Physics-Informed LSTM Network for Flexibility Identification in Evaporative Cooling System

In energy-intensive industrial systems, an evaporative cooling process may introduce operational flexibility. Such flexibility refers to a system’s ability deviate from its scheduled energy consumption. Identifying the flexibility, and therefore, designing control that ensures efficient reliable operation presents great challenge due inherently complex dynamics of systems. Recently, machine learning (ML) models have attracted attention for identifying their model nonlinear behavior. This article ML-based methods integrate system into ML (e.g., neural networks) better adherence physical constraints. We define evaluate physics-informed long-short term memory networks (PhyLSTMs) (PhyNN) identification in process. These approximate time-dependent relationship between input response while enforcing network architecture. Our proposed PhyLSTM provides less than 2% estimation error, converges half iterations compared baseline NN, accurately estimates defined metrics. include detailed analysis impact training data size on performance optimization our models.