Ron Monday said yesterday that the big thing he was having trouble explaining is that the xCoupler is an appliance, not a software system. It is like a hardware router or firewall, not a server. People keep saying, “It can’t be that simple.” But it is.
This brings me to another thought I’ve had increasingly. I think that sometimes we overengineer the bejayzus out of things.
For example, Emerson CSI’s Machinery Health Monitor. Understand me, I think it is an impressive device, and for some applications, especially if you have PlantWeb as your DCS platform, it makes sense. But at $7K a whack, it is way too pricy to use for all but the most critical motor trains. Same for GE Energy’s offering, etc.
The mass market is in the ability to monitor vibration, temperature and pressure on any motor-pump combination and do it for under $1000 per point. The only way to do this is with some sort of wireless system, running on a mesh network. Currently, Zigbee, though massively flawed, is the best of the lot for this kind of thing.
Yes, you say, but what you are paying for is the artificial intelligence built into the device, so that it gives you machinery health updates instead of just lots of data.
But here we can do another “think around the problem” bit, like Ron Monday did.
As far back as 1995, John Boland of Visibit Corp. (http://www.visibit.com) was a experimenting with pattern recognition software for dealing with large volumes of sequential data, like machinery health. His really cool product, Visibit, was a software only product that:
— “automatically and continuously represents data in images, archives the images, then retrieves and displays them for users
— “presents large quantities of data as images because humans are hard-wired for pattern recognition
— “instantly highlights trends, problems, and discontinuities that tabular data cannot present
— “sits outside your databases as a data mining tool
— “can process data from any business process, from machine control to stock market analysis.”
(from Visibit literature)
So help me understand why a simple wireless datalogger with three or four sensors, coupled with data imaging software like Visibit couldn’t do an excellent job of doing machinery health monitoring?