Knowing there was a lack of visibility when collecting and using machine data for continuous improvement, MachineMetrics founders Bill Bither, CEO; Eric Fogg, COO; and Jacob Lauzier, CTO, decided to use their experience in manufacturing to develop a performance-based manufacturing execution system (MES) and analytics platform to fill the void.

“After I sold my software company, I went back to my roots of manufacturing when we co-founded this company,” Bither says. “We started off by surveying manufacturers and realized the common problem among them was there was all this data sitting on their machines with few products out there to help make sense of it in order to make better decisions.”

Launched in early 2015, MachineMetrics’ software performs deep analytics, displaying data in real-time on dashboards across a facility in an easy-to-interpret manner. Customers are seeing upwards of 20% increased productivity – sometimes immediately – all from making machine data visible and transparent to all (see Carolina Precision sidebar, pg. 45).

Changing perceptions

“Across the board there seems to be a perception of how a shop is performing,” Bither says. “A shop manager might state that machines are running at, say 90%, or whatever the number they think it is, claiming they don’t really need to pull data from controls to understand performance. Typically, once we hook up the machine to our system, we find the machines aren’t running at the expected performance.”

Bither notes that the first step to knowing how a facility is operating is understanding machine performance. Once that aspect is visible, it’s easier to make positive changes and drive up production and quality.

The solution to accomplish this is straightforward since it’s a cloud-based product with typical barriers mainly being a company’s network and information technology (IT) infrastructure, such as getting the network cables to the machines. Once the network is in place, the next step is installing a lightweight gateway device supplied by MachineMetrics. This gateway connection collects data from each machine and sends it securely to the Amazon Cloud where users can interact with a cloud application.

Data timelines displays each event – from fault messages to new part starts to the amount of time a machine was in and out of cycle. Addition report options include overrides, downtime, cycles, and quality.

As Bither explains, “Our approach is for customers to keep their machine network separate from the internal company network. Then we connect as many machines as needed. In this manner, old and new machines alike, even those running different operating systems, can all be connected, securely.”

Now, even with continued growth of MTConnect – manufacturing standards that enable equipment to provide data in structured XML rather than proprietary formats – it’s still not the end-all solution for pulling data from machines. It’s certainly helping companies realize the need for connected machines and data, but technicians are still required to install software and upgrades, a process addressed with MachineMetrics’ hardware and software.

“As an example, if the machine has MTConnect or runs on a Fanuc controller, and it’s already enabled on the machine, it is an easy and straight-forward process for us to connect and start monitoring,” Bither says. “However, with older machines and even some new ones, we may need a separate off-the-shelf hardware device that we connect to machines. Each shop requires a slightly different approach depending on the equipment on the floor.”

Logical learning, scalable solution
Workcenter View can be displayed on an Internet-enabled touch screen device, giving operators the ability to input context of events that occurred on that machine.

MachineMetrics was built to be simple to understand and easy to implement, with its displays looking and functioning like web applications. While training is standard once installed, Bither finds that most users already have a grasp since it’s intuitive, unlike enterprise resource planning (ERP) that requires months of training.

“Featuring the user-friendly interface, when it’s up and running, it’s working and helping. While there is an operator interface that allows for human context to be added to the data, it’s not a requirement,” Bither notes. “An example of when this might occur is if the machine is down, an operator can input context as to why it’s down but that’s about the extent needed for operator input.”

On the scalability side, once the system is up and running, every bit of data is stored in the Amazon Cloud, so MachineMetrics can store data from thousands of machines or just one. Then, every second of data on any machine can be aggregated throughout weeks, months, or years to look at performance across time – the key to finding where improvements are needed. Using those performance data and metrics lets a company know where to focus efforts to addresses its problems. Next is to measure those results against MachineMetrics’ historical reporting and Pareto charts put together from the machine’s data – the transparency that makes all involved aware of all functions.

There are other companies in the market offering similar solutions; however, each has its own angle. Bither notes MachineMetric’s focus is on monitoring the performance of the job being run to know if goals are being met. This ties together everyone – from management to the shop floor – so all can work together to achieve the same goals.

“There aren’t that many companies out there doing this, so with all the talk of the Industrial Internet of Things and connected machines, it’s a good time to be here and offer our product,” Bither concludes.

MachineMetrics

www.machinemetrics.com

About the author: Elizabeth Engler Modic is editor of TMD and can be reached at 216.393.0264 or emodic@gie.net.