Component fouling estimation
Incrementally learn to predict pressure curves in the system, use predictive error as indication for fouling.
Relevant topics: Incremental learning, multilayer perceptron, continous error estimation.
Use case archetype: Condition monitoring
Virtual sensor for emissions
Predicting NOx levels based on temperatures, oxygen mix and flow.
Relevant topics: Multilayer perceptron.
Use case archetype: Virtual Sensors
Optimizing burner emissions
Use virtual sensor for emissions and predictive models of burner temperatures to optimize the process.
Relevant topics: Incremental learning, multilayer perceptron.
Use case archetype: Predictive control
Condition monitoring of a valve
Incrementally learn to predict normal behavior, report divergence back as a condition score.
Relevant topics: Incremental learning, concept drift detection.
Use case archetype: Condition monitoring
Vibration analysis of ball bearing
Use Spectral to extract vibration characteristics, monitor them using change detector.
Relevant topics: Ekkono Spectral, change detection.
Use case archetype: Condition monitoring
Remaining useful life of mantle
Incrementally model and predict wear (mms remaining) to estimate when it hits zero.
Relevant topics: Incremental learning, linear regression.
Use case archetype: Predictive maintenance
Remaining range of electrical bus
Incrementally model and predict remaining charge, use model to understand individual consumption.
Relevant topics: Incremental learning, multilayer perceptron.
Use case archetype: Device modeling
Predictive climate control
Incrementally model temperature dynamics in a room, use model for predictive control.
Relevant topics: Incremental learning, multilayer perceptron.
Use case archetype: Predictive control