Computational fluid dynamics (CFD) has been unable to reliably predict aeration and drag torque in clutches. In this study, a CFD method was developed to reliably predict the onset of aeration and drag torque as a function of the clutch's rotational speed. This study showed that though the oil and air behave as if they are incompressible at steady-state, the formulation must account for the compressible nature of the gas and the unsteady processes that occur before reaching steady-state. This study also showed that the dynamic nature of the contact angle between the oil and the stationary disk must be accounted for to predict drag torque and aeration as a function of rotational speed, and a model of the dynamic-contact angle was developed.

Vacuum and low pressures are needed in many applications, and the liquid-ring vacuum pump, which does not have any solid-solid contacts between moving and stationary parts, is widely used because of its low operational cost and long service life. Though progress has been made in advancing this pump, industry still has aggressive goals on improving its efficiency and performance. In this study, a reduced-order model was developed to predict the ability of liquid-ring pumps to ingest air and thereby create lower pressure as a function of pump design and operating parameters. The model developed is semi-empirical — constructed by first analyzing available experimental data to extract features and trends and then encapsulating them into a model through appropriate dimensionless parameters. This model by being in closed form shows the functional relationship between the pump’s design and operating parameters and its ability to ingest air and create a vacuum. To make predictions, this model only requires the following inputs: suction pressure, impeller’s rotational speed, and a few dimensions of the pump. The model developed was assessed by using it to predict the ability of the pump to ingest air for a wide range of suction pressures (cavitation pressure to 760 torr), rotor speeds (up to 1,750 rpm), and dimensions of the pump (radius and span of the impeller blade, hub radius) and then comparing predictions with experimental data not used in the creation of the model. The model developed was found to be accurate within 11% of the experimental data.

When a liquid is forced to flow radially outward in the gap between two coaxial, parallel annular disks—one rotating and one stationary—the liquid occupies the entire gap until the speed of the rotating disk reaches a critical value. Beyond that critical speed, gas from the outer radius starts to enter into the gap, a process referred to as aeration. The higher the rotational speed, the greater is the extent of penetration by the gas into the gap. The extent of gas penetration strongly affects the torque exerted between the two disks because of the large difference in the gas and liquid viscosities. In this study, a reduced-order model is developed to predict the onset of aeration, extent of gas penetration into the gap, and drag torque as a function of the disk's rotational speed, gap between disks, properties of the liquid, and mass flow rate of the liquid forced through the gap. The model developed was validated by comparing predictions with experimental data.

In gas turbines, pressure variations induced by the rotor’s rotation and created by the stator and rotor-stator interactions cause hot gas to enter into the wheelspace (i.e., the region between the stator and rotor disks). Purge flow is used to prevent the hot gas from entering, and rim seals are used to minimize the amount of purge flow needed. Preventing hot-gas ingestion is essential to protect the structural integrity of the rotor and stator disks. In this study, a reduced-order model is developed to predict hot-gas ingestion for a given purge flow and for an axial and a radial rim seal and to guide the design of rim seals. The model developed is semi-empirical — utilizing available data in the literature and recognizing certain trends and universal properties in the data if proper dimensionless parameters are used. The model has the advantage of only requiring the following inputs to make predictions: rotor’s rotation speed, flow rate in the turbine’s hot-gas path, purge flow rate, and dimensions of the seal and the wheelspace. Thus, this model differs from existing models in that an experimental input on parameters such as discharge coefficient, minimum sealing flow rate, or mixing length across the rim seal are not needed. The model developed was assessed by using it to predict hot-gas ingestion for a range of purge flow rates (0 to minimum needed to prevent ingestion), rotor speeds (2,000 to 3,500 rpm), and dimensions of the seal geometries (amount of overlap in the seal and width of the flow path through the seal) and then comparing the prediction with experimental data not used in the creation of the model.

Cooling the trailing edge of a gas-turbine vane/blade typically involves an embedded wedge-shaped duct in a very confined volume, where coolant enters the duct radially and then directed to flow axially by ribs and pin fins to cool the entire trailing-edge region as efficiently and uniformly as possible. CFD simulations based on steady RANS – compressible formulation with temperature-dependent properties closed by the shearstress transport turbulence model – were performed to study the flow and heat transfer in a wedge-shaped duct with ribs and pin fins under rotating and non-rotating conditions. The objective of this study is twofold. The first is to understand the flow mechanisms by which ribs and pin fins turn radially outward flow to flow in the axial direction. The second is to understand what features of the flow and heat transfer obtained under laboratory conditions – where measurements have been made to validate this computational study – can be extrapolated to engine-relevant operating conditions. Results obtained show pin fins judiciously placed around the turn of the wedge-shaped duct to greatly reduce the size of the separated region when the coolant jets radially into the wedge-shaped duct. Also, pin fins provide flow resistance to control the flow direction, and the uniformity of the flow along the cross section of the wedge-shaped duct in addition to enhancing surface heat transfer via horseshoe vortices about each pin fin. The staggered array of ribs along the radial direction of the wedge-shaped duct was found to create up to three sets of recirculating flows that cause the flow entering the duct in the radial direction to spiral towards the axial direction with one created by the stagnation region upstream of each rib, one by the separation at the downstream edge of each rib, and one by the cavity-like flow between the aforementioned recirculating flows. When there is rotation, the staggered array of ribs was found to minimize the adverse effects of centrifugal buoyancy by confining flow separation to be between the ribs on the leading face. On what could be learned under laboratory condition that are meaningful for engine-relevant conditions, all of the aforementioned flow mechanisms are the same and the flow features are qualitatively similar. The exception is the size of the separated region at the tip of the wedge-shaped duct under non-rotating conditions, which is very large for the laboratory condition and very small for the engine-relevant condition.

Pardeshi, Irsha A. MSE, Purdue University, May 2013. Flow and Heat Transfer in an L-Shaped Cooling Passage with Ribs and Pin Fins for the Trailing Edge of a Gas-Turbine Vane and Blade. Major Professor: Tom I-P. Shih. Efficient and effective cooling of the trailing edges of gas-turbine vanes and blades is challenging because there is very little space to work with. In this study, CFD simulations based on steady RANS closed by the shear-stress transport turbulence model were performed to study the flow and heat transfer in an L-shaped duct for the trailing edge under two operating conditions. One operating condition, referred to as the laboratory condition, where experimental measurements were made, has a Reynolds number at the duct inlet of ReD = 15,000, coolant inlet temperature of Tinlet = 300 K, wall temperature of Twall = 335 K, a back pressure of Pb = 1 atm. When rotating, the angular speed was Ω = 1,000 rpm. The other condition, referred to as the engine-relevant condition, has ReD = 150,000 at the duct inlet, Tinlet = 673 K, Twall = 1,173 K, and Pb = 25 atm. When rotating, Ω was 3,600 rpm. The objective is to understand the nature of the flow and heat transfer in an L-shaped cooling passage for the trailing edge that has a combination of ribs and pin fins under rotating and non-rotating conditions with focus on how pin fins and ribs distribute the flow throughout the passage and to understand what features of the flow and heat transfer can or cannot be extrapolated from the laboratory to the engine-relevant operating conditions.