Lessons of Machine Learning
There’s a lot of talk in the industry about the concept of machine learning and its potential as an enabler of a smarter, safer and more intuitive operation. But machines aren’t the only pupils in this learning process. The design and operations teams have much to gain as well, as a recent Machine Design story notes.
“Machine learning also makes it easy to implement predictive maintenance capabilities by remote asset monitoring, which can avoid costly unplanned downtime,” the article notes. “Reducing unplanned downtime by just a few hours would be enough to pay for implementing a predictive maintenance system, including the sensors, edge gateway/controllers, installation, and operation.”
By improving the operation, machine learning also has the potential to improve the bottom line. At a time when managing costs are such an important part of the plant team’s consideration (and compensation), machine learning also can be profitable.
The Ears Have it
Experienced technicians always have had the ability to detect subtle differences in a machine. The idea that something “just doesn’t sound right” may not be a scientific principle, but that doesn’t make it less valid.
One more lesson for machine learning is its ability to interpret what the plant is “communicating” about its health. 3DSignals is a company that uses deep learning on the cloud to enable sound-based diagnostics for various components in a factory, and it takes that technician’s ear and puts some 21st Century technology behind it.
“When we ask a technician, ‘What is your intuition in understanding if a machine is working well or not,’ the technician most commonly says ‘I can hear it’,” said co-founder and CEO Amnon Shenfeld in a Machine Design article. “With this experience, we came up with an idea to send interesting sound samples to ‘sound experts’ that can give us immediate feedback on what a sound means. Using this feedback, we can very quickly teach our learning algorithms to perform the same sound-based diagnostics.”