely used in industries to attain a reliable quality even under existence of various noise variables. Basically, it relies on the interactions between the controllable and the noise factors. Under insignificant interactions between the design parameters and noises, the effectiveness of Taguchi methodology will be doubtful. Furthermore, existence of strong interactions among controllable factors may undermine the power of Taguchi approach. An abundant crop may be expected by a well planned Taguchi approach. A well designed set of experiments is one of the important prerequisites for its successful result. For this, we must have a clear understanding on strengths and weaknesses of Taguchi methodology together with a sound knowledge on the applicable system. This study provides some insights on Taguchi approach based on two case examples of one successful and the other unsuccessful. Eventually, this research aims to facilitate deeper understanding and right application of Taguchi methodology in industries. †

:To determine a proper ASIL for each hazardous event with a proper safety goal, the right classes should first be determined for the three properties of the hazardous event; (i) severity of harm from the resultant accident, (ii) exposure to the relevant operational situation, and (iii) controllability to avoid the induced risks. ASIL can be clearly determined with right classes of these three properties. But no specific methodologies or processes for their classification can be found in ISO 26262, except only a rough guideline with a simplified set of illustrative tables. In this paper, we try to present a systematic model for classifying the three properties of the hazardous event and suggest a refined procedure of ASIL determination. The proposed model provides a specific method to get a more objective ASIL compared with that in the standard. Scrutinizing the current methodology, we develop a refined method and also provide an illustrative example.