Using Crow AMSAA Plots for RCAs

Root Cause Analysis (RCA) Overview

A Root Cause Analysis (RCA) is a structured approach to identifying the underlying factors that result in the unwanted/unexpected outcomes of chronic or sporadic events. It is a methodic and rigorous process. It also highlights what assets, systems or behaviors need to be modified to limit or eliminate recurrence of similar undesirable outcomes. The fundamental driver is to address, correct or mitigate the root causes that lead to the unwanted event(s) rather than addressing the symptom(s). Bob Latino, a renowned RCA expert, summarizes this concept as “the establishing of logically complete, evidence based, tightly coupled chains of factors from the least acceptable consequences to the deepest significant underlying causes.”

RCAs are extremely informative and value-added exercises. They can stand out as “learning moments” for an entire organization. But their depth causes them to be work intensive. Due to limited resources in every organization, the initiation of an RCA has to be carefully evaluated and prioritized.

Crow AMSAA Overview

Crow-AMSAA Plots have a variety of names, such as Reliability Growth Plots or Duane Plots. The term “Crow” comes from Dr Larry H. Crow, who enhanced James T. Duane’s pioneering launch of this methodology, developed in the early 1960s (1). Dr. Crow successfully applied the method in the US Army Materials System Analysis Activity (AMSAA). The technique has blossomed into large amounts of new applications in industry such as, but not limited to:

• Making reliability more visible and manageable hence contributing to Reliability Improvement Programs,
• Monitoring Design Optimization and Quality Performance,
• Catastrophic or adverse event occurrence predictions and trends.

In practical terms, the Crow-AMSAA technique involves plotting, cumulative failures vs cumulative time on a log-log scale resulting in straight line plots (2). The line slope value (or Beta value) indicates improving, deteriorating, or constant failure occurrences. Due to the straight-line nature of the plots, future failure forecasts can be estimated. In plain words, based on the current trend, when is the next failure expected to occur? This method handles mixed failure modes, so it is, therefore, suitable for the complex nature of the generating units.

What is of interest to us in this article is the change in trends. Sometimes undesirable events increase in frequency at a specific moment in time. A higher frequency of occurrence can lead to a change in trend in a Crow AMSAA Plot. When doing an RCA on those events, it is crucial to know when they started. In other words, the timestamp of the event would indicate “what changed” and eventually “when the change occurred”. This allows the RCA team to review decisions made at the time. And speed up the RCA. Remember that RCA exercises are thorough but in so doing, time consuming and tedious. Therefore, any technique that makes the analysis more efficient is welcome. 

Process Trip Example

A basic example of this concept is provided below. Process trips or interrupts were recorded during a specific time interval; from years 2020 to 2025. Those trips occurred following a process overload or component failure. They were concerning to management as they caused stoppages in production leading to downtime, output reduction and potential safety issues. However, initial higher trip intervals were deemed acceptable to management, but later trips increased in frequency causing concern. The time to failure was reduced as shown in the right-hand section of Graph 1 below. Subsequently, a Root Cause Analysis exercise was initiated.

To a certain extent, an analyst can see in Graph 1 above that the increase in failure occurrences is occurring between the 800^th and 1000^th consecutive day. They can choose to move on with the RCA and find out what changed around this interval. However, Crow-AMSAA plots could offer more precise information on the changes as shown below.

Using Crow-AMSAA plots – A more elegant method

The data from Graph 1 above is transformed to a Crow-AMSAA logarithmic plot as illustrated in Graph 2 below. The raw data is attached here should the reader want to use it.  

Graph 2 illustrates the inflection point where the initial line (green) changes to a steeper sloped line (pink). This change or inflection occurs precisely on October 27^th, 2022. The successive slopes are quantified by the Beta value. The higher the beta value, the higher the slope. And the lower the time to failure. The correlation factors (r^2) are high (i.e. very close to 1) indicating good models for each individual segment.  

In this scenario, the operating system was pushed harder (over loaded) due to increases in production demand. The capabilities of the equipment were pushed to the limit, leading to more trips. The RCA will eventually highlight the procedures that were overlooked or lacking when the decision was made to increase production.

Conclusion

There is no dramatic change in inflection in Graph 2. However, the visualization is somewhat neater and more elegant. Providing straight lines and a clear inflection point. Additionally, the MTBF (Mean Time Between Failures) values are identified which can help assess the impact from one process output to another. MTBF values indicate that the average frequency of failures has increased by approximately 50%. This could help the operator quantify the financial impact of the changes in operational load. And financially justifying the RCA exercise as well as the mitigation measures to be implemented.

There is no absolute need to use the Crow AMSAA methodology for this exercise. The chronological data in Graph 1 would suffice. However, the Crow AMSAA plot provides more precise information, better visuals, and more quantitative decision-making values.

References

1. Robert B. Abernethy, The New Weibull Handbook, 5th Edition (2007), Section 9.1-9.2.

2. Nigel Comerford, Crow/AMSAA Reliability Growth Plots, 16th Annual Conference 2005 – Rotorua.