Silafrica has been on a journey to adopt industrial IoT technology and use real-time data to optimize factory operations. Akshay Shah, the Group Executive Director of Silafrica, recently had an interview with the CEO of ThingTrax, a company that supplies this technology, about Silafrica’s journey, challenges and what has been learned. If you are considering investing in similar technology, you may benefit from some of these learnings discussed in this interview. Connect with Akshay on LinkedIn to share ideas.
Please enjoy Part 2 of this three-part interview series.
PART 2: The Journey of the Silafrica and ThingTrax Partnership
Paul – You’ve talked a bit about your background and the general things that you were interested in learning more about. In your mind though, it seems like you have a vision for Silafrica. Could you tell me a little more about where you’re at today, and where you want to get the business to eventually?
Akshay – Anybody who is in the plastics manufacturing space knows that the majority of the costs sit with raw materials, energy usage, equipment and molds, and direct manpower. So, whether opex or capex, it’s really all on the shop floor. Having granular visibility – in a real-time way – of data that lets us become predictive, that helps us to become more efficient and gives more control overall, is the end goal.
Let’s look at your top four costs, you know, raw materials, energy, manpower and maintenance; and then your main investment, which is shop floor machinery. If there is a way to get real-time granular analytics that help to predict where we take our costs and our assets, in order to squeeze out the maximum efficiency, we can then become much more competitive without having to take any shortcuts or waste any more money and resources. Rather, we can improve efficiency, quality and consume less resources – which is good for the planet, for our profit and cashflow.
Before ThingTrax, there is inherent inefficiency; simply because we are looking at yesterday’s information, last week’s information, or last month’s information to make decisions. By the time something changed – you know, ambient conditions change, maybe there was a power fluctuation, maybe there is a change in the product mix – something has already started to change within the minute or within the hour. The earlier we can catch it, the less rejections we have, and the less time and energy we’re wasting. So, this whole idea of being able to identify changes in real-time, and use the data to become predictive and competitive, is key.
Beyond the idea of becoming competitive, it was also about control. Silafrica currently runs four separate manufacturing facilities in three countries. Having that granular visibility across the multiple operations, down to every machine, mold, chiller and compressor, helps us to maintain a high level of control. This is really important in ensuring consistency and accountability across sites.
The third thing was also about talent. There is a different type of talent required on the shop floor to actually go onto the machine and adjust machine settings, do preventive maintenance, mold loading and what not. But the talent that is needed to look at all the data and figure out what is the right combination, when should we be making changeovers and optimize the planning – that’s a totally different skill set and not something that is easily available. Having software be able to do this for you, without the need for additional investment or risks from human error, makes a lot of sense to me.
The final area is around leakage. In plastics, the raw materials, and even energy, is a commodity that can easily go missing if you do not have the right controls in place. Then, there is also theft of productivity which is your cycle time. There is a desire for workers to pad the cycle time a little bit on the shop floor level so that they can show performance within targets. But I would rather take a different approach – let all of these leakages of time, energy and materials get exposed so that we can work together to fix the problems. If they don’t get exposed, we don’t know where the leakage is happening. So, you don’t know where to put your finger into the hole to block that leak. I would prefer that we reveal that leakage, let the data expose everything, be completely transparent with each other, and then become proactive in fixing all these leakages. Getting real-time data from IoT (Internet of Things) devices helps us convert these leaks into meaningful information and insights.
Paul – So one question I had that builds on your comments is that this feels like a journey to get to your end vision. A significant amount of investment, not only in new technology, but also adapting to new ways of working. There is a change of competencies that need to be developed in the organization. I sense that you need to be really compelled to take that first step and start on that journey. What is the main pain, the main motivator or the main driver that resulted in you starting down your IoT journey?
Akshay – Initially, it was about getting our costs down and having transparency around what’s happening in the factory. I would say the main motivator was wanting to know what’s happening in the factory on a real-time basis, and not wanting to depend on people to collect data manually and put that into excel sheets. Even if you have SAP or Syspro, or another ERP system, ultimately there is still reliance on receiving data from a human being unless you’ve got an IoT interface. I was just not comfortable depending on data that was coming from people – I needed it to come from the machine itself.
Paul – And if you were to solve one challenge initially for you to get business value from the first stages of effort, what area do you think that is? What area of opportunity do you think you’d want to focus on from a business value standpoint?
Akshay – It would definitely be around the Overall Equipment Efficiency (OEE), because at the beginning, we’re still doing this manually. We want machines to capture this so that when we look into the data, it’s like looking into a mirror – good, bad or ugly – you get to see the reality. Now, the team in the factory might be reporting an 89% OEE, or they might be reporting a 72% OEE. How do I react to that? By the end of the day, that data could easily change if there was a formula calculation error, data collection error, entry error or whatever. And maybe if they don’t have the data, they don’t put in the data. Just knowing the reality and being able to say okay, well our OEE is lower because we’ve got a low rejection rate, or our OEE is lower because we’ve got much more downtime than anticipate, enables us to then have targeted interventions. Then our OEE can improve in reality, rather than in perception. I would say that’s a quick win, because you are now operating on real data and can make real interventions.
The other quick win would be around energy management. That’s why our next phase with ThingTrax was not just to collect cycle time data, but also the energy consumption of each cycle. Energy in Kenya is particularly expensive; we pay somewhere between 16 to 18 U.S. cents per kilowatt hour. So again, correlating for example your real-time energy consumption to real-time production so that you can get your kilowatt hour per kilogram produced (kWh/kg), is another area were seeing information very transparently, and then being able to have targeted interventions based on that data. This means that we could go to a specific machine or a specific production area, and immediately take intervention when we see spikes happening when it comes to kilowatt hour per kg.
Let’s say a machine is running – and this happens quite easily with grinders or shredders, the machines that chop up the plastic into smaller bits. We have grinders that don’t have a sensor to check whether there is any material coming into the feedthrough. An operator might take a break and forget to switch off the grinder or might just slow down their work and not feed material into the grinder. So, the grinder keeps grinding, the machine keeps consuming energy but there is no material going through it. The same thing happens with your injection molding machine – if you keep it running for too long, and there is no material going through, then you’ve got higher kilowatt hours per kg produced. That spike immediately means that somebody is not paying attention to energy management. Instead of trying to fix that problem after looking at your energy bill from your previous month, and manually comparing this with your production for the previous month, seeing the problem has been there for a month before going to fix it – you’ve likely already lost maybe 10, 20, 30% of your energy bill, and it’s too late to go and do anything. Then you’ve got to do a root cause analysis and figure out why and how, and it’s all too late. So, I think that the ability to have targeted intervention in real time when it comes to energy and efficiency are immediate quick wins.
“The ability to have targeted intervention in real time when it comes to energy and efficiency are immediate quick wins.”
Paul – Do you have an idea of the amount of energy that is wasted, and the opportunity space for improvement? Like, is it 30% and you’re trying to get it down to 20%?
Akshay – No idea, remember, the baseline is manual information. So, if I had to guess, I would be guessing off a fictitious baseline.
Paul – Got it. It’s back to the insight in the first place about knowing where you are first so you can figure out where you can get to.
Akshay – Right now we’ve got 29 IoT devices in the factory. That is a good enough number for us to get a real baseline for energy consumption, at least for those machines. And then we will run it for a period of time until we’ve got a really good baseline, and then after that we’ll be able to figure out, “Does this look abnormal? Does this not look abnormal?” – and then start making those interventions. In 6 months from now, I’ll be able to answer what I could be making in potential energy savings, or material savings or efficiency savings.
Be sure to read the conclusion of the interview in Part 3 of this three-part series.
If you’d like to learn more about implementing ThingTrax in your own factory, simply input your biggest manufacturing challenges into this form and one of our Factory Advisors will come back to you with a list of solutions tailored specifically to your needs.
Or, if you want to see ThingTrax in action first, book a free trial with us here, where we’ll connect your factory and give you access to real-time analytics and insights from your own machines in less than a day.
The time is now to invest in the digital tools that will transform your business and the technology to drive continuous improvements for your factories well into the future.