Bending the Curve on Industrial Energy Use

David Baltusavich, SynaTree

This article is a summary of the talk David presented at Erie Day of Code 2019, Entitled IoT Sensing In The Real World

Figuring out what to share with my fellow developers, designers, programmers, and other tech professionals in a public forum is never easy. Each of us occupies a niche, and the tools and techniques favored by one are not necessarily prescriptive for all.  Therefore, in deciding what would be worthwhile to share with an audience of my peers, I decided to keep the development details relatively abstract and focus instead on the opportunity, as I see it, that we have to really bend the curve on energy use in the industrial sector.  Reducing emissions of greenhouse gases and improving productivity is something that is near and dear to my heart, and I hope yours as well.

I come to this conversation after having spent a few years in the trenches with Industrial technology, specifically in developing an IIoT platform for an industrial client.  What I came to realize is that relatively simple improvements in industrial processes can lead to sizable gains in efficiency, productivity, speed, and ultimately profitability for legacy industrial producers and even relatively technically proficient users of industrial technology.

For example, according to the Department of Energy, a poorly tuned combustion reaction can waste up to 35% of the fuel being burnt.  This is a staggering number, easily adding up to millions of cubic feet of extra fuel burned per year in a large heat treating operation.  This fuel, in turn, releases millions of pounds of carbon dioxide into the atmosphere and exacerbates the climate crisis we know is rapidly coming to pass.

Keeping combustion operating at ideal levels is the purpose of the system I’ve been building with partner PSNergy.  The technology needed to do so is complicated, but it is approachable. And combustion is just one example of insidious inefficiencies that are 1) known to exist, 2) that have readily apparent solutions, 3) are largely not being addressed despite these facts.  If you talk with industrial businesses you will likely find other cases like this.

I don’t believe the source of the problem is a lack of computerized systems.  In fact, many industrial processes are guided by computerized systems such as PLCs.  PLCs are great, and they do what they are designed to do very well. I think the basic reason that we do not solve problems that on the surface seem fairly easy to solve comes down to the fact that PLCs are not designed, in general, to surface data outside of the PLC controller itself.  While many machines are exquisitely designed to record and act on sensed conditions and make adjustments in real time, the fact is that the tools available for PLCs to share that data outside of the machine are rudimentary and badly outdated.

Interfacing with a PLC need not involve directly connecting a laptop!

Interfacing with a PLC need not involve directly connecting a laptop!

Take, for example, a machine that exposes various internal measurements and conditions via the Modbus RTU protocol.  In principle this machine is capable of communicating important facts with plant managers, but unless somebody has taken the time to write a software application to purposefully read the data, it is effectively useless.

Other machines may surface data in CSV or TSV format, but without a specific plan to utilize that data in a visualization, the value of that information is dramatically reduced.

So this article is a call to any developer with even a slight interest in helping to bend the curve on our nation’s badly needed industrial renaissance, or on climate change, or on profitability, or on productivity, to start thinking along these lines.  In many cases the data that would lead to major process improvements already exists, but is not being read or interpreted. In some cases the data is being read, but is in a format that it not easily digested by human managers or analysts.

The skill set required to complete this work is not the skill set required to program a CNC machine or an industrial robot; it’s just the normal skills you may have already developed analyzing traffic on a marketing website, or importing data from Excel, or designing business intelligence systems and ETL processes.  The techniques may be new to you, but the problems are the same classic problems you have likely been dealing with your entire career as a developer; i.e. gathering data from disparate sources, building user interfaces for that data, and figuring out how to find a surface important insights buried in the data.

If you are a designer, the problems I’m describing are fundamentally problems of displaying information in a way that a human can learn something from it.  If you have worked on a project to organize information on screen in a way that make it seem simple, when in reality it is complex, then you likely have the skills to make a difference in this space too.

Finally, the techniques for “jailbreaking” the important data from inside the PLCs need not be programmed in languages unfamiliar to the contemporary developer.  Single Board Computers like the Raspberry Pi are capable of running modern operating systems with modern tools, such as Ruby, Python, Node JS, or whatever other language you’ve been using.  While there is a learning curve, and I discussed some of the specific things I’ve learned along the way in my presentation, these projects are both very approachable and a lot of fun to work on.

What do you need to get started working on a project like this? First, you need a willing industrial partner.  I cannot help you there, but I suspect you can find one if you ask around. Second, you will likely need a partner to help with the hardware side of things if you are not experienced with a soldering iron. If you don’t already know someone in your social network who is a tinkerer or electronics nerd, you will likely be able to find a group of enthusiasts locally, and you may find someone in those groups who would be willing to help.  Wonderful partnerships can develop between a skilled software developer and a skilled hardware person, as they are the perfect complement to each other’s skills.

Good luck, and let’s go make a difference!

David Baltusavich