Coupled near and far field thermal plume analysis using finite element techniques

Abstract

The use of the open cycle cooling process for thermal power plants requires significant effluent discharges into aquatic environments. Both engineering and environmental considerations require accurate prediction of resulting temperature distribution in the receiving waters. Most predictive models have looked at one of two distinct regions of the discharge--the near or the far field--to the neglect of the other.

A methodology is developed in this work to combine the attributes of both near and far field models. A finite element far field code is utilized which calculates both the circulation and heat distribution over a large area of the domain. From the far field coarse grid, a semi-circular area is removed which corresponds to the near field region of the discharge. At the new edge of the domain, which represents the near-far field boundary, mass flux and temperature boundary conditions are specified which simulate both the discharge into and entrainment out of the domain resulting from the surface discharge jet.

Initial verification and testing of the model's characteristics is carried out in a hypothetical idealized domain. A more realistic verification is done at two prototype sites by comparing calculated results to previously acquired field data. The two sites are Millstone Nuclear Power Station (on Long Island Sound near Waterford, Connecticut) operating with two units and Brayton Point Generating Station (in Mt. Hope Bay near Somerset, Massachusetts) operating with three units on open cycle (existing conditions) and with four units on open cycle (proposed future condition). These comparisons suggest that the model can realistically describe the far field flow patterns associated with near field mixing thus making the model a useful tool in evaluating induced circulations, the source of entrained organisms, etc. These flow patterns are a direct function of the near field entrainment and discharge distributions which are specified as model boundary conditions and are thus easily calibrated and, if necessary, modified. Comparison between measured and predicted temperatures indicates that the predicted lengths and areas of isotherms are similar to measured lengths and areas. Predicted temperatures generally indicate greater dispersion than measured temperatures thus leading to overprediction of intake recirculation. Also, because boundary conditions on the near-far boundary have been assumed constant, the shape of predicted isotherms is not as responsive to changes in ambient current direction (e.g., tidal variations) as the measurements indicate.

Future efforts should emphasize grid and program coding refinement to improve computational efficiency, use of methods to reduce numerical dispersion and incorporation of time-varying near-far field boundary conditions.