By Seth DeSantis, Director, Brilliant Manufacturing P&E, GE Aviation
(Excerpted from IIC Tech Brief Digital Transformation in Manufacturing: Key Insights & Future Trends)
GE Aviation started its IIoT journey during the second quarter of 2015. That journey started when a group of engineers were looking to code custom PLC logic to capture a handful of machine variables, with the goal of better understanding the process better and then making it more efficient. While they were performing this work, they realized there were many more variables that could be captured in near real time. These variables could enable not only learning out a process, but also understanding the condition of the machine to better predict the likelihood of it breaking. And the likelihood of it producing a non-confirming part.
The initial scope of this work in 2015 was spread out across ~15 manufacturing assets (ex: Furnaces) at two different sites. The approach involved writing custom drivers for these assets, flowing the data from the PLC into Kepware middleware, and then ultimately into a local SQL Server database. Many stored procedures were used to attempt to action the data and there was no vehicle for easily viewing the data.
That first approach led to some initial lessons learned, including:
- Driver Standardization: Creating custom drivers for manufacturing assets was not an efficient approach. This resulted in a mandate on new asset purchases that they come standardized with MTConnect. This provided a semantic layer for common data understanding
- Data Visualization: An end user not being able to easily access and review the data precludes identifying potential collection issues as well as stalls the cultural change to make manufacturing decisions by using machine data. This resulted in an ongoing data visualization journey.
With some Proof of Concept experience under the team’s belt, 2016 brought in a more centralized approach to sensor enabling manufacturing assets. Part of this approach was ensuring that the assets in scope for sensor enablement were appropriately defined to ensure that a positive ROI would be possible for any data insights gained and actioned. This was made possible by only sensor enabling constrain assets, assets that cause a bottleneck and are a limiting factor to an organization’s performance, the obstacle to the organization achieving its goal across our supply chain.
With scope defined, the teams focused on their tech stack and approach. Having already made the decision to standardize on MTConnect at the driver layer, the team also standardized on their storage vehicle. Here, the team would deploy local GE Digital Historian time series databases locally at each site and then also centrally to aggregate the changes in the data. One central Historian instance is deployed for each different type of data classification (Export Controlled, non-Export Controlled, etc.).
The team also explored different visualization tools, ultimately landing on Cimplicity, a GE Digital product already in use at Aviation for Human Machine Interface but not for machine data visualization. These Cimplicity instances were built per site and were customized based on each site’s needs. Data that one would see would range from the operating status of enabled assets to KPI’s on uptime and availability.
During 2016 and 2017, the number of connected assets grew from ~15 across two sites to ~350 across 7 sites. Benefits that were starting to appear in pockets included:
- The reduction of scrap by stopping a process before it completes a non-conforming part.
- Cost avoidance through increased utilization of constraint machines; from visually seeing when a constrain machine becomes available to being able to analyze asset usage patterns, the teams were able to avoid purchasing new assets by getting more cycles out of existing assets.
- Reduction of maintenance hours in the capability to programmatically monitor an asset’s variables, as opposed to a human checking those variables (via, say, a gauge reading), allowed more assets to be maintained by a smaller number of workers.
In 2018 and beyond, the sensor-enabled footprint continued to grow. The number of connected assets grew from ~350 across 7 sites in 2017 to ~1700 across 27 sites by the end of 2019. In addition to the footprint growing, so did the team’s experience.
The IIC Tech Brief “Digital Transformation in Manufacturing: Key Insights & Future Trends,” was designed to help manufacturing leaders keep pace with the rapid emergence of new technology. It highlights advancements driven by the IIC’s Manufacturing Industry Leadership Council (MILC), IIC working groups, and IIC members. Seth DeSantis is a member of the IIC MILC.