Vibration Machine Learning from the Industrial Internet Consortium
Grant Gerke | April 25, 2017
Some industry analysts aren’t happy with overused buzzwords like “machine learning” or even “deep machine learning” taking the place of “IIoT” in the hype category. I agree these new buzzwords are ubiquitous in many media corners and deep machine learning is mostly found in R&D.
However, a white paper or deep dive is a great way to see what is possible for predictive analytics in the field or factory. A new white paper from the Industrial Internet Consortium, titled, “Making Factories Smarter Through Machine Learning,” offers a great read on how machine learning can allow for better edge analytics, reduce data streams and promote better data fidelity.
The white paper examines the ability of CNC machines to reduce data streams via machine learning with the use of the Plethora IIoT platform and system-on-chip engineering (SoC). The SoC technology allows for customized software to create application-specific requirements, such as data filtering being sent from machines.
A passage from the White Paper below:
The other capability provided by the software is the ability to read complex sensors and perform pre-processing in terms of data reduction: For example, vibration is sampled at least two times the vibration frequency. In this case, a fast Fourier transform is performed and only the frequency of interest is stored. This is an area where there is high opportunity for more efficient processing – effectively using machine learning for pre-processing and feature selection.
Therefore, it (SoC) can sample each variable with smart criterions: For example, temperature may not be measured with the same frequency of vibration
The white paper provides a real roadmap solution on how to move from preventive, SoC machine learning and simple industrial networking solutions to make this happen. The link to the white paper can be found here.