I and my colleauge Onurcan Akıncı completed an Industry 4.0 senior design project on 2017-2018 fall and spring terms in ITU with precious advisory and great support of Dr. Ali Fuat Ergenc. We were both students in ITU Control and Automation Engineering Department, with in passion to do cool stuff. Here is our story of achieving the success in connecting different kinds of industrial equipments to cloud and performing POC's of Industry 4.0 scenarios on it.

A cool course about “industrial data”

I had almost 4 years of interest and experience in IoT solutions and cloud; but never did anything related to that during university. Somehow, I was looking for a way to merge my IoT knowledge with industrial ecosystem. While I was planning my courses for 2017 fall semester, I found the perfect course for an IoT guy like me: KON435E - Industrial Data Communications

Along with the exciting course content, the lecturer was Dr. Ali Fuat Ergenc. As far as I heard from my friends, I already knew that he was a creative academician and had a unique approach to engineering. This was going to be my first course with Dr. Ali Fuat.

When I looked up to course description from the catalogue form, I saw lots of things I didn't know anything about.

Protocols and communication layers, serial communication, physical layers: RS232-RS422-RS485, physical media: copper cables, fiber optics, industrial communication protocols: TCP/IP, Modbus, DNP3(Distributed Network Protocol), industrial networks: CAN-bus, industrial ethernet, ProfiBus, ProfiNet, DeviceNet, CC-LINK, Modbus Plus, AS- i(Actuator Sensor Interface), EtherCAT

I was wondering whether these are old boring industrial stuff from 90s; or something I should know well in order to work on Industry 4.0 scenarios. So, I gave a better look at the course catalogue:

Students who successfully complete this course will have

  1. Knowledge about serial communication physical layers and media,
  2. Knowledge about industrial communication protocols
  3. Knowledge about industrial data networks
  4. Ability to design and install industrial data networks among PLCs
  5. Prepare comprehensive reports individually and in groups,

Well, this sounded pretty good. I felt that I'm at the right place. I registered to the course. Then, the journey has started.

Connecting industrial plants in different countries over internet

During the first week of the semester, I was also thinking about finding a good advisor for my senior design project. Each professor accepts only several students each term and everyone had already started to take seats. I was planning to focus on IoT and Industry 4.0. When I checked senior design project announcements from previous semesters, the situation didn't seem that good. There were just several internet related projects. So, if I wanted to do this, probably Dr.Ali Fuat was going to be the best.

At the first lecture, Dr. Ali Fuat Ergenc mentioned that we could even connect two geographically distinct industrial systems over internet, after some extra work on industrial data communications. This sounded pretty good to me, and a great candidate to be my senior design project topic. I had already done this kind of things; but not in an industrial way. It was going to be a nice application in perspective of Industry 4.0.

Exploring the laboratory

In just first several weeks of semester; I met my colleague Onurcan, completely in luck. He is experienced in industrial devices and PLC networks. While we're studying for the industrial data course, we were also preparing our architecture to connect several industrial devices. Later, we named this a “platform”.

The Power and Motion Control Laboratory in our department, where we were working, is already equipped with an industrial system that can control the lighting in the lab. That equipment was planned to perform some automation using counters and motion sensors in the lab, but probably, it had been eliminated at the user experience phase. We had been using the lights manually for years and I hadn't even know that there was an automation in it. This system seemed perfect for the “connecting legacy industrial system to an Industry 4.0 platform and benefit from it” approach.

We've developed a protocol gateway software for connecting legacy industrial equipment to modern cloud platforms. After lots of trouble, we were able to read or write industrial data on OPC-DA using MQTT over cloud. But there was one more step we needed to pass before benefiting the industrial data on cloud.

First connection established with China

Even we were already excited because of connecting the industrial stuff at the lab to the cloud, Dr. Ali Fuat was still insisting on connecting one more industrial system but from a way different location; his colleague's lab from China. After more and more hard work, including time zone differences, midnight works, (Did I say timezone differences?) installing wechat to be able to communicate with the team at Beihang University. At the end, we managed to read and write from/to two geographically distinct legacy industrial systems. Hey look, we're doing morse alphabet thing with China over cloud!

Simple dashboard

We did a lot

After some time, we needed to handle this seriously. Our primary goal was showing that it's possible to connect different kinds of industrial systems to the cloud without any extra hardware and benefit from great cloud potentials even on the 90's hardware. So, we needed a stream processor to process industrial data in a flexible way. Then, we found some data-driven scenarios to perform on cloud. This was not so hard, the laboratory was already popular. Thus, we could collect lots of data and behaviour from various sensors connected to the system.

The first step was to understand the “everyone is gone” behavior from sensors and when it happens; turning all lights off in the lab. Another step was to understand the “there is a course in the lab” behavior and logging it to the cloud. These kind of scenarios could be done with several timers and flags in PLC's but there are hard limits for them, so it is possible to stuck at some point. Notice that I'm not even mentioning the benefits of data integration of stream processors. You can't integrate your PLC's timers to Zapier, right?

To prepare quick POC's to these scenarios; we've integrated the Siddhi Complex Event Processor to our platform. Well, we could push and pull PLC data directly to a cloud based event processor. It was working great and our lights were controlled by an event processing scenario!

Simple dashboard

Final touches

We were having fun while doing cool stuff but we had to put an end and unite under one roof what we did until that time. Also, it should be shaped to an academic work: this is our story behind an Industry 4.0 senior design project.

Infrastructure of Application-Independent Industry 4.0 Based Connected Smart Platform

We wrapped up everything we had been researching and doing; even including RAMI 4.0 (Reference Architectural Model Industrie 4.0) and OPC to MQTT gateway source codes.

You can read our senior design project, Application-Independent Industry 4.0 Based Connected Smart Platform, and get inspired about building your own Industry 4.0 solution. Notice that we've edited the paper a little bit to share it as a blog post. Please comment your ideas and thoughts; and let's discuss if you have an other approach to our solution.

Thank you!