My work to bridge the IT/OT gap can be divided into two steps. One is my work establishing a networking connection between the bandsaw and a data collection platform. The second is developing a smart manufacturing data model for the saw. Unfortunately, time was not on my side and the company became insolvent before I could put these two facets of the project together. However, on my last day in the shop before the project was put on hold indefinitely, I finally achieved a successful connection between the bandsaw and the data collection platform.

Once the PLC was installed onto the bandsaw, I moved on to establishing an IIOT architecture onto the machine. The first step was to connect the machine to the SMIP, a cloud-based smart manufacturing platform developed by CESMII. This step proved to be possibly the most challenging and frustrating part of the project. What I expected to take a couple weeks took me several months to resolve. While the SMIP primarily uses OPC-UA and MTConnect as its data models, the CLICK PLC utilizes Modbus and MQTT as industrial communication standards. Bridging this communication gap was challenging due to differing IT and OT standards. The SMIP focused on object-based protocols for data categorization, while the PLC relied on serial communication for live data transfer between machines. To address this, I collaborated with Real Time Automation, an Industrial networking expert, and utilized their Modbus Slave-OPC-UA Server gateway. This gateway converted serial Modbus RS-485 data from the PLC into native OPC-UA object data, allowing the SMIP to historize and analyze it. The data, organized as a set of Modbus registers, was designed to be structured into a functional data model on the SMIP. With the help of RTA, CESMII, and ThinkIQ networking experts, I was able to establish a successful connection to the SMIP and verify that data was being captured from the machine. One important step was the use of a static IP on the host computer to maintain a connection to the cloud database. Another important step was making sure that the SMIP was historizing data variables that were setup. Without this, data was not being shown on the platform. These problems, along with several other important steps that were only figured out through trial and error needed to be setup correctly to establish and maintain a valid connection.

It was critical to structure the data from the PLC into a functional data model to contextualize the live-data and derive operational improvements. A fishbone and UML diagram were used to develop a novel data model for a standard industrial hydraulic band saw. In addition, a process control table was created to map the collected attributes to the standard operational parameters for data-driven optimizations. Parameters that were considered include OEE, power consumption, and machine utilization. I was able to obtain raw bandsaw data on the SMIP indicating a successful connection between the bandsaw and the SMIP, but before I was able to implement the data model, the company became insolvent and had to close shop.