Computational fluid dynamics (CFD) methods are applied to the analysis of a low-pressure turbine exhaust hood at a typical steam power generating station. A Navier-Stokes solver, capable of modeling all the viscous terms, in a Reynolds-averaged formulation, was used. The work had two major goals. The first was to develop a comprehensive understanding of the complex three-dimensional flow fields that exist in the exhaust hood at representative operating conditions. The second was to evaluate the relative benefits of a flow guide modification to optimize performance at a selected operating condition. Also, the influence of simulated turbine discharge characteristics, relative to uniform hood entrance conditions, was evaluated. The calculations show several interesting and possibly unique results. They support use of an integrated approach to the design of turbine exhaust stage blading and hood geometry for optimum efficiency.

This article describes the effects of two methods for representing the nonuniform distribution of flow properties across a steam turbine discharge annulus, on the hood loss coefficient. One method uses a mass-weighted integration of the property across the station, while the other is based on a mass-derived representative value of the property. The former has the potential for very high accuracy provided a sufficient number of points are integrated. The latter, while less accurate, is easier to apply and therefore more commonly used. The analytical modeling includes a simplistic step profile of pressure across the annulus, as well as a three-dimensional exhaust hood, flow-field simulation calculated using a Navier–Stokes code. Results show that significant errors can occur in the hood loss coefficient with the mass-derived approach. Although the analysis centers on hood loss coefficient as the performance parameter whose sensitivity is being monitored, the results highlight the pitfalls of improper application of measured data for any internal flow system.