Considering all the hype there's been in recent years about the potential of the Internet of Things (IoT) in industry, it's refreshing to hear an example where the technology is having a real-world impact. Speaking during this week's PowerPlex virtual event, Bob Bierwagen, VP of Strategy for MPI Corporation, described an IoT pilot that is not only shaving costs from a core manufacturing process but has also created extra revenue.
MPI is a group of companies based in Indiana that provides metal fabrication and distribution, much of it for the automotive industry. Bierwagen described an IoT pilot at HTI, the group's heat treatment business.
HTI is a leading specialist in a process known as Austempering. This is a hardening process for iron-based alloys that first passes the metal parts through a high-temperature furnace to harden them, and then quenches them in a molten salt bath kept at a precisely controlled temperature. The process leaves the part highly resistant to shock and metal fatigue — crucial for the many thrust washers MPI provides to go into vehicle transmissions, among other parts.
MPI chose to do its IoT pilot at the HTI factory because the team there was able to identify a clear business case. There are four furnaces at HTI that perform the Austempering process. Parts are carried through these furnaces on nickel titanium belts that are able to withstand operating temperatures of 1400-1800 degrees Fahrenheit (750-1000ºC). But if those belts get out of alignment, it can lead to adverse wear to the belts or damage to the furnace, which means production has to be halted. Bierwagen explains:
When you've got a furnace at 1800 degrees, it can take 24 to 36 hours to cool. So unplanned downtime was not a good thing for us. We identified that we could save potentially three or four occurrences of this a year if we had better control.
IoT sensors monitor critical conditions
The IoT project installed infra-red sensors at the furnace to track the edges of the belt, which fed data into a Rockwell controller to calculate the center point of the belt. That data then feeds into a Kepware server which acts as a gateway into Plex IoT. A dashboard in the Plex software provides visual alerts for abnormal or out-of-bounds metrics, and users can sign up to get email or cellphone alerts for either condition. That's invaluable, says Bierwagen:
Being able to push that out to email or to cell phones that our maintenance or operations people can see in real time is really an exciting place to be for us.
A second element of the pilot was to add an infra-red sensor at the autoloader which drops parts onto the belt. Its function was to check the height of the part because overstacking can lead to quality issues. As well as alerts, having the data recorded makes it possible to go back and investigate what happened, says Bierwagen.
What's important with any of that is that we have the data to go back and look at. So if something comes up that we need to investigate, we can then go take a look at that data and see how it's running.
Based on the results of the pilot, MPI expects a significant impact in its next financial year which starts from October, he says.
Based on what we've seen of ROI so far, we expect a 30% reduction in unplanned downtime cost, about a 10% reduction of plant maintenance, about a 10% reduction in job transition time.
Reducing the transition time between jobs has made it possible to put more parts through the furnace in a given day. That has already translated into higher monthly revenue, says Bierwagen.
Gap time is one of those real key metrics for us. We want the operators putting parts on the belt in time, so that when the salt is at its correct temperature the parts are already there.
There are other potential revenue gains, he adds, through improving quality:
We can reduce part loss for our customers, which means they have to provide fewer parts to us as a service provider.
We know that we're going to be able to improve product quality through much tighter control of the production process. That's a major competitive advantage when you're talking about second-tier parts like transmission washers. Obviously, when the washer fails and the transmission fails, a car fails. That's not a good thing. So being able to improve quality output is a very important aspect of our customer-facing approach.
Incremental changes with minimal disruption
One often overlooked aspect of this kind of IoT project is that it can be implemented with very little disruption. Aided by advice from Rockwell Automation, the pilot was set up to work alongside existing legacy sensors and required very little new capital expenditure. Bierwagen says:
One of the nice things about IoT is it's not a rip-and-replace like you have with an ERP project where everything changes. You can do these in smaller pieces, and as part of this we were able to validate our IoT components to ensure compatibility and capability as we move forward.
Now that the pilot has been successful, MPI will be doing some further spending to upgrade some of the legacy controllers on the HTI furnaces, he adds.
Right now, all of the controllers to set the temperature settings, the belt speeds, the quench tank settings, all of the atmosphere controls, everything is manual. These are probably 30-year-old controllers. And with the success of the pilot we've decided that we will do some additional capex.
The ultimate aim is to pull together all the sensor and machine data with the ERP and MES data on jobs and production, so that everything can be managed automatically.
Where we want to be is the ability to set the furnaces, to set all of the equipment that we have within HTI, using a recipe system and to totally close the loop. So we've got a closed loop manufacturing process.
In the meantime, there's more that the IoT system can do, such as reacting more quickly to defects.
The furnace in our case is a very complex beast and much more so than I would have guessed as we got started in this ...
There's about 150 data points in total that we've identified that we can keep track of. And we may not do all those right away over the next six months, but we have, with the operations team, identified the key components that we want to track.
For example, by monitoring the temperatures throughout the process, the operator can be notified immediately of anomalies that might mean a batch of parts needs to be set aside. That significantly reduces the size of the batch that needs to be pulled.
Defining real-world objectives
The final goal is to start using artificial intelligence to make recommendations for more efficient maintenance and operations.
We want to get to not just predictive maintenance but to prescriptive maintenance, to the point that the system is telling us 'Hey, you're gonna have to take a look at this belt position or look at your salt tank, might have your heating elements need some analysis.'
Having the data is great for being able to have the system drive insights to you, as opposed to just waiting for something to happen or having somebody analyze that data.
The starting point for the success of the pilot was sitting down with the operations team at HTI to define some real-world objectives at the outset, says Bierwagen.
The key point here is that this is not something I created or that corporate created ...
Working with the team there, we actually identified the major opportunities that we had to generate ROI, both from a maintenance standpoint and from an operational standpoint. From that, we then generated a pilot project objective.
This was within a broader strategic goal of using technology to become more proactive as a business, says Bierwagen — not just recording data but actually using it to make business decisions, and putting tools into employees' hands that enable them to be more engaged in delivering better outcomes. It's essential to see IoT as part of this kind of bigger strategic picture, he believes.
It's more than just IoT. It's really about transforming business. I think as you look at any IoT project that you do, you need to take that approach.
Any technology innovation has to be grounded in the real world if it's going to produce meaningful business outcomes. MPI's story shows the importance of starting from actual needs and pain points when designing a pilot project. There's nothing fancy here — just adding some sensors and controllers that collect crucial extra data to improve the efficiency and accuracy of a core manufacturing process.