19th AIAI 2023, 14 - 17 June 2023, León, Spain

Characterization of an absorption machine using artificial neural networks

Alfonso Jesús Ferre Montoya, María del Mar Castilla Nieto, José Antonio Carballo López, José Domingo Álvarez Hervás


  To fulfil current energy objectives established by govern ments and public institutions is mandatory to develop zero-emission buildings or refurbish existing ones using, to this aim, systems based on renewable energy sources. Absorption machines are becoming an increas ingly important alternative to conventional vapour-compression chillers due to the possibility of being powered with heat from renewable energy sources, such as solar energy, biomass or others. Thus, this kind of system can be the keystone of the air-conditioning system in residential, public or commercial buildings. Most current approaches to model absorption machines are made on stationary conditions, with less attention paid to the dynamic phenomena inherent to renewable energy sources. Besides that, the lack of internal measures in this kind of plant makes the de velopment of a dynamic model a challenging task. This work presents a dynamic model of a LiBr Absorption Machine developed through Artifi cial Neural Networks. Comparison with real data shows an appropriate behaviour of the obtained model. Simulations were done to evaluate the performance of external temperatures and mass flow rates; moreover, the Coefficient Of Performance (COP) is also included in the model predictions.  

*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.