Internet of Things – Stories
Industrial Internet of Things (IIoT) Journey
A large North American plant we work with had to reduce the amount of scrap they were generating to 2%. By using Azure Machine Learning we helped them to create models with historical sensor data, which were then used on real-time sensor data to predict scrap. We worked with them to build a predictive machine learning solution.
Identifying Data Tags & Cleaning Data
During the initial part of this project, System One worked to pinpoint the data tags, which identified a product, that were available and of interest. We connected the Rockwell Gateway to Azure and processed the data in Azure to be consumed by Power BI dashboards. Identifying the tags and cleaning the data is critical for the value of the data in the dashboards and Machine Learning models.
System One developed production dashboards that were deployed on the manufacturing line for operators to use. Such dashboards can be created for any type of organization/enterprise. These dashboards provide the operators with insight on how their portion of the line is performing and aid them in where to make changes, when necessary.
Thank you to Microsoft for coming to Hershey to capture a real-life IoT success story, which our team was part of.
Demystifying IIoT –The Weather Station
For most organizations today, the concept of IIoT is still new. System One developed a method that allows decision makers the opportunity to create their own first IIoT solution. We call this ‘The Weather Station’, which are essentially labs that allow users to get their hands dirty and build plug sensors with breadboards, and to program Raspberry Pis.
The Running of
The Red Dye
When an inspector stands at the end of the production line and looks at the #1 unit of product, he says, “How many seconds ago did it go through the other 200 sensors?”. To answer this question, System One conducted a time study at the facility we were working with. We did this by pouring red dye into the mixer when our team was waiting and watching at each location to see when the red dye passed through.
The War Room
This is where System One goes on site and creates a war room to test the models. At one facility, we were able to set up multiple big screen monitors throughout the room, with each one displaying a dashboard. The dashboards displayed control charts from the sensors which were streaming live data.
The What If
System One developed the What If App to help organizations identify what sensor values to change and what values are optimum. In a simulation, users can pull historical values each time there was a scrap event happening. Then within the What If App, users can adjust the sensor values in an application and rerun the predicted model to see if they are able to reduce the scrap.