What is Asset Performance Management (APM)?

Let’s start by working the acronym backwards. APM is the management of the performance of your assets. Management needs data to make decisions with. Performance is measured with data. Assets and people can be sources of data. Now it all makes sense… it’s all about data! APM is certainly a big piece of your approach to digital and operational excellence and needs to have some structure around it to be successful.

Knowing the performance of your assets only has value if it leads to insights into looming failures, reduces unplanned or unnecessary work, extends equipment life, increases margin, or in general helps you manage the risks the equipment presents to the organization. It’s what you do with the performance information that is important… not simply that you have it.

It is very possible that pieces of your approach to APM do not originate from your organization. Vendor or contractor information about the performance of your assets may be a key element in your decision-making process. In fact, you may have outsiders managing aspects of your assets that have a direct impact on their performance and benefit your operation. These outside “services,” need to be well understood and continually evaluated for relevance and usefulness.

Managing the ever-increasing amount of data we have access to is a significant task that is often over looked. Synonymous to your Inbox… a constant flow of “good,” information from which you get to decide when to reach into the river and pull out a gold nugget to help with a decision. You will need tools, people, processes, and techniques in place to govern and transform your data into something useful. Gone are the days where a shared spreadsheet is sufficient.

No matter the industry you are in… yes, yes, of course you are special… the assets under your purview were purchased to perform a task to the benefit of the organization. Your industry may have guidelines, standards, and even regulations that specify acceptable envelopes of performance. This is a good thing! You may have a foundation that others do not. You may already have a long list of insights you must have from your data sets. Now you just need the tools to make the connection.

Given that you know what your assets are, can measure their performance in the context of your business, and if someone is minding the store, you are well on your way to achieving enterprise wide APM.

What are the core elements of APM?

Like any other industry concept, we can argue all day long about what it consists of or what the boundaries are. Our view of APM looks like this: Activities that include the monitoring of assets to then identify, diagnose, and prioritize impending equipment problems. These activities are performed continuously and in real time, facilitating the determination of asset strategies such as repair, replace, defer, etc. APM goals include the reduction in unplanned or unnecessary work, extend equipment life, and increase company margin.

These activities are certainly supported in part by our current set of traditional CMMS/EAM systems. But in recent years a plethora of data science and machine learning tools and techniques have emerged. It is very possible that the amount of data coming from your equipment has increased to a point where you are questioning how it might be used to the advantage of the organization.

A new way of looking at this growing stack of data is as a self-generated beneficial resource that should be mined for insights. Buried deep in those hard drives are trends, anomalies, failures, and other useful data points that can be very valuable given the right tools and skills leveraged upon them.

Does optimization play a role in APM?

Yes, optimization plays a significant role. Any process that produces data can be analyzed and then optimized. For example, maintenance (corrective or preventive), scheduling, inventory, budgeting, and even capital investments, can all be looked at with the bright lights of insights from properly prepared and analyzed data.

No matter the type of process, the most common business problems are usually inefficiency or non-repeatability. Collected data will very often have the secrets to solve these… you just need to find them.

One good example is around customer service operations. As tickets are entered either by or for clients, a hidden clock begins to tick. There are sweet spots for response and resolution that makes a happy customer. Knowing how well the organization is operating within these subjective ranges can make or break the perception your clients have. Maybe you respond quickly but it takes a long time (for several reasons/excuses) to resolve their issue. Maybe it is the opposite. No matter, finding the source of the trouble and then optimizing the local flow in the process are two critical activities.

A specific example for the maintenance world is that of when to perform preventive maintenance activities. Most operations loosely follow the manufacturers’ recommendations for fluid or other replacements based upon calendar or operations times/cycles. While following these recommendations might work out well for some equipment, they do not consider the operating context or environment the equipment is in. A motor operating in a basement has very different needs than the very same model up on the roof.

By taking advantage of the telemetry collected from these two motors through analysis and machine learning, you will be able to detect anomalies that can infer impending failure. There are even hidden trends and correlations that can also warn of trouble ahead. Further, using advanced analysis tools, you can easily understand what, “normal,” is and make appropriate changes to your preventive maintenance approach. Executing the optimal preventive maintenance activities, based upon the context of the equipment and real data can have significant value to an organization. Unnecessary tasks, inventory levels, and many other costs to maintain the equipment can be greatly reduced or even eliminated by mining your data.

Catching up… artificial intelligence and machine learning

We briefly mentioned AI and machine learning earlier, but let’s discuss what AI can do today and how it might benefit our operations and approach to APM. Consider the large, diverse, and ever-growing data sets that our operation is generating every second of every day. Also, as we mentioned earlier, these data sets are a valuable resource that is “just happening,” as we go about our business. It has truly reached a point where humans cannot just look at a spreadsheet and find the tall poles. There are nuances and patterns hiding in the data that we simply cannot find on our own. Using machine learning that leverages tried and true analytical methods, we now have tools that can dissect our data sets, learn about them, and show us insights and predictions that only a few years ago we didn’t know about.

Keep in mind that many of the common tools and techniques used by current data scientists are well known in the statistical and analysis disciplines. Anomaly detection models, correlation models, multivariate time-based data sets, supervised vs unsupervised data sets, Poisson, and Weibull distributions, etc. are approaches that have proven their reliability many years ago. What is new is how these techniques are being used to perform tasks such as early warning of failures.

If the performance of your equipment is measured by more than a few variables, then adopting AI/ML tools will certainly benefit the bottom line with less downtime and reduced operating costs. Having a high degree of confidence in the likelihood of a predicted failure gives you time to plan and schedule for the event, and greatly reduces the stress on the organization. Your decision-making process will be far more “sure,” given these new insights you didn’t have before.

Wrap up

We hope it is clear at this point that Asset Performance Management is wide ranging and can include a collection of historically proven methods along with a growing list of advanced tools and techniques which leverage machine learning and artificial intelligence. The investment you have made in the assets that make your business function can be at risk if they are not performing as you are expecting. The key is knowing what that performance is over time, and what you are doing about it if it falls below the levels you have set.

TRM and IDCON have been working to assist clients to squeeze the most they can out of their liveries of assets for over 30 years. There isn’t a type of equipment that we have not encountered, so we are uniquely qualified to assist you in your APM activities.