Our AI. Their AI. Is there really any difference?
In case you haven’t noticed, AI is changing the way we live, work and play every day. From customized playlists based on your music listening habits, to the creation of photorealistic graphics and even videos generated by simply speaking what you want into a chat window, AI has opened new horizons as it finds its way into almost every segment of society.
Asset health management is no exception. In fact, right now, it’s hard to talk to any provider of asset health solutions that isn’t touting the power of “their” AI.
In this article, I’d like to dive a little deeper into the question of what sets one provider’s AI apart from another’s. However, my short answer may surprise you…
The difference is less about the AI itself
If we consider the core AI that detects anomalies in data, the truth is that it isn’t that much different from one provider to the next. What truly matters is all the other things wrapped around the core AI.
AI solutions are analogous. All the things that surround the core “anomaly detection” function of AI differentiate one solution from another. This infographic showcases how the AI we are delivering today is different than what we’ve done in the past – and indeed different than what others in this industry are offering – resulting in six key ways that our AI sets itself apart.
In this article, we’ll explore just one of those differences: how integrated work management tools drive action from AI insights and amplify outcomes at scale. You can download the infographic to explore five other areas in which the power of our AI is leveraged to deliver superior outcomes.
The Trifecta
As context, let’s first consider that AI solutions must answer three fundamental questions:
- Is something wrong?
- What specifically is wrong?
- How should I respond to the problem?
It’s well understood that AI is exceedingly good at answering the first question – often much better than threshold-based alarms.
In a future article, I’ll address question #2 more fully. Suffice to say for now that this is where our approach of combining deep domain knowledge, codified in physics-based models, agentic workflows, and scalable enterprise intelligence, accelerates and, in many cases, automates the journey from “something is wrong” to “this is what your problem is”.
Question #3 can be thought of as context. How important is this problem in the grand scheme of things? How urgent is it that I address it? What should be done to address it? We refer to questions 2 and 3 collectively as “prescriptive analytics” because at the end of the day, what a customer needs is not just to know that something is wrong, but specifically what is wrong and ultimately what to do about it. A prescription, if you will, for how to proceed.
Case Management
Knowing what is wrong and what to do about it raises another important point: do you have an environment with which to collaborate with people, assign actions, provide supporting data, and track issues to closure? In short, you need some way to manage and optimize the work resulting from an identified machinery anomaly. We refer to this as “case management” and without this extremely important capability, you must resort to building your own DIY solution or trying to force a generic solution into the specific needs of asset health where one size almost never fits all.
Case management is a core capability built into all our Cordant™ software offerings. Once our asset health analytics have detected a problem and notified you of the specifics and the actions, you need a way to manage the issue all the way from inception to closure. Our case management environment delivers exactly that.
Walk-Through
The easiest way to understand case management is by way of an example where we show how an issue has been detected by our Cordant™ Asset Health Analytics engine, and how a user would respond by opening a case and tracking their own actions, as well as the actions of others, all the way to closure.
Let’s imagine you are the machinery engineer responsible for equipment across numerous offshore platforms in the North Sea. You’ve been away over the weekend and upon getting back to the office, the first thing you do is open your Cordant™ Asset Health application and notice that the dashboard is not quite as you left it on Friday.
First, you see that the same asset is in “intervene” status – and you were already planning to address that one first thing today. You also see that the same 11 assets for which maintenance has been scheduled are still pending the execution of that maintenance. However, there are several new assets that carry the “Monitor” status and these deserve your attention.
You drill down and decide to go after the notification marked “cavitation” first. You find that it is an AI-detected issue for recurring cavitation on a water injection pump on one of your platforms.
At this juncture, it’s also crucial to understand that Cordant™ Asset Health Analytics is not confined to the use of just vibration data. In fact, in this example, the water injection pump doesn’t have any vibration monitoring equipment on it – it has only process data available from the control system. However, because the AI is “agnostic”, it doesn’t care what kind of data it is or where it originated – it was able to spot an anomaly and – by looking at multiple process variables – determine that the issue was cavitation based on expected versus actual Net Positive Suction Head (NPSH).
Next, you look at the history and see that cavitation has occurred several times in the last 3 months on this asset.
You then examine some of the supporting evidence and see that there have indeed been issues with NPSH prior to each incident. You also examine the work history and see that on each of the prior incidents, debris was found in the pump’s suction filter, causing the cavitation because when the debris was removed, the problem went away.
However, since the cavitation has been a recurring issue, simply cleaning the suction filter yet again won’t address why it is repeatedly becoming clogged. Armed with this information, you open a case by right clicking and selecting “Create Case”.
This time, however, you ensure that the instructions to the maintenance team are not just to inspect the filter for debris, but to investigate the situation further upstream to see where the debris might be coming from and to thus address root cause.
A panel opens and you begin populating the newly initiated case with appropriate information by:
Giving the case a name
Adding a description
Assigning to a category such as “Mechanical Health,” “Process Health,” “Instrument Health,” “Emissions,” “HSE,” “Strategy,” etc.
Assigning a risk type (condition)
Assigning a priority
Assigning to one or more individuals
You then view the case you have just created and, if necessary, add supporting evidence like images, documents, and links to plots, associated events, etc. You can also annotate the case with observations, assign additional actions, view work notifications, and see the history.
You can also see the status of assigned actions such as “in progress,” “complete,” “not started,” etc. You change the status to “Schedule” which drives additional investigation to take place of not just the filter, but the elements upstream of the pump to see if the source of the debris can be isolated and mitigated.
Lastly (and not shown here), you look closely at the other new “Monitor” statuses that came in over the weekend and disposition each of those as appropriate by using the case management tools at your disposal – while also moving forward with the asset in “intervene” status to ensure it receives the attention it deserves.
Outcomes from Case Management
In this example, case management capabilities not only allowed you to schedule the necessary work in your EAM/CMMS system but also capture the costs and savings. This is returned to the Cordant™ environment, allowing much easier and more accurate communication with management that monetizes the actions taken within the system. It also assists you in prioritizing issues in the future because the economics of decisions are better understood. In this case, your instincts were correct: a source of debris was indeed found upstream of the pump and the situation was remedied. The issue that had previously occurred six times in three months has not reoccurred in the last 12 weeks – precisely because you were able to address root cause rather than treating symptoms – and it does not seem likely to reoccur as a result.
The bottom line
Case management delivers a complete set of tools to initiate and coordinate actions – and seamlessly integrates with CMMS and EAM solutions like SAP, IBM-Maximo*, and others.
Cordant™ Asset Performance Management blends asset health, asset strategy, and asset defect elimination into a single, powerful ecosystem and allows you to take your efforts to the next level of effectiveness. It addresses every class of asset regardless of criticality, and every type of asset challenge – whether performance, process, emissions, mechanical, or instrument. It also allows you to bring the power of AI to every one of these assets and every one of these challenges – but in a way that differentiates itself in numerous ways.
In this article, I have looked at just one of those differentiators – case management – but there are at least five other ways that Cordant™ Asset Health Analytics sets itself apart and you can read about those in our infographic and in our fact sheet. You can also contact your nearest Baker Hughes sales professional for a guided tour of the entire Cordant™ solution – including Cordant™ Asset Health and Cordant™ Asset Health Analytics.