Machine Learning Algorithm

For more than a decade RSL have been implementing advanced methodologies & unique algorithms in order to perfect the decision processes of its diagnostic and control systems.

In order to achieve effective, reliable, sensitive and quick diagnostic performance throughout all the monitored plant operational modes, both steady state and transitional, RSL has extensively perfected its Machine Learning, Artificial Intelligence and decision making diagnostic support algorithms..

Pre-processing algorithm is extensively used in order to eliminate spurious sampled data to reach any of the diagnostic processes. This algorithm  make use of different mathematical procedures for manipulation of data such as fast RPM synchronization, data dimensionality reduction with non-linear PCA, data demodulation and source separation.

Knowledge-based algorithms are implemented in either deterministic or fuzzy logic are used for efficient detection of operational regimes and advanced anomaly detection and classification.

Self-Learning algorithms which constitute the major algorithms used for prediction of analytical and diagnostic decision processes, using automatically generated powerful Data Driven Models which dynamically simulate the machines’  behavior, which is then compared in real time to the actual machine behavior enabling anomalies detection at early stages.

This powerful capability also allows the incorporation of the output of the Machine Learning to the control elements of the machines thereby preventing uninformed and undesirable entry into specific operational modes which might be hazardous to the machine.