I think one of the biggest changes in predictive maintenance in recent years, William K is that sensors can now monitor the condition of individual parts to determine the actual condition of those parts.
Predictive maintenance, the process of doing repairs when a system starts to fail instead of either on a schedule or when the failure stops operation, is not a new concept. For many years, at least in some plants, machine operators and the maintenance group would communicate frequently about machine performance, and the result would be that when things drifted "just a bit", or when a machine made a different sound, the mantenance group would be made aware of the change, and could take the corrective measures during the normally scheduled maintenance period between shifts. Of course this happy mode required a lot of "team spirit" as well as operators familiar with their machines.
Today that condition is mostly not available for many reasons and so the control systems and the new added monitoring systems must do the same thing that people used to do. Alomost as good as the old system, possibly more sensitive than the old system, and probably the only way to go in the fully automated industries. But not really a new concept.
I think we already have much of this concept in many automobiles today. We just don't have the interface and the flexibility to check all the data that is available, as we may want to. We simply get what the car comes with. My car now tells me how much life I have left in the oil. It's not just mileage that it uses any more. It monitors the type of driving style, engine temperatures, duration of trips, and various factors a modern vehicle has sensor data from. Depending on the conditions it has ranged from 8 to 15,000 miles between it's recommended changes. This may still be considered predictive, since it's not directly testing the oil, but there are other sensors that directly measure performance on today's vehicles. Mine tells me what tire need air, and if the engine needs service. You just need to plug in the interface to know the details on what service is exactly needed. And most of this it does long before a actual breakdown occurs. It amazes me how much simpler modern programming is making maintenance today. I had a Danish supplier for a large process machine email me to check a valve on our machine back in the 90's. But that was based on a dial up connection, and periodic monitoring of performance by one of their technicians. The article made me smile and think how much they would have loved this interface!
Rob, this is a great example of the Internet of Things (IoT) and predictive analytics being applied to industrial machines. Your comment about software at the end is instructive as well. The thing about software is that it can be amortized across a large number of customers. This makes it cost effective to put a lot of effort into the software.
I did R&D for a custom system like this many years ago for a simulator manufacturer. This was very specific to the manufacturer, but worked across all their products. The system would tell you where there were potential failures very early. This saved lots of money.
What Siemens is doing here will have a great impact. It brings in big data, PLM and other technologies. This will allow it to become better over time. Very impressive.
Thanks for this, Rob--I knew about the shift to predictive maintenance but hadn't caught up yet with condition-based maintenance. Sounds a lot more efficient. What exactly are the hardware monitoring devices that supply the software with data? Cameras? Sensors? Both?
Engineers at Fuel Cell Energy have found a way to take advantage of a side reaction, unique to their carbonate fuel cell that has nothing to do with energy production, as a potential, cost-effective solution to capturing carbon from fossil fuel power plants.
To get to a trillion sensors in the IoT that we all look forward to, there are many challenges to commercialization that still remain, including interoperability, the lack of standards, and the issue of security, to name a few.
This is part one of an article discussing the University of Washington’s nationally ranked FSAE electric car (eCar) and combustible car (cCar). Stay tuned for part two, tomorrow, which will discuss the four unique PCBs used in both the eCar and cCars.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.