What vibration tells us about Predictive Maintenance & why no two vibrations are the same.
The ability to extract value from data has become a game-changer for companies seeking competitive advantages.
As we know, when using any product, there are typically key indicators of its performance. Some of these are easily identifiable while others are more obscure. I’d like to talk about an example of a detectable issue related to vibration. Many mechanical issues can be highlighted or detected by the sounds or vibrations coming from the product. Consider driving a car, and you feel the steering wheel begin to shake. This shaking wheel is a classic indication of an underlying issue – tire balance, wheel alignment, etc. Complex machinery provides the same types of indicators through its component’s vibrations. However, these types of vibrations are usually indistinguishable to human sight or touch. But sensors available to us today can perform vibration analysis that detect anomalies and measure change both in vibration sensitivity and temperature – among other metrics. When this data is incorporated with machine learning, you can pinpoint areas of wear and exposure, allowing proactive maintenance that maximizes the health of your investment.
Consider an aircraft maintenance provider who is responsible for thousands of mechanical inspections and repairs across a fleet of hundreds of jets. Maintenance providers for aircraft are typically compensated for excellent mechanical work, naturally, and also having aircraft available to fly. Inspection coordination is absolutely critical to minimize aircraft downtime and maximize performance.
Sensors spread throughout the aircraft’s componentry provide invaluable vibration data points, which tell a unique story about what is going on inside that aircraft. For example, the health of the engine can be monitored, piece by piece. When this information is consumed by machine learning algorithms, we begin to unfold hidden patterns that allow predictions, with given confidence intervals, to be made. Predictions based on vibration patterns tell us whether a component requires adjustment or something much more serious. Getting ahead of mechanical problems and incorporating into planned maintenance schedules provides a substantial performance advantage. In this aircraft use case, the types of proactive actions minimize costly downtime, increase mechanical reliability, and most importantly, provide further safety precautionary measures for those operating the aircraft.
This is an example of an aircraft sensor telemetry control center measuring specific aircraft health and overall fleet well-being, all based on machine learning predictions from the fleet’s telemetry. But it’s just one example... There are thousands of other use cases for predictive maintenance for manufacturers. When machine learning tells us there is an upcoming problem, the command center gives an action to remediate. We use that data to predict when a machine is at risk for downtime and save potentially millions.
Getting ahead of maintenance keeps your operation & equipment running smoothly.