An applicability of ANFIS approach for depicting energetic performance of VCRS using mixture of R134a and LPG as refrigerant. (January 2018)

• Record Type: Journal Article
• Authors: Gill, Jatinder; Singh, Jagdev
• Is Part Of: International journal of refrigeration. Volume 85(2018)
• Page Start: 353
• Page End: 375
• Publication Date: 2018-01
• Subjects: R134a/LPG -- Energy analysis -- Mathematical model -- Adaptive Neuro-Fuzzy Inference System (ANFIS); R134a/GPL -- Analyse énergétique -- Modèle mathématique -- Système d'inférence neuro-flou adaptatif -- Système ANFIS; Refrigeration and refrigerating machinery -- Periodicals; 621.56
• DOI: 10.1016/j.ijrefrig.2017.10.012
• Languages: English
• ISSNs: 0140-7007

Abstract

Highlights: Energy analysis was carried out on VCRS working with R134a/LPG and R134a. Energetic performance parameters obtained with R134a/LPG were found better than R134a. Mathematical and ANFIS models for prediction of energetic performance parameters were developed. Statistical performance analysis of mathematical and ANFIS models was measured. ANFIS model predictions showed better agreement with the experimental results. Abstract: In this work, an energy analysis of vapor compression refrigeration system (VCRS) using R134a and LPG refrigerant mixture as an alternative to R134a is carried out. Performance tests were performed with different evaporator and condenser temperatures under controlled ambient conditions. The results showed that R134a and LPG refrigerant mixture has higher coefficient of performance than R134a by about 15.1–17.82%. Mathematical and ANFIS models validated by experimental data were developed to predict the energetic performance of VCRS. It was found that mathematical and ANFIS models predictions agreed well with experimental results and brought out the absolute fraction of variance in range of (0.9253–0.9567 and 0.9953–0.9961), root mean square error in range of (0.0167–7.688 and 0.0018–1.025) and mean absolute percentage error in range of (0.606–2.243 % and 0.103–0.362 %) respectively. The results suggest that ANFIS models show better statistical prediction efficiency.