Evaluation on the heat transfer performance of a vertical ground U-shaped tube heat exchanger buried in soil-polyacrylamide

• Published in: Experimental heat transfer Vol. 30; no. 5; pp. 427 - 440
• Main Authors: Wan, Rui, Chen, Meiqian, Huang, Youwang, Zhou, Tong, Liang, Bin, Luo, Haifeng
• Format: Journal Article
• Language: English
• Published: Philadelphia Taylor & Francis 03.09.2017; Taylor & Francis Ltd
• Subjects: Backfill; Blending effects; Energy conservation; Fluid flow; ground heat exchanger; Ground source heat pump; Heat exchangers; Heat transfer; Heat transfer coefficients; polyacrylamide (PAM); Reynolds number; soil temperature; Soil water; Soils; Water heaters; Water temperature
• ISSN: 0891-6152; 1521-0480
• DOI: 10.1080/08916152.2016.1276647

Ground source heat pump (GSHP) systems have been applied widely because of their environmental-friendly, energy-saving, and sustainable nature. In this work, heat transfer performance of a single vertical small-scale U-shaped tube ground heat exchanger under hot climatic condition is addressed considering the influences of inlet water temperature, Reynolds number, and backfill materials (raw soil; soil-polyacrylamide (PAM) blend (0.27% blending ratio for PAM). The backfill materials had an important effect on the heat transfer of the ground heat exchanger. At an inlet water temperature of 45°C and Reynolds numbers of 3104 and 4656, the temperature drops of water in the tube in the soil-PAM blend increased by about 0.3 and 0.4°C compared to that in the raw soil. Within Reynolds number from 3104 to 6208, the average surface heat transfer coefficients of the water in the tube in the soil-PAM blend and raw soil at an inlet water temperature of 45°C were 411 and 231 W m −2 K −1 , respectively. The results suggest that adding the PAM into soil can be an effective manner for enhancing the heat transfer of the ground heat exchanger. The dimensionless surface heat transfer correlation of the water in the U-tube heat exchanger in the soil-PAM blend was obtained. The model could better fit the experimental data within ±10% deviation.