Durham University have embarked on a new research initiative in low pressure (LP) steam turbine exhaust hoods. An extensive literature review has highlighted the importance of applying representative conditions at inlet to the exhaust diffuser, specifically accurate total pressure and swirl angle distributions, in numerical simulations to generate flow fields within the hood in line with field data. With commercial sensitivity surrounding industrial designs for exhaust hoods and last stage blades, Durham University have developed a generic, research-level exhaust diffuser geometry and accompanying last stage blade to encourage the expansion of academic research in the field. Preliminary CFD calculations on the LP exhaust has shown the design produces a representative flow structure, capturing the primary sources of loss; the separation along the bearing cone and at the tip of the flow guide.

This paper summarizes the findings from research studies carried out over the last 30 years, to better understand the flows in steam turbine low pressure exhaust hoods and diffusers. The work aims to highlight the areas where further study is still required. A detailed description of the flow structure is outlined and the influence of the last turbine stage and the hood geometry on loss coefficient is explored. At present, the key challenge faced is numerically modeling the three-dimensional, unsteady, transonic, wet steam exhaust hood flow given the impractically high computational power requirement. Multiple calculation simplifications to reduce the computational demand have been successfully verified with experimental data, but at present there is no 'best-practice' approach to reduce the computational time for routine design exercises. This paper highlights the importance of coupling the exhaust hood to the last stage steam turbine blades to capture the interaction; ensuring the total pressure and swirl angle profiles, along with the tip leakage jet are accurately applied to the diffuser inlet. The nonaxial symmetry of the exhaust hood means it is also important to model the full blade annulus. More studies have emerged modeling the wet steam and unsteady flow effects, but more work is required in this area to fully understand the impact on the flow structure.

It has been widely recognized for some decades that it is essential to accurately represent the strong coupling between the last stage blades (LSB) and the diffuser inlet, in order to correctly capture the flow through the exhaust hoods of steam turbine low pressure cylinders. This applies to any form of simulation of the flow, i.e., numerical or experimental. The exhaust hood flow structure is highly three-dimensional and appropriate coupling will enable the important influence of this asymmetry to be transferred to the rotor. This, however, presents challenges as the calculation size grows rapidly when the full annulus is calculated. The size of the simulation means researchers are constantly searching for methods to reduce the computational effort without compromising solution accuracy. However, this can result in excessive computational demands in numerical simulations. Unsteady full-annulus CFD calculation will remain infeasible for routine design calculations for the foreseeable future. More computationally efficient methods for coupling the unsteady rotor flow to the hood flow are required that bring computational expense within realizable limits while still maintaining sufficient accuracy for meaningful design calculations. Research activity in this area is focused on developing new methods and techniques to improve accuracy and reduce computational expense. A novel approach for coupling the turbine last stage to the exhaust hood employing the nonlinear harmonic (NLH) method is presented in this paper. The generic, IP free, exhaust hood and last stage blade geometries from Burton et al. (2012. "A Generic Low Pressure Exhaust Diffuser for Steam Turbine Research,"Proceedings of the ASME Turbo Expo, Copenhagen, Denmark, Paper No. GT2012-68485) that are representative of modern designs, are used to demonstrate the effectiveness of the method. This is achieved by comparing results obtained with the NLH to those obtained with a more conventional mixing-plane approach. The results show that the circumferential asymmetry can be successfully transferred in both directions between the exhaust hood flow and that through the LSB, by using the NLH. This paper also suggests that for exhaust hoods of generous axial length, little change in C is observed when the circumferential asymmetry is captured. However, the predicted flow structure is significantly different, which will influence the design and placement of the exhaust hood internal "furniture."

The exhaust hood of a steam turbine is a vital area of turbomachinery research its performance strongly influences the power output of the last stage blades. It is well known that accurate CFD simulations are only achieved when the last stage blades are coupled to the exhaust hood to capture the strong interaction. This however presents challenges as the calculation size grows rapidly when the full annulus is calculated. The size of the simulation means researchers are constantly searching of methods to reduce the computational effort without compromising solution accuracy. This work uses a novel approach, by coupling the last stage blades and exhaust hood by the Non-Linear Harmonic Method, a technique widely used to reduce calculation size in high pressure turbine blades and axial compressors. This has been benchmarked against the widely adopted Mixing Plane method. The test case used is the Generic Geometry, a representative exhaust hood and last stage blade geometry that is free from confidentiality and IP restrictions and for which first calculations were presented at last year’s conference [1]. The results show that the non-uniform exhaust hood inlet flow can be captured using the non-liner harmonic method, an effect not previously achievable with single passage coupled calculations such as the mixing plane approach. This offers a significant computational saving, estimated to be a quarter of the computation time compared with alternative methods of capturing the asymmetry with full annulus frozen rotor calculations.

The exhaust hood of a steam turbine is an important area of turbomachinery research as its performance strongly influences the power output of the last stage blades (LSB). This paper compares results from 3D simulations using a novel application of the nonlinear harmonic (NLH) method with more computationally demanding predictions obtained using frozen rotor techniques. Accurate simulation of exhausts is only achieved when simulations of LSB are coupled to the exhaust hood to capture the strong interaction. One such method is the NLH method. In this paper, the NLH approach is compared against the current standard for capturing the inlet circumferential asymmetry, the frozen rotor approach. The NLH method is shown to predict a similar exhaust hood static pressure recovery and flow asymmetry compared with the frozen rotor approach using less than half the memory requirement of a full annulus calculation. A second option for reducing the computational demand of the full annulus frozen rotor method is explored where a single stator passage is modeled coupled to the full annulus rotor by a mixing plane. Provided the stage is choked, this was shown to produce very similar results to the full annulus frozen rotor approach but with a computational demand similar to that of the NLH method. In terms of industrial practice, the results show that for a typical well designed exhaust hood at nominal load conditions, the pressure recovery predicted by all methods (including those which do not account for circumferential uniformities) is similar. However, this is not the case at off-design conditions where more complex interfacing methods are required to capture circumferential asymmetry.

Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. This paper aims to highlight the importance of the accurate computational modelling of both the inlet and outlet exhaust hood boundary conditions. The computations presented are calculated using the public domain LP exhaust diffuser test case proposed by Burton in 2012. The original test case did not include the effect of tip leakage on diffuser flows, this paper describes the inclusion of tip leakage and the results are shown to be in-line with the outputs produced by other authors. The key advance in this paper is that calculations were conducted with a representative condenser pressure gradient caused by the temperature variation inside the condenser tube nest. It is shown that accurately modelling the exit boundary calculation has a large influence on the flow structure and a smaller influence on the pressure recovery inside the exhaust diffuser. This influence is smaller than that seen by other authors when including unsteady effects or accounting for the circumferential non-uniformity of the turbine exit flow but will need to be included in design calculations as diffuser design advances. Nomenclature C p = P 2 −P 1 P T 1 −P 1 Static Pressure Recovery c p