No description available.

\Yith the current develop111ent in computer technology and Computational Fluid D)'n<'tlllics techniques, t.he si11utlation within axial flow compressors becomes 1110re and 1110re pract.ical and beneficial to the compressor designs. Due to the insufficient capabilit)' of today's COll1put.ers for three-dimensional unsteady flow 1110delling of 111Ult i~Llg(' axial flow compressors, sophisticated models of steady state flow and perfor111ance 1110delling of the C0111prcssors deserve to be thoroughly investigated. In l1utltistage C0111pressor sinlulations with steady state methods, frame of reference is fixed on blades and the c0111putational domains for rotors and stators haye relati\'e rotation. One of the difficulties in such simulations is how to pass information across the interfaces between blade rows without losing continuity. Two 111ajor stead)' state modelling approaches, a mixing plane approach based on Denton's circu111ferentially non-uniform mixing plane model and a deterministic stress approach based on Adamczyk's average passage model, are investigated and compared with each other through the flow predictions of the third stage of Cranfield Low Speed Research Compressor at peak efficiency operating condition. In the deterministic stress approach, overlapped solution domains are introduced to calculate deterministic stresses in order to 'close' the time-averaged governing equation system and the influence of the downstream blade row of the blade row under investigation has to be imposed through the simulation of bodyforce and blade blockage effect of the downstream blade row. An effective method of simulating bodyforce and blade blockage effect has been developed and proven to be simple in programming. ConYentionally, boundary conditions are specified in CFD calculations based on experimental data or other empirical calculations. By taking advantage of the special flow features in rear stages of multistage axial flow compressors where each rear stage behaves like a repeating stage of its neighbouring stages in terms of flow pattern at the inlet and the exit of these stages, a repeating stage model has been developed aiming at significantly simplifying the boundary conditions when simulating rear stages of a multistage axial flow compressor with only mass flow rate and stage exit average static pressure required as global input. A computer simulation system 1'/ STurbo3D has been developed to investigate a11d assess different steady state simulation models within multistage compressor environment. It has been proven that with the mixing plane model M STurbo3D is able to predict flows in multistage low speed axial flow compressors with acceptable accuracy. Application of the repeating stage model to the third stage of LS RC shows that the prediction with this model has equivalent accuracy to the prediction with the conventional boundary setting, and proves that the repeating stage model is an effective alternative to the expensive complete compressor simulation. The deterministic stress model provides more information of rotor-stator interaction and slightly better performance prediction than the mixing plane model, but the benefits of the model is not significant when applied to low speed axial flow compressors.

No description available.

No description available.

No description available.

Approved for public release, distribution unlimited / This thesis presents a Computation Fluid Dynamics (CFD) analysis of the aerodynamic performance of circular and elliptical leading edges of compressor blades, with a range of leading-edge droop angles. Specifically, simulations were conducted, with a free stream Mach number of 0.65 to quantify the change in pressure distributions and boundary layer momentum thickness in the leading edge region for a range of incidences on a flat plate with various leading-edge ellipticity ratios, ranging from unity (circular) to 5.5. In addition, the impact of drooping the leading edge was analyzed over a range of incidence angles from zero to 13 degrees. Pressure distributions indicate that elliptical leading edges can eliminate separation bubbles at zero incidence. The results indicated that the minimum loss occurred at an ellipticity ratio of about 3 and elliptical leading edges with a droop slightly greater than the average incidence can significantly decreased aerodynamic losses over a wider range of incidences.

Approved for public release, distribution unlimited / A performance comparison between smooth-cast and rough-cast blades was conducted in a lowspeed multistage compressor. The purpose was to show that rough-cast blades can be used for initial performance tests of low-speed compressors. A baseline performance was established with smooth-cast epoxy blades and then compared to rough-cast aluminum blades. The pressure-rise coefficient versus flow coefficient and velocity triangles were used as the measure for comparison between the different blade types. The velocity triangles were constructed and compared across the span (hub-to-tip) of the secondstage. This was done by using two probes that could traverse radially along the blade and sense the flow velocity and angle relative to the compressor axis. Measurements were taken at six different locations from the hub-to-tip of the blade. The performance tests were conducted at four different throttle settings: near stall, open throttle, nominal operating point (NOP), and near the NOP. Results showed that although there were some differences between the smooth-cast and rough-cast blades, the overall performance was very similar. As a result, rough-cast blades could be used for initial performance tests or as the initial stages of a compressor in order to set up the flow for the evaluation of a new stage of smooth blades. The use of rough-cast blades in early stages would save the time and money needed to produce large numbers of high precision smooth-cast blades.

Approved for public release; distribution is unlimited / This thesis presents a simple, meanline analysis of the impact of blade roughness on the mass flow, work coefficient, and efficiency of a three-stage axial compressor as a function of the location of fouling. First, an extensive review is presented on the state-of-the-art of measuring compressor degradation and on the impact of roughness on loss and deviation in a compressor cascade. The performance of a baseline, three-stage compressor, which has hydrodynamically smooth blades, is predicted. Using this baseline geometry, the influence of roughness in the front, middle and rear stages is calculated using empirical data for the enhanced losses and increased deviation, with a stage stacking technique. Influence coefficients that relate percentage changes in one parameter to percentage changes in other parameters are calculated. This analysis predicts that the most sensitive parameter for predicting fouling in the front stages is the percentage change in mass flow and the most sensitive parameter for predicting fouling in the rear stages is the efficiency. / Lieutenant, United States Coast Guard

Thesis (B.S.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 1982. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING / Includes bibliographical references. / by David Roland Mustelier. / B.S.

No description available.