Fri, Sep 19, 2025
[Archive]
Jaypee University of Engineering and Technology
Abstract: (46 Views)
In light of growing energy demands and environmental concerns, increasing the efficiency of thermal power plants continues to be a crucial problem. One of the most used thermodynamic cycles in power production, the Rankine cycle, is modelled and optimised in this work using a Python-based framework that makes use of traditional machine learning (ML) algorithms. Basic thermodynamic principles were used to create a huge synthetic dataset that simulated a variety of operating circumstances. Key performance indicators including Mean Squared Error (MSE) and R2 score were used to train and assess a variety of regression models, including ensemble approaches, decision trees, support vector regressors, and linear regression. Complex nonlinear interactions between variables including turbine efficiency, boiler pressure, and condenser pressure were well captured by the Decision Tree and Random Forest models, which demonstrated the highest predicted accuracy and interpretability. The robustness of the model was further confirmed by residual diagnostics and feature significance analysis. The work shows that classical machine learning models provide a quick, comprehensible, and scalable substitute for conventional thermodynamic simulations, opening the door for incorporation into digital twins, predictive maintenance platforms, and real-time control systems. This method is particularly applicable to contemporary, data-rich energy applications since it may be expanded to various thermal systems such as Brayton or organic Rankine cycles.
Keywords: Rankine Cycle, Machine Learning, Performance Prediction, Python Programming, Thermodynamics Optimization
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
