Data-Driven Modeling and Explainable Machine Learning in Energy and Engine Systems

Data-Driven Modeling and Explainable Machine Learning in Energy and Engine Systems Synopsis

This edited volume, Data-Driven Modeling and Explainable Machine Learning in Energy and Engine Systems, includes seven chapters that bring together recent methodological advances and practical applications of intelligent modeling techniques across a diverse range of engineering domains. Chapter 1 investigates short-term wind power prediction using real SCADA data, comparing traditional linear regression with recurrent neural networks to highlight the importance of temporal learning in capturing dynamic power generation behavior.

Chapter 2 proposes a hybrid Prophet-XGBoost framework enriched with SHAP-based explainability, demonstrating how ensemble learning can simultaneously achieve high accuracy and interpretability at the inverter level. Chapter 3 presents a comprehensive acoustic emission-based bearing fault diagnosis framework using time-domain, time-frequency, and frequency-domain analyses. Chapter 4 focuses on aircraft engine NOx emissions.

Chapter 5 introduces an Optuna-optimized XGBoost model for predicting power output in combined cycle power plants. Chapter 6 explores deep learning-based time series forecasting for nuclear electricity generation. Finally, Chapter 7 examines sensorless temperature estimation in permanent magnet synchronous motors.