In a recent article from the MachineMetrics Tech Blog, the data science teamlikened high-frequency data to being able to hear the complete symphony of notes from the machines. The metaphor wasn’t far off…we can quite literally hear the sounds our machines are making through the raw PLC (programmable logic control) data we’re collecting, without any microphones.

Due to the interference of other noises in a machine shop however, audio cues that machines provide are often disrupted and rendered less useful.

To get around this, our data science team proposed an approach where we don’t useanyexternal sensors. Our current advancements in research have brought us to a point where we can create an analog to sound with just the PLC data, allowing us to extract the pure, unadulterated sounds of machining without factory noise.

In essence, we discovered a technique that can extract the latent harmonics on the load and spindle speed data we’re getting from the PLC, with the result that we “hear” the machine through just these two signals. In "The Inaugural MachineMetrics 8-Track," we’ll go over how we did this and some of the implications for industry.