Actions
Engagement
  • Author Author: cstanton
  • Date Created: Date Created
  • Last Updated Last Updated:
  • Views 2804 views
  • Likes 1 like
  • Comments 24 comments
Related
Recommended

What's Stopping the Internet of Industrial Things?

Is it even Needed?

There, I said it, is that controversial enough? It might be. You have likely heard of the Internet of Things (IoT) by now and what it is, "let's connect everything to the internet! so we then know what it is!" and my suspicion is, that it is not necessarily needed. Or perhaps the concept is just too big to comprehend how it applies right now, let us think about and discuss this, I suggest first we start small.

 

Starting Small

For a lot of people the first thought that comes to mind with IoT is making a smarter home via home automation. There is nothing more awesome than feeling in control and being informed whether your plants are in need of a water, if your home is secure or keeping a close eye on your front door via a camera when Amazon ring the doorbell. We have even read that a controlled sprinkler system allowed a home owner to save their premises from being burnt down where their neighbours suffered great losses. The success stories are slowly pouring in, we are surely better connected, are we not? We can even look after our pets. Though for every success story there is often a horror tale, but we shall not talk about cars (or Die Hard 4 and its hacked power plant).

 

However, when we are looking at the scale of the home, we have an isolated scenario, a sandbox, a location where frankly we can do whatever we like. Here in begins to lie the crux of the scale of IoT and its increase to a level where it could be considered to be applied to, let us say for example, a humongous steel production plant.

 

Thinking Big

I was introduced to industrial environments from a young age, my father being an industrial shift engineer working on the huge and heavy electronics of British Steel. Regularly I would hear tales of trudging up long ladders, getting covered in extra fine coke dust and crud, to take care of the electronics for a conveyor belt system that would put a shopping till to shame, and motors the size of a large living room with junction boxes the size of a human. So, too, came the tales of blast furnaces and metal working, but also of the systems that controlled these. From the computer systems that people sat at to systems which you would not hear about from home.

 

Programmable Logic Controllers (PLC) are the electronics that are the heavy duty, industrial armoured electronics against the elements and truly harsh conditions. Your typical Beaglebone Black would not cut it where a PLC can do the grunt work, (though with the industrial BBB we're getting better). This is what I see as the type of system which the Internet of Things is trying to replace or supplement in an industrial environment, and it is not going to be cheap to do so.

 

So just how do you go from a system that is setup to use PLCs to one that wants to be connected to the internet? Or is it already on the internet? Unfortunately for this information, my father retired from Corus as it was then known, before becoming Tata Steel and did not see what upgrades were done to the system.

 

However, I suspect that the engineers of this community may have seen similar or even more.

 

What is Stopping the Internet of Industrial Things?

You tell me! You do not have to be specific, I suspect that a number of people are under Non-Disclosure Agreements with their company, but let us be general about this shall we?

Parents

  • Well, my first thought upon reading this post was -- [Industrial IoT] "not needed" is a curious notion, especially for a community site of engineers who are creating the embedded edge devices that will foster the global adoption of a most important technology --- one that promises to impact the lives of human beings all over the world in a positive way.

     

    So, is the Industrial Internet of Things needed? Well, a few years ago there was a massive explosion at an oil refinery in Texas. I believe it killed 30. Given the cost of rebuilding the plant, the subsequent litigation and the loss of  loved ones and breadwinners of their families, in hindsight, I bet the plant wished it had an IoT-enabled gas monitor that could have alerted, in real-time the presumably increase in explosive gases. Sure, they probably had gas monitors, which were obviously not good enough.

     

    When I was field engineer years ago for Eaton Electric Drives Division, I went to work at nuclear plants, wastewater treatment plants, steel mills, meat-packing plants, corn processing plants, pharma plants and countless other hot, dirty, smelly and dangerous facilities. I recall once going to a steel mill that was losing $150,000 every time power was lost from the tap-switching (SCR) power conditioner I was sent to repair. The managers were not very happy when I arrived and were already threatening lawsuits. It would have been nice if that steel mill had had an smart-power monitoring system.

     

    Talk to an product liability attorney (or a victim of a product failure) and you will quickly learn there are problems with product quality that were never caught despite the fact that these factories have quality certifications such as ISO9001 or Six Sigma.

     

    Yeah, I'd say  Industrial IoT is needed. But what exactly is it? What can it do for you? Well, let's run through a basic use case example. I've provided a basic diagram of this Industrial IoT system. In this use case, the company wanted to predict failures of components even before anyone noticed a problem. They wanted to avoid warranty repairs, product recalls, lawsuits from product failures and most importantly avoid damage to their reputation and brand.

    image

    In this Industrial IoT system, the Internet-enabled edge sensors transmit massive amounts of data wirelessly to a gateway. The gateway uses and IP-based connection to the Data management system, a Hadoop-based Data Lake. The data management system crunches the numbers and processes it through the data analytics to give insights.

     

    Insights are what separates an Industrial IoT system from a traditional PLC-based, P-2-P process controller. The insights provide the quantitative analysis where decisions are made and business changes are implemented.

     

    While this engineering community is more or less focused on developing the embedded edge sensors, it is the data lake and the data analytics that allow predictive analysis and insights that otherwise would not be possible.

     

    Now given this example and the framework I have introduced, What are the Obstacles or impediments to the adoption of the Industrial Internet of Things. [BTW, nothing is stopping the adaption; impeding is not stopping.]

     

    You need to get the sensors right to collect the data; you need to transmit it in real-time since process controls cannot tolerate a lot latency; you need a data lake; and, finally, you need the right data analytics to determine the Insights. And you need a forward looking management team that will be willing to implement the change -- perhaps the biggest obstacle.

  • in reply to rscasny

    I bet the plant wished it had an IoT-enabled gas monitor that could have alerted, in real-time the presumably increase in explosive gases. Sure, they probably had gas monitors, which were obviously not good enough.

    Would having these 'obviously not good enough" monitors connected to IIoT help.?

    Was the failure in the sensor or the control system.?

     

     

    I'm not sure that having important infrastructure connected to the internet would make anyone happy.

    Given that hackers seem to have little restrictions, imagine the effect of someone was targetting the system for terrorist type activity.

     

    Mark

Comment
  • in reply to rscasny

    I bet the plant wished it had an IoT-enabled gas monitor that could have alerted, in real-time the presumably increase in explosive gases. Sure, they probably had gas monitors, which were obviously not good enough.

    Would having these 'obviously not good enough" monitors connected to IIoT help.?

    Was the failure in the sensor or the control system.?

     

     

    I'm not sure that having important infrastructure connected to the internet would make anyone happy.

    Given that hackers seem to have little restrictions, imagine the effect of someone was targetting the system for terrorist type activity.

     

    Mark

Children
  • in reply to mcb1

    Hi Mark.

     

    Well, you raise two issues:

     

    • The effectiveness of IoT based control systems vs. non-IoT based / standard gas monitoring systems
    • The security of critical  infrastructure

     

    Both these issues are very important, but separate considerations.

     

    W would need to have the blueprints of the placement of the current gas monitors (upstream, downstream, throughout the plant) and what their set points were and the hysteresis of the sensors and how they are connected to the alert control system to really answer your question. I think the Internet provides a means to not only provide more edge sensors at more strategic locations throughout the plant, but the key element is the Big Data management system that receives the sensor data would be more predictive in determining when a problem of an abrupt increase in explosive gas concentration is occurring

     

    As far as the security of critical infrastructure....this of course is a big concern and not just for oil refineries. I'd suggest you read Mission Critical magazine and Critical Infrastructure mag. This article from Control Engineering mag is a good primer on the topic:

    http://www.controleng.com/single-article/cyber-security-in-process-plants-recognizing-risks-addressing-current-threats/a…

  • in reply to rscasny

    I think the Internet provides a means to not only provide more edge sensors at more strategic locations throughout the plant

    So does the piece of wire running around it.

    Many of these plants require intrinsically safe wiring, and using RF solutions (yes WiFi is RF) may not be an option.

     

    I'm not sure about this 'data lake' business.

    A gas monitor (to use your example) is there to monitor the level of a gas, or gases, and when the level reaches xx, then the system reacts.

    Sending all this into a 'data lake' doesn't improve the detection aspect of the sensor.

     

    Thanks for the reference but I work in a fairly secure environment and I'm fairly aware of the risks.

    We don't even allow our two main networks to interact with each other, let alone the outside world.

     

     

    Mark

  • in reply to mcb1

    Mark,

     

    Sensors and the Internet are only one piece of the big picture of Industrial IoT. The data lake really makes it all happen and offers much more than a better control of a machine. The Industrial Internet of Things is really meant to obtain meaning from huge amounts of data to drive the decision making processes in ways that could not be done before.

     

    If someone only wants to control machines with the Industrial Iot, I'd say, "why bother"-- one would be wasting the true value of the IoT technology and spending lots of money. But if you are proactive and are interested in predictive analytics, in other words, see trends of potential breakdowns or defects before the symptoms even become obvious, then the power of Industrial IoT is worth your time and money.

     

    Here's what GE Aviation is doing: it does thousands of jet engine overhauls. In the past, the analysis was done with the plane on the ground with an engineering team. Anyone who has spent any time troubleshooting systems, knows the troubleshooting can be slow. But now GE is using a data lake to automate the analysis. GE got 25 airlines to stream data to aggregate terabytes of full flight data with a data lake to decrease the analysis time and see problems/defects before they become headaches. Sure, it's a work in progress, but it's the road to a big pay off.

     

    I'm an optimist. I see a lot of potential with Industrial IoT. Maybe it won't happen tomorrow. It may not event be as big as Gartner predicts. And there are plenty of challenges to work out (e.g., security, latency, reliability, etc.) but IIoT is new and test cases are rolling out and being implemented. I think it's an exciting time and a technology with immense promise. Keep the faith.

  • in reply to rscasny

    Sorry but I think you are confusing the purpose of two totally different scenarios and trying to use it to promote another.

     

    For GE they have lots of parameters and variables, not just a couple of sensors that require action when a threshold is reached.

     

    The end game could be to reduce the cost of overhauls from a strickly by the book hours/takeoffs to a dynamic model based on true workload of an engine.

    In this case there are millions of dollars to be saved from the overhaul costs alone (materials + time out of service).

     

    I'd hate to think that they would be ignoring a high temperature just because x other aircraft also experience it.

     

     

    We run a similar system here in NZ where we save airlines millions in fuel, reduce delays and CO2 emissions.

    In this case many variables are collected or already known and the system does some analysis to make corrections to the flight during cruise to match the desired end target.

    In our systems we also collect lots of information that can be accessed to determine xyz.

    We just don't use a fancy name ie 'data lake', we call it System Health Processor.

     

     

    I'd also like to suggest that the article is a PR page and has no real information about how they are storing the data so you can actually know that its the engine temperature you're looking at, rather than the onboard battery system.

     

    I'm sure that eventually someone will decide that connecting everything to the Internet is a great idea, and it will solve everything.

    Given that many years ago people said we would be talking to our computers and mice and keyboards would not be needed, and I've yet to see that, I'm not holding my breath on this IIoT solution.

     

    Mark