Quantitative Assessment of the Overall Heat Transfer Coefficient U

Abstract

This investigation was performed in order to quantify the validity of the assumed constancy of the overall heat transfer coefficient U in heat exchanger design. The prototypical twofluid heat exchanger, the doublepipe configuration, was selected for study. Heat transfer rates based on the U = constant model were compared with those from highly accurate numerical simulations for 60 different operating conditions. These conditions included: (a) parallel and counter flow, (b) turbulent flow in both the pipe and the annulus, (c) turbulent flow in the pipe and laminar flow in the annulus and the vice versa situation, (d) laminar flow in both the pipe and the annulus, and (e) different heat exchanger lengths. For increased generality, these categories were further broken down into matched and unmatched Reynolds numbers in the individual flow passages. The numerical simulations eschewed the unrealistic uniforminletvelocityprofile model by focusing on pressuredriven flows. The largest errors attributable to the U = constant model were encountered for laminar flow in both the pipe and the annulus and for laminar flow in one of these passages and turbulent flow in the other passage. This finding is relevant to microchannel flows and other lowspeed flow scenarios. Errors as large as 50% occurred. The least impacted were cases in which the flow is turbulent in both the pipe and the annulus. The general level of the errors due to the U = constant model were on the order of 10% and less for those cases. This outcome is of great practical importance because heatexchanger flows are more commonly turbulent than laminar. Another significant outcome of this investigation is the quantification of the axial variations of the temperature and heat flux along the wall separating the pipe and annulus flows. It is noteworthy that these distributions do not fit either the uniform wall temperature or uniform heat flux models.