Condition monitoring provides insights about system health beyond raw sensor values. It helps you prevent severe failures and unplanned production stops by exposing the true health of your machine. When using incremental learning, we don’t need historical data. Instead, we leverage the the power of real-time data.
Edge Opportunities
With Ekkono’s edge machine learning SDK it is possible to incrementally train a condition monitor. We do this based on streaming data in your device. Incremental learning eliminates the need for historical data.
The actual learning is performed inside your individual devices, using real-time data collected out on the field. The learning starts with a short initial training period. This initial data is used to determine the normal data state when the machinery is healthy. In the next step, we use this normal data state as a baseline to detect deviations or irregularities in the streaming data. The detected deviations are compounded to an overall health score. In the end, you can use this health score to determine the need for manual inspection of data or device.
Usually, extensive data collection in the field is expensive. Let our incremental learning capabilities do the job for you. We have no need for the historical data required by traditional batch training.
Condition Monitoring
Problem
The classic batch training approach to condition monitoring requires a lot of historical data. We often find that there is a lack of historical data that includes enough failures to support traditional machine learning models.
You could use simple threshold based models to determine health. But there is a risk of missing important indications of degradation that do not exceed the specified thresholds. On the other hand, those deviations are readily detectable by our machine learning algorithms.
Solution
Learn the healthy state of the device and monitor how the real-time data deviates from this. The input sensor or array of sensors used should be related to the health of the device. Using incremental learning, the result is a unique health model for each individual device.
Application
Any deviation from the normal state indicates that something has changed in the device. You can use our health monitor to determine when manual inspection of the device is needed.
Read more about our Use Case Archetypes.