Used at scale, AI-driven systems that continuously monitor machine health can transform industrial maintenance — but like every new technology, there are hurdles to clear before their full value can be harnessed. In part 3 of this four-part guide, we tackled the challenge of changing your systems and processes to fully benefit from digital asset health monitoring. In this post, we’ll share our top tips on how to choose condition-based monitoring technology that will scale with you for the long term.
There is no single condition monitoring technology that will catch every developing fault on every machine you have. But if you install five complementary systems on every single asset, you’ll end up with more alerts than you can handle. Perform a criticality analysis, research the different condition monitoring technologies, then separate your asset base into categories based on what it is you’re looking to do.
For example, if you’ve got assets that are both highly critical from a downtime perspective but also really big energy users, it makes a ton of sense to go with an electrical signature analysis (ESA) solution, which can simultaneously address both the health side and the efficiency side. But ESA systems can’t catch every possible failure mode, so for such critical equipment, you may also want to install vibration, lubricant and/or infrared systems to monitor the components they’re most suited for.
For equipment that has a chronic single point of mechanical failure, the best choice would be a localized vibration sensor that enables specific insights into what’s going on at that exact weak spot in your setup, instead of a more catchall tool like ESA. (Also, you might ask yourself why this keeps happening.)
At the low end of the spectrum, there may be assets where there just isn’t a case for CBM; they’re simply not critical enough. Even in a fully mature CBM setup, there will be alerts that end up at the bottom of the priority list. If certain sites or machines routinely find themselves there, you should evaluate whether condition monitoring even makes sense for these assets. If you’re never getting to the point where you act on the system’s alerts, then a run-to-failure approach may be the best way to go.
As we noted above, you’re going to end up with a landscape of condition monitoring technologies that complement each other in order to achieve the target you have in mind. But you don’t want ten separate dashboards; you want to centralize all that data. You want the insights from your vibration analysis system, the insights from your lubricant analysis system, and the insights from your electrical signature analysis system to show up together and give you a full picture in one central, familiar place. They can only do that if you buy technology that’s designed to integrate with existing systems.
Tying back into our previous post on the process aspect: easily integratable systems will also require the least change to everyday business as usual, because they’ll only be adding new information into systems your teams are already using.
On a related note, don’t forget to look at your existing systems that are generating tons of data you just haven’t yet been using for CBM. Many condition monitoring platforms are able to analyze data you already have available, instead of or in addition to capturing new data with their own sensors.
A technology partner wants to help you achieve your best possible CBM system, not “trick” you into installing their system everywhere. Good suppliers don’t need to do that. They are confident in where their system works well, and they’re not threatened by other technologies because the landscape is big enough for them all.
Pick a supplier that doesn’t stop at just selling you the service but actually works with you to make sure the implementation is as successful as possible. A good supplier will be there for the whole journey, and will challenge you when they think you could be doing something more successfully than you currently are.
Look for suppliers that have a good footprint in your particular industry. The transition to a more CBM-driven approach is something your whole sector is moving toward all at the same time. It’s a great idea to share notes, but other companies in your industry are also figuring it out as they go. A tech vendor with experience in your industry has seen what works well and not so well at your counterparts, and they can use that knowledge to set you on the right path from the start, instead of having you invent the wheel once or twice over again.
(For a step-by-step guide to choosing robust condition monitoring technology, including worksheets you can use to assess suppliers, see our quick-start guide to predictive maintenance.)
In the previous post in this series, you added the procedural element to your implementation plan. Now we’ll expand the plan to address the technological side of the rollout.
- Go back to your stakeholder communication plans and add a section on where to install the new technology. Again, the goal here is to put the local team in control of their own implementation. Who knows better which machines are critical, which ones fail repeatedly, and what failures happen most often? Be specific about the scope (how many machines can they include?), the deadline, the format they should deliver (such as a spreadsheet listing assets) and what will happen next (are they deciding or just recommending?).
- Determine what resources each group will need during the asset selection process: information on each technology you’re rolling out, a criticality analysis tool, a help desk they can call with questions, etc. Add these to the resource section of each stakeholder communication plan. (Much of this will be the same across stakeholder groups.) Be sure to include a process for providing new resources as the need comes up. It’s key to make each group feel supported, not dictated to, so be sure that intent comes across in your wording.
- Based on each team’s proposal, update the timeline you created in the last round for rolling out the new technology. Will team A need more time than you initially planned? Does it still make sense for teams B and C to roll out simultaneously? Has team D shared technical issues that have to be resolved first?
- Here’s some good news: you’ve been making the plan resilient all along, by revisiting earlier decisions and adjusting them based on new information. Here’s some news that might feel less good: You’ll keep doing that for the lifetime of the project. No plan is ever truly complete.
- Use that fact to your advantage — by planning to adjust course as needed. First, decide how you’ll know you need to change something. How will you encourage and collect feedback? How will you measure employee buy-in and adaptation? How will you evaluate progress toward full rollout?
- Next, plan how you’ll use that information. This step is a meta-exercise; you won’t know what hurdles you’ll face until you face them. But you can already plan how you’ll hit the ground running. How often will the core team meet to discuss these metrics? Which executive sponsors can you call in to help?
- Last but not least: add in a sense of celebration. Have each team (including you, the core team) define milestones that you’ll celebrate, and how you’ll celebrate them. There’s nothing more motivating than seeing your efforts achieve their goals. Celebrating milestones makes teams much more resilient to setbacks, and much more likely to brainstorm solutions so they can keep the project on track.