Many maintenance organizations around the world are reactive. They experience excessive breakdowns, low productivity, and high maintenance costs. To break this reactive cycle, a common solution organizations implement is Condition-Based Maintenance; however, this strategy rarely works. They pay for CBM solutions but get the results of a run-to-failure strategy. Here’s how to avoid that.
Condition-Based Maintenance (CBM) technologies are some of the most sold solutions in the maintenance industry, and with the recent interest in Artificial Intelligence (AI) and the Industrial Internet of Things (IIoT) combined with Predictive Maintenance (PdM), the market has expanded substantially.
It's easy to understand the reasons why organizations opt for CBM or PdM solutions, such as:
Clearly, these CBM and PdM platforms are the best things since sliced bread. But the truth is when a reactive organization invests in one of these solutions, it rarely delivers the promised result, and the organization continues to experience too many unacceptable and preventable failures.
Here’s the shocker — it's not because the technology is faulty or that the vendor has delivered a poor-quality solution. The problem is that most reactive organizations are not ready to manage the solution they purchased. They are too reactive and too busy with today’s emergencies to respond to the information their CBM system is providing.
In the rest of this article, I will discuss Condition-Based Maintenance in greater detail, why you can't achieve a reliable plant without it, the reason why most CBM efforts fail, and what you need to do to ensure that your Condition-Based Maintenance delivers value to your organization.
First, let’s have a quick look at what Condition-Based Maintenance really is.
Most failure modes are not age-related; however, they do provide some warning that they are about to occur. If we can find these signals, we can take preventive action to minimize the consequences of failure.
As a strategy, CBM looks for physical evidence that a failure is happening or is about to happen. When thought of this way, CBM has broader applications outside its more common uses, like vibration monitoring.
An important concept within Condition-Based Maintenance is the P-F curve shown in the figure below:
Image Source: Road to Reliability
This conceptual diagram shows that the functional performance of equipment deteriorates over time to a point where a failure can now be detected (point “P”). If the failure isn’t mitigated, the equipment will degrade until a functional failure occurs (point “F”). The time range between P and F, commonly called the P-F interval, is your window of opportunity to detect imminent failure and address it.
The length of the P-F interval varies depending on the specific failure mode and the condition monitoring technique used. You want to use a condition monitoring solution that gives you a reliable, repeatable indication early enough for you to respond appropriately.
Condition-Based Maintenance lets you know a failure is about to happen, allowing you to plan for it and strategically choose the best time to address the impending failure. For example, let’s say you have a scheduled outage next month, and according to your CBM, the drive-end bearing of your export pump will fail in roughly three months.
Without CBM, that failure would have been unexpected, leaving you to deal with it as a breakdown, which is inefficient and expensive. But with CBM, you can properly plan and schedule the fix during the scheduled outage, meaning the equipment is repaired safely, quickly, and at a lower cost, all of which reduces the impact on production.
To be clear, the value of CBM is not in failure prevention; it’s in identifying failures before they occur. As a result, you can make a timely intervention that reduces the impact on production, decreases costs, and improves worker safety.
But not all failure modes or equipment will benefit from Condition-Based Maintenance.
Failure modes that are clearly age-related are often best managed with a time-based repair or replacement task. If you know a component will wear out every 12 months, you simply replace it in time. You don’t need a condition assessment every few months to determine if the component needs replacing.
But research across different industries (airline and military) has shown that between 70% and 90% of all failure modes are not age-related, and for these failure modes, the likelihood of failure doesn’t (significantly) increase over time.
This research is summarized in the six different failure patterns shown below:
Image Source: Road to Reliability
This has major implications for our PM programs, which are filled with time-based maintenance tasks. It reveals that if most failure modes are not age-related, then we should be addressing them with some type of Condition-Based Maintenance.
As a result, a significant part of your PM program should be Condition-Based Maintenance. In other words, CBM is essential to creating a reliable plant. You will not succeed without it.
Effective condition monitoring generates “weak signals,” or indicators of potential failure weeks, if not months, in advance. But what happens to those weak signals in a reactive maintenance environment? They get ignored.
People are too busy dealing with today's emergencies to worry about a pump bearing that might fail in six months’ time. So, six months later, that exact same pump breaks down due to a bearing failure.
Management asks why that bearing problem wasn't captured through the condition monitoring program, only to be told that it was: the report with the vibration trends is there, and the CMMS work requests were raised, but nothing was done about it.
This is the worst of both worlds. You're practicing run-to-failure — except it's more expensive because you're adding the cost of CBM to it.
Sadly, this is a reality for facilities, and to be honest, I have personally experienced this myself. To resolve this, organizations need to ensure they have an effective planning and scheduling process in place so the work environment becomes more stable and can act on the “weak signals” that come from the CBM program.
As we saw, for CBM to be an effective strategy, early intervention is essential. This requires an efficient and effective process for gathering data, data analysis, decision-making, and timely intervention. It is this last step, timely intervention, where reactive organizations typically fall short.
These early interventions just don’t survive in the daily reality of many organizations; they aren’t seen as urgent enough until it is too late. For your CBM program to be effective, you need to have a stable working environment with:
In other words, you need an effective planning and scheduling process in place before you implement a CBM strategy. It's like building a house; you need to have the foundation in place before you add anything else.
In the same way, you must have planning and scheduling in place so that when you specify CBM as part of your PM program, you can guarantee that any required interventions will be executed on time. When I speak to CBM and PdM solution providers, I hear this issue repeatedly — their technology is flagging impending failures, but the maintenance teams are not getting to them in time. As a result, their clients are not getting the value they expected and are not happy.
In most maintenance organizations, productivity is often poor, typically around 20% – 30%. This means that during a typical ten-hour workday, your technicians only spend two to three hours doing actual maintenance work, sometimes even less. This is not because people don’t work hard enough; it’s because our systems and processes are inefficient and ineffective. Our day-to-day maintenance work is full of all kinds of waste, such as:
Maintenance organizations are often tightly resourced as it is, and then having a typical productivity of only 20% to 30% means our maintenance teams really struggle to get the work done. That eventually leads to a vicious cycle of reactive maintenance where the work that gets done first is the work that is shouted about the loudest — and guess what: that won’t be our CBM work request because nobody is super excited about an impending failure that is three, six, or even nine months away.
But with an effective maintenance planning and scheduling process, you remove a lot of this waste. My experience shows that with relative ease, you can increase productivity by 30% to 45%, and that is the equivalent of increasing your team’s productivity by 35%, all without hiring anyone.
By implementing a robust, effective process, your work environment also becomes less reactive. Teams finally find the time to get rid of those recurring problems, your people are less frustrated and start to see the benefits of a reliability-based culture, and those early warnings from your CBM program will get the attention they deserve.
If you want to know where your gaps are when it comes to maintenance planning and scheduling, then you can utilize a maintenance and planning scorecard that provides personalized feedback and recommendations on how to improve.
Let’s be clear, planning and scheduling are essential prerequisites before you implement a Condition-Based Maintenance program — without them, your CBM efforts will fail. But there is, of course, a great deal more to consider when creating an effective CBM program. You must:
Condition-Based Maintenance is essential to creating a reliable plant because most failure modes are not age-related and can only be effectively managed with CBM. The problem most organizations face, however, is that they operate in a reactive maintenance environment and are unable to effectively deal with the “weak signals” that CBM provides.
Implementing CBM too early is a waste of time and money. To make CBM successful in your organization, you first need to have a stable working environment in place so that you can act on the early warning signals provided by your CBM program in a timely fashion. In doing so, you effectively create a stable working environment with an effective maintenance planning and scheduling process.