The Connected Factory and More: 5 Examples of How IIoT is Changing Manufacturing
Isaac Maw posted on June 15, 2018 |

What’s the Promise of the Connected Factory? 

There’s a well-known old parable about some blind men who, upon meeting an elephant for the first time, try to appraise it and figure out just what exactly an elephant is.  

One man feels the elephant’s trunk, and says, “Oh, an elephant is like a big snake.” 

One touches the elephant’s flank, and says, “An elephant is a type of wall.” 

The next feels a tusk, and decides, “Elephants are like spears.” 

The next grabs a leg, and says, “An elephant is like a tree.” 

You get the idea. Because the blind men have limited information available to them, they’re unable to form a complete concept of the elephant. Furthermore, if any of the men tried to make decisions based on their concept of the elephant, it could have catastrophic results.  

To keep this metaphor brief, the elephant is your business, and the blind folks are your executives, department heads, directors. Without a system of providing high-level, detailed insight into what’s going on, it can be difficult or impossible to make big-picture decisions effectively and efficiently. 

With the paradigm of the connected factory, the parable is a little different. The elephant can turn to the blind men and tell them exactly what he is and his entire life story.  

So, with the corny introduction finished, this article will walk through some of industry’s best use cases for industrial internet technology. IIoT technology is a key part of the fourth industrial revolution. For large multinational corporations, it’s an essential part of doing business. For small and medium enterprises, it may not be necessary, but a strategic application can find cost savings and improve profitability.  

Application 1: True Predictive Maintenance 

Predictive maintenance is one of the most disruptive applications for the internet of things in industry. However, because of its potential, the term has become somewhat of a buzzword. If you walk by your conveyor motor and your eyes start to water with smoke and ozone, saying, “Uh… I think this motor is about to quit,” does not count as predictive maintenance. True predictive maintenance boils down to machine learning analysis, using as many sensor data points as possible, such as vibration, temperatures, currents and voltages. Some machine learning algorithms can accurately predict failure as far as four months or more in advance.  

“Companies always want to buy the three zeros,” said Mike Brooks, senior director of asset performance management at Aspen Technology. Brooks told me about Aspen’s work in predictive maintenance solutions. “They want to basically buy zero safety incidents, they don't want to hurt people; they want zero accidents and environmental incidents, and they want to have zero breakdowns. And they always go in that order.”  

Predictive Maintenance vs. Condition Monitoring: What’s the Difference? 

Sensors on your equipment can check machine health data points such as temperature against what’s essentially a rules engine, and give you warnings or make process changes based on those rules. However, what this type of system can’t do is tell you why or when the system will cross over from ‘go’ to ‘no go’. The idea of predictive maintenance systems is to build accurate probability predictions on the data, rather than simply reporting it. Brooks said it best: “It's like going to the doctor. The doctor doesn’t just look at the stethoscope, the doctor looks at the patient history to understand what happened, how you were prescribed, what the results were. We do exactly the same with the machine. You can't do it just from the sensors, you have to understand the history.” 

Profit Potential of Predictive Maintenance

 The possibilities for this technology abound. For example, a manufacturer could hand over full control of maintenance part inventory to an automated system. Rather than keeping a stock of replacement parts, a predictive system could predict when a component was going to fail and place an order for the replacement part so that it arrives in time for maintenance crew to replace the component during a scheduled downtime, virtually eliminating the well known “Make it happen!” drama of unplanned downtime. That’s just one hypothetical example of what can be done. Many more possibilities abound. 

Predictive maintenance technology links back to my introductory analogy of the elephant: when you have access to the big picture, you gain full control. Brooks gave another example: “When you tell the operator that you've got 45 days until a compressor fails, the first question he’s going to ask is, ‘well, what do I do?” The obvious first step is to decide whether to interrupt production now or later. If production demands require it, you may need to run the compressor to failure, and be ready for a replacement. Otherwise, it may be possible to preventatively replace or repair the compressor before failure. Various factors go into this decision, including market demands, safety, and other costs. 

However, Brooks’ example shows the utility of root cause analytics, which is a product offered by Aspen Technology. In the example of the compressor, root cause analytics can show you that the reason for the imminent failure in 45 days is insufficient cooling. The insufficient cooling is leading to water droplets in the compressor, which is leading to pitting and deposition on the blades, which is leading to vibration damage, which will ultimately lead to failure, in 45 days. By identifying the root cause, the problem can be addressed, possibly eliminating or delaying that failure.  

Application 2: Controlling an Operation Remotely 

Supervisory control and data acquisition (SCADA) systems are widespread in the power generation, oil & gas and telecommunications industries. These systems are designed to bring control over valves, switches, and other field devices to a central digital system. But SCADA systems have been in use for decades. There's untapped potential behind the idea of central control—and the IIoT is the key. 

I spoke to Dan O'Brien, Director of Strategy and Marketing, and Elinor Price, Sr. Product marketing manager at Honeywell Connected Plant, the division of Honeywell specialized in industry 4.0 and digital transformation. They filled me in on some more powerful examples of remote control. 

In the oil and gas industry, assets are often widely spread across a geographic area. According to Price, network-connected sensor data means, "Being able to actually see what's going on, in real time, across all your assets. Being able to have that 'enterprise view' across all the assets, understanding where there's a problem, where you might need to direct maintenance, and where things are performing well." 

Connected, monitored pumping stations and pipelines have clear benefits: reduced need for maintenance personnel, as crews wouldn't need to waste time making the rounds to functioning equipment. Reduced risk of environmental accidents and reduced unplanned downtime are other potential benefits. 

However, there's still more to the story. "Oil and gas is not only on land, but also on offshore platforms," said Price. "How do you support the platforms offshore, and how can you do that more effectively with lower manpower offshore, from a remote-control center onshore? You can get the best skilled workers, because they may not want to be living offshore, you can have that expertise available across any of your remote assets. So that's something that we have customers around the world doing today with our solution."  

With connected devices on board an offshore platform, landlubber subject matter experts can communicate with offshore workers or even operate controls remotely.  

Considering these exciting possibilities, I asked O'Brien about the state of the field of IIoT and the connected plant. "Nascent is the best word for where we (as an industry) are today," O'Brien said. "We're just scratching the surface of how we're going to figure out the business benefits from all the information, all of the connectivity, including control, in the future for devices and information. I think IIoT, from an industrial perspective and probably from a consumer perspective as well, is so much in it's infancy. And that supports everything we're seeing at Honeywell Process Solutions," said O'Brien.  

"I would definitely second that," said Price. I think a lot of people are sort of window shopping and scratching the surface of what potentially could be massive."  

"There are miles and miles to go. There's lots of room to run," added O'Brien. "But notice that neither of us said that this is hype, that this is theory, that this is whatever. It is being adopted. I guess that's the key part. It's not just hype or theory anymore." 

Application 3: Improved Internal Collaboration 

Next, I spoke with AVEVA, a multinational engineering and industrial IT company. AVEVA is focused on driving digital transformation in their clients to improve profitability, efficiency and performance. I spoke with Matt Newton, senior portfolio marketing manager, asset performance at AVEVA.  

In many companies, the phrase "the left hand doesn't know what the right hand is doing" is too true. The production department doesn't know what's coming down the pipeline from sales. The sales department doesn't know what the marketing department will have for them next quarter. The IT department doesn't know about the new project coming down from the boardroom. It's a story ubiquitous in corporate culture, and everybody gripes about the inefficiencies, redundancies, and miscommunications that can arise from this siloed approach to business.  

But wait: Isn't the IIoT all about connecting assets to gain visibility, then control of the big picture? 

Some organizations have had success, such as Roy Hill Mining, as documented in this video prepared by AVEVA:

Granted, Roy Hill was built from the ground up to incorporate digital technology and an integrated operating model. It's possible to make the argument that starting from scratch is easier than upgrading your existing ways of doing things. However, it's pretty amazing what Roy Hill was able to achieve with help from Schneider Electric's industrial software, which has now merged with AVEVA.  

Rather than allow each department to operate in silos, stifling collaboration and data access, the company utilized IIoT technology to maximize the collaboration and efficiency of their control center. 

The company operates from a remote operations center, located 1300km from the mine. The remote operations center features a layout designed to match the entire end to end value chain of the company, "from pit to port." The right-hand side is where they control the mine, and the left side controls the port and shipping. This means everyone is sitting next to team members in control of related processes. The room is also organized from front to back, with more time-sensitive roles in the front. 

Roy Hill understands the value of having maximum access to the data needed to execute intelligent decisions. According to Roy Hill, the company uses the digital twin, machine learning, and augmented reality technologies to lower the barriers between workers and value-added work.  

Newton described it this way: "IIoT technology gives you opportunities to look at different processes or different departments operating in silos and find ways for them to collaborate to create more efficiency and value. In some cases, it's not just about what's happening inside the business, it's also about what's happening in real time with the market. So, in the case of Roy Hill, we've helped digitize their entire value chain right from where they're pulling iron ore out of the ground, to shipping it at the port, to looking at what the market demand is, and really tie that whole digital value loop together. They can adjust operations in real time to drive their business to generate a better return on our investment, and improve profitability as much as possible, all based on real-time data that they're getting from the market. It's really interesting to see what advances have already been made possible." 

Application 4: Artificial Intelligence 

Yes, this A.I. stock photo is pretty rediculous
Yes, this A.I. stock photo is pretty ridiculous
It's impossible to have a conversation about Internet of Things without bringing up machine learning and AI. All the experts I spoke to for this article agreed that AI will play an important role in the IIoT tools of the future. However, there are some misconceptions about AI today. To learn more, I spoke with Dr. Richard Soley, Executive Director of the Industrial Internet Consortium (IIC). Dr. Soley has experience developing neural network systems, which are now commonly known as machine learning systems. In other words, he knows the state of the AI and machine learning field inside and out. 

The terms artificial intelligence and machine learning are often conflated. However, there is a critical difference. "Machine learning cannot tell you why it did what it did. Machine learning systems gulp in data and statistically correlate results to inputs," explained Soley. "AI systems can tell you why they did what they did, and what rules were followed. Machine learning systems, all you can do is feed in more data and try to correct it."  

Machine learning systems are commonly used to deal with the massive volumes of high-frequency data involved in asset monitoring. However, this 'black box' issue Soley describes above, where you can't see how or why a machine learning algorithm arrived at a result, can pose a problem when things go wrong.  

However, that's not to say that machine learning is simplistic or not useful. "The algorithms and heuristics have not changed much over the past 35 years, but the computers are a million times cheaper and faster," said Soley. "It's tremendously useful. It can recognize shapes or faces, for example. But when based on corrupt data, it can spit out bad results."  

This can be a problem in systems which intake human operator input as well as sensor input. If operators aren’t committed or able to input accurate data, the whole system can be thrown off course. Garbage in, Garbage out.  

Dr. Soley emphasizes that well-trained personnel are essential in IIoT applications, and indeed that a lack of skilled personnel is a major factor holding industry back from taking full advantage of IIoT solutions. 

"If you don't understand what the data is, no machine learning system is going to fix it for you. You need trained personnel to interpret the input data, as well as the results, to get good use out of machine learning systems," He advised. 

More Bleeding-Edge New IIoT Applications 

The Industrial Internet Consortium and its members are dedicated to developing new uses of IIoT technology. The testbeds are a major focus and activity of the IIC, and they can be viewed at this page.  

Application 5: Full Digital Transformation 

There are a few buzzwords circulating around the topic of Industrial IoT: 'Connected Factory', 'Industry 4.0' are two popular ones. Digital transformation is another, and its precise meaning can be confusing. To shed light on the true meaning of digital transformation, I spoke with Stephan Ihmels of Siemens Mindsphere.   

The company describes Mindsphere as 'a cloud-based, open IoT operating system.' Ihmels said it this way: "[Applications built on Mindsphere] can be as simple as integrating a PLC that has a connectivity device integrated, or a drive that has a connectivity device integrated, with an app which helps the Siemens customer to understand the condition of the drive. But it also extends to the digital grid business, the mobility business, where entire new business models emerge to support the products that are being built." 

Digital transformation is more than going paperless or replacing a clipboard with an iPad. Digital Transformation refers to the revamping of a business model to incorporate new digital technologies. In the competitive world of manufacturing, it's adapt or die. It's not enough to bring in digital technology as an add-on to your business as usual. Like Roy Hill Mining, true adoption of IIoT solutions could mean new organizational changes, new personnel, new culture, and even new processes.  

For more news and information about the Industrial Internet of Things, consider subscribing to our newsletter! 


Recommended For You