Sign in Subscribe
Artificial neural networks

Artificial Neural Networks Optimize Waste Heat Recovery in Energy Consumption and Greenhouse Gas Emissions

Strategic Insights

— 1 min read

Investigations into waste heat recovery have reached a crucial milestone with the publication of a new report on artificial neural networks. The researchers from the Autonomous University of Morelos in Mexico, supported by the SecretariA De Ciencia Humanidades, TecnologiA E InnovacioN, have made significant strides in optimizing energy consumption and reducing greenhouse gas emissions. By leveraging the capabilities of artificial neural networks, the team has successfully demonstrated the application of these networks in managing nonlinearities and complex interactions, making them ideal for controlling a double-stage absorption heat transformer.

Key Takeaways:

  • The researchers utilized artificial neural networks to optimize waste heat recovery in a double-stage absorption heat transformer, achieving improved thermodynamic performance and increasing energy recovery from waste heat.
  • The Levenberg-Marquardt algorithm was found to be the most effective method for optimizing the performance of the heat transformer, resulting in a 20% increase in the correlation coefficient and enabling greater energy recovery from waste heat.
  • The study employed R2024a MATLAB programming, real-time data acquisition, and visual engineering environment software to control the double-stage absorption heat transformer and gather data.
  • The researchers compared the Levenberg-Marquardt and scaled conjugated gradient algorithms, testing their performance with different numbers of neurons (5-25) to determine the optimal operating conditions.
  • The study highlighted the importance of the circular economy in reducing waste, increasing energy efficiency, and minimizing environmental impact, emphasizing the need for innovative solutions like artificial neural networks to achieve these goals.

Statistics:

  • The researchers found that applying the Levenberg-Marquardt algorithm resulted in a 20% increase in the correlation coefficient.
  • The study demonstrated the potential for artificial neural networks to improve thermodynamic performance by up to 20%.
  • The researchers used 5-25 neurons in their experiment, testing and comparing the performance of the Levenberg-Marquardt and scaled conjugated gradient algorithms.
  • The study utilized R2024a MATLAB programming, real-time data acquisition, and visual engineering environment software to collect and analyze data.

Sources:

  • NewsRx LLC (2025, July 7). Researchers from Autonomous University of Morelos Describe Research in Artificial Neural Networks (Optimizing a Double Stage Heat Transformer Performance by Levenberg-Marquardt Artificial Neural Network). Journal of Engineering. p. 4131.
  • Romero, J., Diaz-Gonzalez, L., Montiel-Gonzalez, M., Cerezo, J., & Vazquez-Aveledo, S. (2025, 7(2), 29). Optimizing a Double Stage Heat Transformer Performance by Levenberg-Marquardt Artificial Neural Network. Machine Learning and Knowledge Extraction, 2025, 7(2), 29.

Read more