"load-torque profile"

"The widespread use of electric motors across diverse load level applications makes it challenging to select an appropriate motor rating for a given load level. Consequently, accurate load torque prediction and classification are essential for reliable motor operation. This work develops a real-time experimental framework using three types of three-phase motors, namely the squirrel-cage induction motor (3 \u2013 SCIM), slip-ring induction motor (3 \u2013 SRIM) and synchronous motor (3 \u2013 SM) operating under multiple load conditions. Load torque prediction and classification are performed using real-time current signal data under various load levels acquired through a digital storage oscilloscope (DSO). A novel Physics-Guided State Space Graphical Neural Network (PG-SSGNet) is proposed to simultaneously predict and classify the load torque of an operating motor. Prediction performance is evaluated with Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and Coefficient of Determination (R2), while classification performance is assessed using accuracy, confusion matrices, precision, recall and F1-score. The superiority of the proposed model is validated through comparative analysis with conventional networks such as Multilayer Perceptron (MLP), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), along with advanced networks like Graph Neural Networks (GNN) and State Space Model (SSM)."