A study of an air convection pile, a thermosyphon permafrost protection device

An experimental investigation was conducted on an air thermosyphon permafrost protection device called the air convection pile. Velocity and temperature profiles were measured at several vertical locations on a 10 and a 20 foot air convection pile for three different operating ranges. The velocity measurement showed that a low frequency high amplitude oscillating flow existed throughout the flow field. The frequency and amplitude of this turbulence were measured at the maximum velocity points in the flow field. Some effects of summer winds were also simulated. These showed that wind forced convection in the summer could be a significant problem. Measurements during reverse flow operating conditions were also taken. When this occurs, ambient air enters the pile exits and leaves through the inlets. A theoretical investigation was also conducted.

A theoretical model was developed to analyze several of the proposed causes of reverse flow. This model consisted of numerical solutions of the governing differential equations for conduction and radiation heat transfer in the start-up conditions prior to beginning of air flow through the pile. The integral continuity and energy equations were solved using experimental data to calculate the rate at which heat was transferred to the air and the mass flow rate of the air through the air convection pile. The governing boundary layer equations were solved using the model developed by Reid and Tennant. This model was improved based on information obtained during the experimental phase of this research. In the few applicable cases the experimental results compared well with the test results of other studies. The theoretical investigation eliminated most of the proposed causes for reverse flow and indicated that there was only one general cause for reverse flow. The results of the improved Reid and Tennant model were generally within ±20% when predicting the total heat transferred to the air. The velocity and temperature profiles generated were generally different from those measured experimentally showing a more rapid rate of development.