How Real-Time Information & Predictive Analytics Unlock Value
Augmented Reality | 19 October 2022 | Team EACPDS
They say a picture is worth a thousand words, so here’s a hypothetical situation to paint the story ‘how real-time information and predictive analytics unlock value.’
To start, imagine a fully functioning assembly line with a robot, pneumatic system, a series of conveyors, and a vision system.
Lets pretend the supply station in the back is bringing in our raw materials. The robot is assembling those materials with precision. The resulting assemblies are than passed on to the quality station, and the vision system inspects each of those assemblies to insure proper alignment of the parts.
This is a pretty generic operation, but it can show how unified real-time information and predictive analytics unlock value.
Now imagine yourself as a maintenance engineer, who wants to check the status of your asset pool.
Using a software, such as ThingWorx Navigate by PTC for example, you launch a role-based maintenance application. All of a sudden you see a complete list of your assets with real-time performance stats and relevant alerts or notifications. You also have a complete list of all your outstanding maintenance work orders.
From here, you have the ability to drill into any of your assets, but you start with the quality station. You immediately see the key characteristics of the station. You see that speed vibration and temperature are all operating within their specified range. You could also see notifications of any warnings, malfunctions, or potential future problems.
Next, you use your device to take a look at the pneumatic system. The pneumatic system also looks fine. Both pressure and flow are operating within the specified range, and there are no outstanding maintenance tickets or work order notifications on your screen.
Now, let’s consider a situation where there was a leak in the pneumatic system. Let’s say a loose fitting was releasing pressure, a fairly common problem in pneumatic systems. Now, rather than looking fine, your device displays flow readings outside of the designated operating range. Furthermore, an alert has automatically been sent to notify you of a system has an error. The overall status indicator on your screen has now switched from green to orange – operational, but not optimal.
Your software solution’s machine learning is now predicting that this air leak, if not repaired, will result in a pneumatic gate failure in approximately 10 day’s time. The good news for you is the system has already issued you a maintenance work order address the problem before asset failure and unplanned downtime.
This scenario is made possible by a system equipped with primary and secondary sensors, and a complete Industrial Internet of Things (IIoT) solution that can turn raw machine data into valuable information.
For example, your pneumatic system has an air flow sensor, as well as a pressure sensor. The conveyor systems are equipped with motor temperature sensors and vibration sensors.
In addition to the sensors, the rest of the assets on your line are controlled by typical PLCs, which are connected to a software such as ThingWorx and Kepware.
You have also used your software to integrate manufacturing floor systems with a real-time IT applications, asset maintenance tools, and ERP systems. This provides you with a real-time alignment of your IT and OT systems.
Now, all of your systems are throwing data out at a staggering 800 data points per second.
Your software’s machine learning then uses that real-time streaming data to establish a baseline of normal operating conditions. This way it can immediately connect and broadcast any anomalies that occur. It uses these anomalies, in conjunction with its prediction capabilities to notify you of future problems, just as in the case of the pneumatic failure.
Now that you have an understanding of what is happening under the hood, let’s take a look at how all this comes together to enable real-time operational intelligence.
Pretend you are a production manager. Using software like ThingWorx Navigate and Kepware you have complete visibility into all of your factory operations. You can see all of your work orders, lines, and all of their critical KPI’s.
On your device you notice an orange status indicator on line one (that was created from the air leak earlier). Once that air leak has been repaired, everything returns back to normal, just as you would expect.
Let’s explore one more hypothetical situation. Consider yourself to be an operator. In this case, you have just been assigned a new order for a thousand units that need to be delivered and expedited for an end of day delivery.
You’re notified of the order and in this smart connected scenario you, as an operator have a single portal from which you can see and execute all of your work. Through a single pane of glass you now have access to your business systems information and your operational data including the KPIs from your line.
On your device you also have up to the minute visibility of the OEE (Overall Equipment Effectiveness). You see real-time data measurements of your manufacturing operation’s availability, quality, and performance.
Let’s see how some of these metrics might change if we go ahead and speed up the line to accelerate the current order, in order to make room for that expedited order.
To do that you switch the line speed from level one to level two. What you see in seconds on your device is that line speed has increased, and your assemblies are still passing the quality check.
Within a couple minutes and a few additional cycles, on your device you see both your performance and OEE trending upwards.
As an operator you now are assured that you are going to meet your end of the day deadline.
Using these hypothetical situations, together we have painted a picture demonstrating how you can connect disparate assets from different vendors, to provide real-time information.
You’ve also seen how you can leverage role-based applications that combine business systems information and operational data to empower your workforce with real-time actionable intelligence.
By integrating machine-learning capabilities you brought a whole new level of predictive intelligence to your factory floor, identified problems, and resolved issues with minimal impact on operational performance.
This is exactly how real-time information and predictive analytics can unlock value for your organization.