How accurately can you answer these questions about your facilities?
- What piece of equipment has had the most chronic (low cost, high frequency) failures in the past six months?
What type of equipment has had the most failures between scheduled preventative maintenance (PMs) in the past two years?
If you can't answer these questions and don't know how to find the information quickly and easily, you should think about implementing failure codes.
Understanding Failure Codes
Failure codes are predetermined, consistent categories into which partial and total failures are organized to facilitate data mining.To understand what this means in practice, we can first look at maintenance systems that don't have failure codes.
Every work order is a reaction to a problem, but there's often inconsistency in how these problems are described. It's easy to see why; different people explain things differently. And because this is just human nature, the confusion this causes can follow you even when making the jump from traditional paper-and-pencil or spreadsheet systems to more sophisticated CMMS.
Say there's a type of electrical motor that's vibrating more than it should. The first work order might describe the problem as "excessive vibration," but when the problem reappears a week later, the next work order refers to "unexpected shaking." And the third one, a month later, uses "undesired wobbling." There's a trend here, but you couldn't easily see it without having the three work orders laid out in front of you, because each technician is describing the same problem but using a different expression.
Would it be possible to see trends without the help of failure codes? Sure, but to uncover them you'd have to painstakingly go through every work order by hand (or click by click if you're using spreadsheets or work order management software) to build your own lists of cross-referenced, similar-sounding failures.
Now what about maintenance systems with failure codes? Here you can document problems consistently. Everyone is using the same descriptions because everyone is working from the same limited list of failure codes.
Since the data is uniform, it is much easier to go back and organize it into something truly useful. In fact, many CMMS software and EAM software packages can generate data mining reports in just a few clicks.
Ensuring Sufficient Data
Implementing failure codes for your most common problems is a good start, but if you stop there, you're not going to reap all the possible benefits. If you're using a small number of maintenance failure codes, you're only ever getting a superficial map of your maintenance landscape. Here's why:
The smaller the number of failure codes, the wider the category of failures each code represents. Once the categories become too wide, failures that are actually different from one another get lumped together because they all sort of match the same broad description. Having more codes lets you narrow down each category, and that lets you more finely separate failures by type. To ensure departments are collecting useful data, a lot of code libraries are fairly extensive. On top of a wide spectrum of failure codes, they include two additional code types. A quick way to understand each one is by thinking about the questions they help you answer.
- What are the symptoms?
- What appears to be the problem?
- What problems led to the failure?
- What's the root cause of the failure?
- How did I fix it?
Goldilocks-ing Your Failure Codes
We know that having too few codes leaves you with only a surface-level understanding of how your department is spending its resources. But having too many codes also causes problems.
Imagine a technician notices a pump is not cycling properly. When they open the equipment maintenance software to write up the work order, they're faced with a pull-down menu packed with hundreds of failure codes.
They're either going to choose the first one that sounds even remotely descriptive of the current problem, or they're going to quickly scroll down to the last option, "Other."You can't really blame them, because what they really want to be doing is fixing that pump. when you have too many codes, technicians can feel overwhelmed, and the quality of your data goes down. Soon you're bumping up against the old adage "Garbage in, garbage out."
So, is there a magic number? There likely is, but finding it could take a bit of experimentation on the part of your maintenance department, both with the number of codes and how they're presented in the CMMS or facility management software. You could, for example, break up the codes according to asset type. If a technician is working on a pump, like in the example above, only the subset of codes specifically related to that asset type appear on the pull-down menu.
Digging Deep into the Data
Remember the questions at the start of the post? At this point, it's clear that failure codes would help you answer them quickly and easily. Let's look at those questions one more time, and think about how knowing the answers saves you time, money, and frustration.
What piece of equipment has had the most chronic (low cost, high frequency) failures in the past six months?
Now that you know which equipment is the most prone to failure, you can dig into the cause codes and find out why it's giving you so much trouble. It might be that the current operator requires additional training. Or, you might look at the remedy codes and notice that the same failure is being fixed two different ways, and only one of the fixes tends to last.
Digging into the answer to the first question is important because fixing chronic failures is crucial to cutting cost. Charles Latino, when he was president of the Reliability Center, famously estimated that eliminating chronic failures in American industry would reduce maintenance costs by 40% to 60%.
What type of equipment has had the most failures between preventative maintenance (PMs) in the past three years?
Looking at your data, you might find you need to adjust the PM schedule in your preventive maintenance software. For example, If the most common failure is caused by parts seizing, you can lubricate more frequently.
Taking the First Steps
It's possible to start small and implement your own basic system of failure codes, steadily building up to a more expansive system you tailor to your department's specific needs. Off-the-shelf code libraries are also available. Regardless of where the codes are from, the most important thing is to trust that there is value in collecting and examining data. It requires a department-wide effort, but can deliver department-wide savings in time, money, and frustration