IIoT

Maybe this forum is not the best place to ask if IIOT is coming. I think maybe we should be asking Allen Bradley, Omron, Siemens, Etc., and see if they are not developing products that will function with IIOT technologies.

Those people will be who pushes this industry down what road?

I just think we all need to be prepared for it.
 
But ANYTHING on the net, and I mean Anything can be hacked.

Agreed. Thus staying off the net is the best defense.

Of course, "abstinence" probably works as well for control systems as it does for teenagers. Even with isolated systems -- dedicated laptops (no exceptions for vendors / contractors), no USBs that haven't been thoroughly vetted (preferably with Format E:), there will still likely be vectors.

People are notoriously sloppy. That's why we have PLCs.
 
Last edited:
I remember that in the decade of the 90s it was commented that in the near future everything would be connected to the internet, now 20 years later they have given it a name.
 
My impression so far is that IIOT is a solution that is looking for a problem to solve. It seems to have very limited value and some ugly side effects. I would rank it up there with the cell phone connected appliances as an open invitation to hackers.
 
IOT does not bring anything new to this.

I think most of what it comes down to is that people are pointing to these ideas that have been around for years (but relatively uncommon) and saying "Hey Everybody! In the next 5 years we should all start doing these things!", and then they give it different names.

Any time I hear IIoT or Digital Enterprise or Industry 4.0, 50% of the time it's marketing and 50% of the time its a journalist trying to look knowledgeable. HOWEVER, that doesn't make a lot of the concepts bad ideas, even if they aren't new ones.

I can think of a whole bunch of projects I've been attached to where something like a Digital Twin could have helped save a lot of project risk. If you can simulate your PLC and the robots and the rest of the plant, then you can test everything before you install it, instead of programming on the fly and hoping it all works. More cost up front, requires WAY better project menagement, but less risk of delay at the end.
 
About the only thing I see in those jargon-filled documents that hasn't been available to savvy users for years is "predictive analytics". I have an idea what that means, but is anyone offering a product that can mine a historical database and make those predictions? Are there any case histories available of facilities that have adopted any of these techniques?
 
You guys have too much technical knowledge, that's the problem.

To me a smart phone is nothing but a small computer with some RF components and a touch screen.

Facebook is just a some webservers connected to a user database.

Twitter is just a message database with a webserver frontend.

IIoT are just industrial equipment hooked up to the internet.

You just got to put a name on something so you can market it and the non-technical people become afraid that they are missing out on something they don't understand so they throw money at it and a market has been created and somebody makes more money. And everybody is happy.
 
One of the few benefits of having floated around for a long time is that my perspective on "new" "exciting" developments is the repetitive nature of them.

How may of you remember the hubbub about MAP back in the 80s? MAP (Manufacturing Automation Protocol) was the "IIoT" of its age, the grand unifying theory of all things industrial talking with all other things. It was developed by a consortium of GM, IBM and Boeing IIRC, and was WAY better than that new upstart, Ethernet. "Yessiree Bob, this is going to be the end-all / be-all to every industry..." Fzzzzttt.... gone by Y2K!

Then it was DeviceNet... "Yessiree Bob, I gotta tell you, THIS is going to be it..." Oh... never mind...

Then it was Profinet (if you were anywhere but the US). "Jahwolee Hans, zis is goink to solve ALL of zee problems mit DeviceNet..." Nien, it didn't...

Then Rockwell countered with ControlNet... "Yesiree Bob (again), he were are..." Oh wait, never mind again...

I will say though that Ethernet seems to have outlasted all of them and as it approaches the nirvana of true determinism, it's making things seem a lot more real this time. The ability of a network to manage firmware and programming changes in its connected devices starts to make the detractors of using complex devices become outweighed by the benefits. And lately I've been seeing predictive maintenance features being added to new designs that takes advantage of all this communications capability. I just looked at some features in some new VFDs that will PREDICT when a cooling fan is about to fail, warning the operator of it (via Ethernet) so that PREVENTATIVE action can be taken BEFORE an unscheduled shutdown takes place. THAT is where the rubber starts to hit the road in my experience. I also just implemented a system wherein a pump motor run by a "smart overload" is reporting shaft kW, then in the SCADA system we are comparing that to flow out of the pump and if flow/PU of shaft power starts to deviate, we can predict when the motor bearings are going to need to be replaced. Again, TANGIBLE benefits. It's getting there, it's just going to take some edumacation of the workforce.

And that's never a simple task...
 
I looked up MQTT. It doesn't look very efficient.
ProfiNet was a huge help in my opinion. It was so much easier to use than AGSEND and AGRECV and Profibus. Very few of my customers could make the older protocols work without a lot of hand holding and support. That is why I hated Siemens so much back then.

I remember being at a IPC show at the Cobo Center in Detroit in the early 1990s. MAP was being pushed as the next big thing and if GM was doing it all the vendors had to also. I think MAP faded away long before 2000.
 
IIoT without Industrie 4.0 is like an engine without a car and without a driver.
Sure if all you want is a rotating shaft, an engine is perfect for you! I think some industries have loads of very simple sensors and actuators that IIoT is a nice way of managing them. Can it be done with SCADA? Of course it can, we have done it in the past. It was probably more expensive per sensor, but I guess "cheap feature rich scada" is harder to sell than just changing its name. "We have made it 80% easier to download firmware updates." is still a pretty cool thing.

Industrie 4.0 talks about a lot of things besides just connecting data from disparate sources. It talks about analysing that data. You have to set up something in your business to analyse your data and make business decisions.
It talks about making machines more versatile so they can do more than one thing. To me, that sometimes makes sense and sometimes doesn't. If I want to make M8 bolts, then an M8 bolt making machine will be cheaper and faster and easier to use than a machine that can make M1.6 to M250 bolts. And Steve who struggled at school but tries really hard will be able to fix my M8 bolt machine, rather than relying on people that frequent IIoTalk.net.

Industrie 4.0 goes further, and suggests streamlining your ordering system so that a customer orders something from your website and in seconds it is scheduled in your machines part production system. There's possibly some IIoT used to achieve this? Oh and maybe your machine can send a message to your customer saying production has started and temperature sensor 3 is at a balmy 64°C.

Put some vibration monitoring on there. Compare it to other similar machines and you have a decent predictive maintenance maybe? Is the hard part doing the engineering on where best to put the vibration sensors and what it means? Can BigData concepts maybe handle the engineering for you? Maybe, but a few minutes with a pencil and a calculator at the moment is a lot cheaper than cloud connected vibration sensors with AI analysis to offer optimal predictive maintenance forecasts. Sometimes breakdowns are cheaper.

My favourite Industrie 4.0 concept is the digital twin. Not really a new concept, but it suggests using the historical data from the physical system to improve the simulation accuracy of the cyber system. Good idea. Almost like common sense.

These things though, they are all long term. You will have to invest in feasibility before you make a penny. Invest in design before you make a penny. Strategise. Still no pennies.

In conclusion, IIoT offers no technologies that we didn't have access to before. If you have long term thinking and long term investment to back you, you can possibly be a very versatile producer of parts, which can make you more resilient to changing markets. However, an M8 bolt making machine can spit out a lot more bolts than a 3d printer, and you can service it monthly and have no issues. M8 bolts are not going to go out of style anytime soon. A 1/2" bolt machine though, people might give up on them tomorrow and you will have wasted your money!
 
About the only thing I see in those jargon-filled documents that hasn't been available to savvy users for years is "predictive analytics". I have an idea what that means, but is anyone offering a product that can mine a historical database and make those predictions? Are there any case histories available of facilities that have adopted any of these techniques?

Yes predictive analytics exists and is very popular now..
To name a couple...

GE "Predix"
https://www.ge.com/digital/predix-platform-foundation-digital-industrial-applications

Wonderware "Insight" or "Online Insight".
https://www.wonderware.com/industrial-information-management/online/

I am sure there are YouTube videos on them..
They use their historian to push tags to SQL for data analysis (using the "Predix or "Insight" software addins and reporting) for changes over time.

Both of these can be installed locally (with their own Historian and SQL), but they are touted to have way more power in "The Cloud". Due to the fact that more data is available for comparison. HOWEVER....this involves a healthy yearly fee ($10,000 range....minimum) increasing the more tags you want tracked.
 
Last edited:
Usually when you see MQ it means message queue. I think it originated with IBM MQ Series. Microsoft has MSMQ. I've used MSMQ a fair amount. I didn't like it. The queue gets corrupted and it can be a mess, especially if it's running on a production server with other services. You may have to bounce the server. If you are interested in enterprise wide data, check out technologies like WCF, and SignalR. I know, WCF is dead right? No it's not. If you want to move data around the factory that is not HTTP, in other words TCP, Named pipes MSMQ then WCF IS good. It makes moving data between things easy. SignalR is pretty awesome. With it you can now implement a pure web based HMI without Ajax polling. It's publish subcribe and can push data to a web page or many web pages. WCF can also do publish subscribe. For HTTP Services there is stuff like web API, which I know very little about. Anywho, separate you requirements between HTTP and the good stuff. :)
 

Similar Topics

I found this article in today's news interesting ... as if we don't already have enough to worry about these days ... headline: here's a link...
Replies
5
Views
2,381
On with the dawn of IIOT and the linking of legacy systems to the "cloud" what is the opinion on Edge Of Network Connectors, These are typically...
Replies
10
Views
4,045
I'm having trouble finding an OPC Server besides the OptoOPCserver that will communicate with the LCM4 controller. I know the LCM4 is an obsolete...
Replies
2
Views
1,827
This is a GREAT read about MQTT, I wanted to share it.
Replies
1
Views
1,450
Hello, Can someone please provide a difference between a Standard SCADA control system (Sensors/Controllers/SCADA Servers...) and an IIOT based...
Replies
9
Views
2,705
Back
Top Bottom