Remote Factory Monitoring Installation Using Arch

Watch or listen to this short video explaining the process of a fully remote installation of our factory-monitoring platform.

Now, More Than Ever

“Remote” is on everyone’s mind these days. From remote work to remote management, monitoring, and servicing, decision makers are looking for best practices to adopt at every level of their organizations. One of the most common questions we get from customers is, “Can you really do a remote installation of your factory monitoring platform?” The answer is yes, we can, and we do. Some manufacturers either aren’t able or choose not to open up their facilities to an outside team. We’ve had great success helping Tier 1 contract manufacturers (CM) establish Global KPIs, all without sending any member of our team onsite. It takes experience, but if COVID-19 taught us anything, it’s that any business able to continue operations with little downtime had a clear advantage. Here, we show you what a remote installation looks like and how it is not only possible- in some cases it’s critical.

STEP 1: Identify Needs and Goals

The first thing we do, right out of the box, is work with the company to establish what the high-level project goals are: What are we trying to do? Why are we doing this? What’s the value?

STEP 2: Identify Project Constituents and Acceptable Timelines

Then, we identify the owners from each constituent group. We are going to be pulling people in from IT, operations, often engineering, sometimes security, potentially from a group that oversees multiple sites, down to the individual teams in the sites we are going to work with.

Then, we define project specifications and timelines with that extended team. That way, everyone is on board and understands and agrees on what the objectives are before we ever begin.

STEP 3: Identify Desired Data

When we have that established, we can go in and identify the desired data from the equipment we want, or the data that already exists and what the extraction methods are going to be. We are not trying to pull all the data in a site. We are trying to pull the data that is going to give us that end value the customer wants.

Those methods may be direct digital connections to the machines. Alternatively, we may need hardware to tap to some of the equipment to pull data out of it. From that, we ship the hardware (if any is needed) for the data extraction. Most customers have smart machines where we can pull the data digitally as well as legacy machines where we may need to add customized hardware. Whatever is needed for us to get all the key data- we do it.

STEP 4: Install Digital Infrastructure

In parallel to Step 3, we set up the digital infrastructure. This normally consists of the customer setting up a virtual machine on a server inside their firewall where it has a connection to pull data directly from the smart machines. We install our ArchFX Factory Exchange Broker software on that machine remotely, which means we need remote access to the server where we can log in. It is possible to package things up for the customer to do that without us. However, particularly when we are trying to do everything remote, it becomes very challenging if we don’t have a way through a credentialed VPN to log in to the machines, set things up, and most importantly, ensure everything is working properly. It’s the debugging steps that are the most challenging if we don’t have access.

This was the situation with a CM in China where we had planned to do an installation in person, and we couldn’t because of COVID-19. So, we proceeded with doing it remote-only. In that case, we were helping an OEM decrease their build times by giving them remote visibility into the CM’s production line. We’re two levels removed from a direct digital connection, so doing a remote installation blindly presents some unique challenges. It requires an extra level of documentation and coordination to ensure it happens correctly, but it is definitely doable.

STEP 5: Harvest the Data

If we’re pulling digital data from a smart machine, we either have existing connectors we have written to interface to the machine from the ArchFX Broker, or we will write a new connector. We may be pulling performance and quality data for Global KPIs, status and cycle times for local dashboards, or job complete signals to feed into an MES. If we’re working with standardized data, like IPC-CFX for electronics manufacturing or OPC-UA for mechanicals, it’s easy to do. In some cases we have to create a custom connector for a given machine, which is more work, but once again, it’s all part of the package.

For some of the legacy equipment, we connect our custom hardware to extract the required data. We walk the customer through the installation process and work with their shop floor engineer. We have written installation instructions for how to set up and configure our equipment. We also ship across Power Over Ethernet Access Points that wirelessly connect to our hardware and provide a data path to the ArchFX Broker. Our customers don’t have to pull wires to each machine to get the data out. The importance of wireless connections to the machines can’t be overstated.

So, into the ArchFX Broker we now have digital data that’s coming directly from the smart machines and hardware-extracted data from the legacy machines.

STEP 6: Standardize the Data per Client Need

From here, it depends on where the data is going – what are we going to do with the data to get the desired end result? If we are pulling the data out and building a Global KPI system, the company normally opens up a specific secure port for us to push that data into the ArchFX Cloud, something we can do on-premises as well if needed. We’re then able to build the dashboards for the specific metrics of interest.

If we’re doing a predictive maintenance or predictive quality application, we’re gathering a certain amount of data or the company is sharing historical data that they may have for the equipment so we can build a ML or AI model. We analyze the data, using techniques like time series spectroscopy, and look for patterns. We can identify the signal (out of the many different data streams we may have) that shows us an imminent failure, and be able to display those signals on a control chart that’s easy for the operator to understand. Knowing that there are 30 days to do maintenance for that piece of equipment, to get it out before it causes a failure for the factory is but one example of the information we are helping to establish. It is helpful if there are examples of machine failure, but we can build solutions with a positive ROI without those examples.

On the other hand, we may be connecting into a local manufacturing execution system (MES). For example, we’ve connected into various mechanicals operations where they’re doing metal cutting or bending, or creating automotive parts in plastic injection molding. What we’re doing is getting the count when each job is finished and sending that into the system so they know when the job has started, when the job is done, etc. From a scheduling point of view, the customer has all the needed information to be able to know that a machine is free, when they can schedule another job into it, understand what materials are being used, view the scrap, etc.

In short, a remote-only installation requires good communication and coordination, but much of that would be required when going onsite to do the installation as well. So, there’s a little more communication that needs to happen up front, but there’s no travel. We’ve done the additional documentation, we’ve put the pieces together so that we are able to send any needed hardware to do the remote installations for the legacy equipment. We can do the connections and communication, and we are able to put together the solution for the customer that provides the value we defined up front.