Details

Autor(en) / Beteiligte

• Gkouletsos, Dimitris
• Papadopoulos, Athanasios I.
• Seferlis, Panos
• Hassan, Ibrahim

Titel

Systematic modeling under uncertainty of single, double and triple effect absorption refrigeration processes

Ist Teil von

• Energy (Oxford), 2019-09, Vol.183, p.262-278

Ort / Verlag

Oxford: Elsevier Ltd

Erscheinungsjahr

2019

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• A systematic approach is proposed to investigate the influence of different thermodynamic models and multi-parametric variations propagated through single, double and triple effect absorption refrigeration (ABR) processes and to identify the models with the least sensitivity to variability. The approach highlights the parameters which mainly affect the ABR process performance in all indicators simultaneously. A sensitivity index quantifies the variation range of the ABR performance indicators with respect to parameter variability. Cost functions are developed which combine both performance indicators and state process variables. These are used to identify thermodynamic models which simultaneously a) enable close match of simulation results and reference data, b) maintain low process variability and c) give rise to process state and performance profiles of low non-linearity, despite changes in model parameters. We employ up to 12 thermodynamic models for NH3/H2O, comprising different eNRTL models and equations of state such as Redlich-Kwong, Peng-Robinson, Schwartzentruber-Renon and Cubic-Plus-Association. We consider uncertainty in up to 6 component and process parameters and up to 7 ABR performance indicators. The e NRTL-Helgenson model is selected in all ABR process cases. Peng-Robinson for hydrocarbon water systems is selected in single-effect ABR, whereas Redlich-Kwong is selected for the double and triple-effect cases. •Consideration of multiple thermodynamic models in absorption refrigeration.•Systematic identification of appropriate model using sensitivity analysis.•Application to single, double and triple effect systems.•Model selected based on proximity to reference data, parameter variability and non-linearity.