The assessment of the probability of detecting a defect or cause has always been an important element of FMEA analysis. In the new AIAG & VDA manual, however, significant changes have been made to the way D values are determined - they are discusse
The new FMEA manual, jointly developed by AIAG and VDA, has introduced some significant changes to the way risk analysis is conducted. First, the RPN indicator was replaced with a new measure of AP (action priority). AP is not a calculated value (so far RPN, which was the product of three indices), but it is determined from a special table. Secondly, the standards for assessing all components of risk assessment: severity (S), frequency (O) and detectability (D) have changed. A particularly innovative and debatable approach was applied to the assessment of this third criterion.
The criteria for assessing the probability of detecting an error or its cause are given in the new AIAG & VDA manual ed. 1, clearly refer to the FMEA AIAG standard ed. 4, but introduce a significantly new point of view. An abbreviated version of these D selection criteria for PFMEA according to the AIAG & VDA standard ed. 1 is shown in Table 1.
Table 1. Criteria for the assessment of the detection rate for PFMEA according to the FMEA AIAG & VDA textbook, ed. 1
|D index||Control method|
|8||Visual, tactile or auditory inspection of a fault, or inspection using indicators – without proven effectiveness.|
|7||Automatic control with signaling or semi-automatic – without proven effectiveness.|
|6||Visual, tactile or auditory inspection of a fault, or inspection using indicators – without proven effectiveness.|
|5||Automatic control with signaling or semi-automatic – without proven effectiveness.|
|4-3-2||Automatic controls with stop.|
|1||Non-conforming pieces cannot be produced thanks to the safeguards in place, or there is evidence that the inspection method will always detect the error or the cause.|
Source: own study based on FMEA AIAG & VDA ed. 1, 2019
It is worth noting that, for example, visual control currently has two rating options: 8 or 6, but the distinction between these ratings is not clear in FMEA AIAG ed. 4 from where the control is performed (on the position that generated the error or later), but on whether the effectiveness of the control is confirmed or not. Similarly, automatic and semi-automatic controls may have values of 7 or 5 depending on the assessment of their effectiveness. Therefore, a completely new point of view for FMEA on detectability was introduced.
A big novelty in the AIAG & VDA standard is also the fact that when determining the D index, the assessment (reliability) of the visual control was equated with the control of gauges and measuring instruments (so far these were values, e.g. 8, 7 and 6, respectively). Now, therefore, it is not important what control is carried out by, for example, a quality controller, it is only important whether it has proven effectiveness or not.
A very interesting aspect of the practical approach to designing control systems in the company also results from the analysis of the table designed to determine the AP value (table 2).
Table 2. Fragment of the table with AP values
Source: FMEA AIAG & VDA ed. 1, 2019
It is worth paying attention to two “blocking” sets of D values, which result in the same impact on the AP level:
You can have big doubts especially about the second block 5-6, because are you sure that both of these controls have the same potential for detecting non-compliance? If so, it can be ironically said that it is worth dismantling and selling in all companies visual control systems based on cameras (D = 5) and replace them with a cheaper solution – a quality controller sitting at the production line (D = 6). It seems that the locking of the D 5-6 values is a certain omission by the authors of the FMEA manual. They left the “idea” in the final version of the standard, which was shown for the first time in the draft of this document issued in 2017. There, the grades 5 and 6 were combined, although it was already debatable at that time, because the control with confirmed effectiveness (D = 5) was combined with the one without confirmed effectiveness (D = 6), and additionally one was carried out on the stand, the other later parts of the process. In the final version of the standard, however, very divergent controls were combined: manual and automatic – hence the aforementioned doubts.
Finally, it can be concluded that the new criteria for assessing the detection of errors and their causes presented in the manual FMEA AIAG & VDA ed. 1 are an interesting solution, especially due to the fact that the D assessment was linked not only to the type of control, but also its effectiveness.
Author: Tomasz Greber [PROQUAL]