b'Data Miningto user IT systems to facilitate transmission of machine- and production-related data. Last year, Anca took part in an umati demonstration involving about 60 compa-nies that shared machine information.In addition to efforts to standardize data to enable sharing information, he sees a good deal of work being done in the area of data gathering for predictive main-tenance. This involves collecting information from ma-chinesvibration, temperature, electrical signals, power draw, etc.and identifying problems as ma-chines fall out of tolerance, as well as informing users of what to do to solve the problems.That will allow more autonomous manufacturing so companies can reduce the number of people on-site, Hale said. This works well in the COVID environment and will also work well once economies get up and run-ning again. Everyone was pushing in that direction be-cause of labor shortages, and I expect that to come back very strongly on the other side of this crisis.Automated Part ChecksAutomated data collection and analysis also can be employed to assess the quality of machined parts. At a plant in South Carolina, an artificial intelligence sys-tem developed by the North American research and development team of Charlotte, North Carolina-based Bosch Rexroth Corp. is gathering and analyzing data to spot bad automotive components made by 38 CNC machines. The machines drill two different holes in each part. After a machine drills a hole in one side of a part, a tower holding the workpieces rotates and a sec-ond hole is drilled in the other side of the component. In the process, chips that are sent flying during cutting operations can end up in the already drilled cavities and produce undesired part vibration.The goal of the R&D team was to eliminate the need for humans to manually check each part for chips in cavities.Its a tedious and time-consuming operation, said Lead Innovation Strategist Matteo Dariol of Bosch Rexroth.The system created by him and his colleagues in-cludes Bosch sensors mounted on a drill to capture vi-bration data, which is collected by an edge device near the machines and eventually travels to an Amazon Web Services cloud platform for analysis.To create an AI neural network capable of evaluat-ing the parts, the Bosch Rexroth team had to extract certain information from the vibration signal produced during drilling. The team also needed labels created Anca60AUGUST 2020MachineControl.indd 60 7/17/20 3:02 PM'