Statistical Tools – Pareto

Statistical Tools

Statistical tools can play a huge role in helping us with failure issues. These tools however are not so much about tracing the root cause as they are about warning of pending problems and issues. An early warning system is invaluable and will enable us to schedule and plan to prevent a failure from happening.

Root Cause Analysis as we have seen is a vital tool to assist in solving problems. However, there are also many other statistical methods and tools available that will help us to prevent failures and assist in this war against The Destructive Force of Failure. All of these methods are dependent on the vital recording of asset and maintenance data in a central database repository and this is a further reminder that for any type of preventive failure measure, we need to be running an effective maintenance management system to collect and analyze vital data. Without this information, most of the preventive statistical tools will be ineffective and at best produce results that will not necessarily be reliable.

A few methods to consider are:

• Pareto Graphs or Charts
• Mean Time Between Failure – MTBF
• Condition Monitoring

Pareto Charts

Statistical information is extremely effective to identify trends of equipment performance such as mean time between failure (MTBF), mean time to repair (MTTR), schedule compliance, work order backlog, work order age, and a host of other types of reports and graphical information. Statistical tools are typically not thought of as failure analysis tool, but their importance in identifying trends and deviations from the desired outcome is critical for a successful maintenance effort.

One important statistical tool that is frequently used in equipment failure analysis is the Pareto chart.

The Pareto theory was developed by Italian economist Vilfredo Pareto in 1897 to explain the uneven distribution of wealth! Dr. J.M. Juran started applying this principle to defect analysis, separating the ‘vital few’ from the ‘trivial many’, and called it the ‘Pareto chart’.

This is often referred to as the 80-20 rule, as 20% of the issues cause 80% of the problems, or a relatively small number of issues account for an overwhelming share of the problems.

The Pareto chart shows the relative frequency of defects in rank order, allowing one to organize reliability efforts to ‘get the most bang for the buck’ or ‘pick the low-hanging fruit’.

A Pareto chart can be generated using virtually any spreadsheet or charting software. It is a simple-to-use and powerful graphic to identify where most problems in a plant originated.

It won’t help with catastrophic failures but is an extremely useful tool for finding the chronic problems that over time consume as many reliability and maintenance resources as catastrophic failures.

From the above image, we can see the following

• A list of Issues at the bottom of the graph
• The number of errors on the left horizontal axis
• A percentage scale on the right-hand side of the graph

The red trend line starts on the left-hand side with an error count of about 300, which equates to about 50% on the percentage scale.

When the trend line hits the 80% mark, the error count is at just under 500 but more important is that the actual errors that fall into that bracket are the first three, which represent the majority of errors the organization is dealing with. These are identified as the ‘Vital Few’ and these are the errors that are impacting the organization most and as such are the ones where effort needs to be placed to rectify them.

The effort that would be required to chase down the remaining errors would be considerable but the overall impact on the organization would be trivial (‘Trivial Many’), as the error count on these items is very low.

Effectively by focussing on the errors below the 80% mark, the organization will be making a huge impact on their efficiency in terms of reducing their high occurring errors considerably. The remaining errors that are in the 80% to 100% bracket are minimal and if necessary can be tackled at a later stage. In terms of Pareto, they are correctly identified as the ‘Trivial Few’.

Let’s move on to our final two Statistical Tools.

For more information email us at info@strobe-al.co.za or visit www.strobe-al.co.za