A method and system for detecting operational irregularities in a thermal energy exchange system, and a method of training a machine learning model

The patent replaces manual analysis methods with automated machine learning models that process measurement data from thermal energy exchange systems. The system uses trained machine learning models to automatically detect operational irregularities, substituting human analysts with computational algorithms that can process large volumes of data quickly and accurately without time constraints.

The patent implements a feedback mechanism where machine learning models continuously monitor measurement data from thermal energy exchange systems, detect operational irregularities, and enable corrective actions. This closed-loop feedback system identifies sub-optimal performance early, allowing operators to adjust system parameters before temperatures need to be increased, thereby maintaining energy efficiency while still meeting heat demand.

The patent applies preliminary action by training machine learning models offline using historical measurement data before deploying them for real-time detection. The models are pre-trained on comprehensive datasets that include various operational conditions and fault scenarios, enabling them to process new data efficiently without requiring complex real-time analysis infrastructure. This offline training phase separates the complexity of data processing from the operational detection phase.