Failure Modes in FMEA for a Six Sigma Initiative - dummies

Failure Modes in FMEA for a Six Sigma Initiative

By Craig Gygi, Bruce Williams, Neil DeCarlo, Stephen R. Covey

FMEA can be very valuable for identifying failure modes in a Six Sigma Initiative. After scoring the severity of the possible effects, your cross-functional FMEA team brainstorms potential causes of the identified failure mode.

Think of causes for the failure mode, not for the effect. In the pizza example, you need to think of causes for why the phone is answered on or after the fifth ring, not causes for why a customer hangs up or why a customer becomes disgruntled.

Be sure to list potential causes of the failure mode in separate rows of the FMEA table, which usually means you have to insert new rows into your FMEA table.

Score the occurrence of the cause

The next important task is determining how often each listed cause occurs with a 1 to 10 occurrence score; 10 represents a very frequent occurrence and 1 a very rare occurrence. In a typical FMEA occurrence scoring scale, the second column includes information from the AIAG on the consequences of the frequency, and the last column gives information on a yield level associated with the frequency of occurrence.

This stage is one where the participation of the cross-functional team is critical. Members of the team from different aspects of the process will have a good idea of how often each listed cause occurs. Has the listed cause ever happened before? About how many times over the last month? The last year?

In the unlikely event that no one on the team knows how often a particular cause occurs, the team should go to gemba (that is, the place where the work occurs) and measure for themselves how often the cause is occurring.

Identify current controls

The next step in the FMEA process is for the team to identify what controls or protections are in place within the current configuration of the process to prevent the cause from happening or at least to detect and provide an alarm when the cause does happen. Typical controls include operator training, process instructions, automated in-process inspection, and so on.

Causes often have more than one control. In this case, list all the controls in one table cell and separate them with a + or &.

Score the detection of the controls

Just as with the severity and occurrence scores, the cross-functional team now gives each listing of controls a score for the controls’ effectiveness at preventing or detecting the associated cause. This scale utilizes scores from 1 to 10, with 10 indicating no ability to prevent or detect and 1 signifying perfect ability to prevent or detect.

SPC refers to statistical process control, which is the method of using specialized statistical probability calculations to monitor and detect external influences on a system’s performance.

Be careful not to accidentally flip the numbers of the detection scale. Remember, a detection score of 10 is bad and means you have no detection, while a score of 1 is good and indicates that you have perfect control.

When giving detection scores, keep in mind that training, reminders, or even formal discipline of employees aren’t effective means of preventing or detecting causes of failure. When training or operator awareness is the only or the best available control, the detection score is usually given a fairly high score — maybe a 7 or an 8.

Calculate and review RPN scores

With severity, occurrence, and detection scores for each row, you can calculate an overall composite risk priority number, or RPN. An RPN is simply the severity score multiplied by the occurrence score multiplied by the detection score. The lowest possible RPN is 1 (1 × 1 × 1), while the highest is 1,000 (10 × 10 × 10).

With the FMEA filled out for each process step and with an RPN for each row, the cross-functional team can quickly identify where risk currently resides within the process.

Because the RPN is a composite score, rows in the FMEA table that have same or similar RPNs may need to be addressed very differently. Be intelligent in your review; one process step with a particular score may require action to reduce the occurrence, while another with the same score may require improved detection.

How to devise and assign improvement actions

The far right-hand column of the FMEA table is for coming up with improvement actions that will improve the current risk situation. These actions may include taking additional data or measurements to more accurately understand an occurrence of a specific cause or implementing improved detection actions. You can record these assignments and target completion dates directly into the FMEA with revised severity, occurrence, and detection scores.