A parametric study was performed of the charging thermal performance of a full-scale pipe diffuser in a single cylindrical stratified chilled water storage tank by applying factorial experimental theory to the results of simulations performed with a validated computational fluid dynamics (CFD) model. Dimensional parameters having the potential to influence charging inlet performance were identified and formed into dimensionless groups using the method of repeating variables. Parameters included: the inlet Richardson number based on inlet slot width (Ril), inlet Reynolds number (Rei), ratio of inlet width to diffuser height (l/hi), ratio of inlet diffuser height to tank radius (hi/RW), and ratio of diffuser radius to tank radius (RD/RW). Thermal performance was measured in terms of equivalent lost tank height (ELH). A full 2k factorial experiment of thirty-two simulations was performed and analyzed. Parameter ranges were: 0.05-2 for Ril, 500–5000 for Rei, 0.1–1 for l/hi, 0.005–0.05 for hi/RW, and 0.707–0.866 for RD/RW. Within these ranges, Ril, l/hi, and hi/RW were found to be of first-order significance, while Rei and RD/RW were not. Two-factor interactions involving Ril, l/hi, and hi/RW were also significant. Regression models of equivalent lost tank height as functions of Ril, l/hi, and hi/RW were developed. The predictive capabilities of the regression models were tested against the results of five additional CFD simulations having parameter values different from the 2k factorial experiment cases. On average, regression models predicted factorial experiment data to within 10% with maximum error of 30% to 60%, depending on the model.
All Science Journal Classification (ASJC) codes
- Building and Construction