Factory IoT Saves Intel $9 Million

jfruh writes Want a good way to sell someone a new technology? Prove to them that you believe in it enough to use it yourself. Intel has been trying to get customers to buy into the concept of the "Internet of Things," in which tiny distributed networked sensors would improve manufacturing processes. To prove its point, they implemented such a system in one of their Malaysian factories, and claimed $9 million in savings.

Comments

I only need three things in my Internet

[ArcadeMan]

And that's a theme park, with blackjack and hookers.

THAT IS NOT AN IoT CASE!

[Jeremiah Cornelius]

It's a LANoT.

Re:THAT IS NOT AN IoT CASE!

[UnderCoverPenguin]

Also, not new. I've visited several factories over the last 15 years. The systems on the production lines were all connected over ethernet running TCP/IP. Granted, 15 years ago, sensors and other small devices would be in clusters, each cluster connected to a (large) shoebox sized controller, but over time, as the networked controllers got smaller and less expensive, there were more controllers with fewer devices connected to each.

I suspect the plant Intel "installed" IoT into was just in need of a major update to its existing IoT (by whatever name it was called).

The Internet of Things, aka

[Rosco P. Coltrane]

the next giant leap in ubiquitous mass surveillance.

I just can't wait for all the devices that surround me to snitch on me and report all my life habits to their corporate or state masters 24/7...

Re:The Internet of Things, aka

[Mr D from 63]

This is just simple equipment monitoring with networked devices.This has been done for decades. Just a PR puff to get some brand names out.

Re:The Internet of Things, aka

[sl3xd]

It's just a toilet seat that reports when somebody's on it. Everybody poops! There's nothing to worry about!

Until you realize that it's able to find usage patterns, and your insurance rates go up because they think you may be getting colon cancer.

Everything's connected, and I don't want every facet of my life being reported to some corporate overlord.

Investment?

[Quantus347]

As somebody who designs networks of sensors and controls for manufacturing processes, I want to know what the investment was, and what payback period they are using to calculate those savings. Depending on the size of the plant $9 million might not even come close to covering that kind of mass retrofit.

Re:Investment?

[sandytaru]

I RTFA. They didn't go into specifics, but it seems to have been cheap WiFi point communications coupled with thermal sensors. So if a machine was running hotter than expected, they could stop that line and fix the problem before it broke completely and took that line down for a few days. So the "savings" could be what the cost of that line going down previously had been. Or something.

umm. details?

[supernova87a]

"...CPU tester modules in a semiconductor manufacturing line at the plant were retrofitted with sensors. They then sent data to Mitsubishi Electric C Controller gateway devices powered by Intel Atom chips. After some filtering, the data were then processed using software from Revolution Analytics. Putting the data results into practice resulted in a reduction in component failures, increased equipment uptime and productivity, according to Intel....."

Could someone who actually knows something about what they did write the fucking article please? I have no idea what was improved using this technique by reading these sentences which are the only concrete part of the entire story linked.

Internet of Stupidity

[PopeRatzo]

This story has pretty much nothing to do with the "Internet of Things" they are trying to sell us.

I seriously doubt that any of the WiFi sensors in Intel's machinery required an account with a third party company which then collected data on how Intel used their machines.

We already have an Internet of Things. It's called, "things".

Re:Internet of Stupidity

[Animats]

This story has pretty much nothing to do with the "Internet of Things" they are trying to sell us.

Right. It's ordinary industrial automation. It's also strange that Intel would have CPU testers that weren't networked and reporting to some machine aggregating statistics and looking for process variance. It's pretty much routine in factories today to network the machines. That's been going on since the 1980s.

The Mitsubishi C Controller mentioned is just a CPU board packaged as a Mitsubishi Electric industrial automation module for convenient mounting in industrial automation cabinets. "It includes two Ethernet ports, an RS232 port, a USB port, a CompactFlash card slot and a 7-segment display for debugging and diagnostics. The (Intel Atom) CPU comes with the Wind River VXWorks real time operating system pre-installed." It's programmed in C.

uh huh...

[Charliemopps]

I've been through these sales pitches before.
Ok Intel, how much did it COST to install?
Did you factor in that you sent in all of your Intel experts for free? And that you'll charge me $200 per hour just to ask them what kind of outlet to plug this into?
What was the volume of that plant? Is it producing $10million in product? Or $300 million? Scale matters.

$9 million in savings in a large production plants is shit. They have single machines that cost more than that. To take a gamble on a large change like this, the savings need to be insane. Cut my costs in half and it might be worth the risk. Saving $9 million when my costs average $300 million and, yes... that's nice... but its not worth the risk of new tech.

Re:uh huh...

[thegarbz]

$9 million in savings in a large production plants is shit. They have single machines that cost more than that. To take a gamble on a large change like this, the savings need to be insane. Cut my costs in half and it might be worth the risk. Saving $9 million when my costs average $300 million and, yes... that's nice... but its not worth the risk of new tech.

Actually $9million is $9million regardless how how you cut the pie. Just because a business turns over several orders of magnitude more means they should stop factoring in potential savings as *small* as $9million?

I ask you, what is a gamble? How are you gambling when you monitor your equipment? What is the risk when it goes wrong? Back to potential $9million outages? Oh calamity!

The only reason people are up in arms about this is because someone used the phrase "Internet of Things". If this article was started with "Lean Six Sigma", "Kaizen", or "a Continuous Improvement Project" no one would bat an eye. I am a reliability engineer and creating this type of monitoring is my day to day activities. Sometimes they pay off really well, sometimes they produce no benefit and we wasted a few $100k, but all of a sudden when someone says it's an "Internet of Things" project rather than project everybody shouts about risk?

Get a grip. Oh and I work for a plant that turns over approximately $4bn in product annually, yet if I could save $9m I guarantee there would be prizes, parades, and all sorts of untold honours directed my way. Never under-estimate how hard it is to squeeze the last bit of financial efficiency out of a place.