Figure 10 – uploaded by Ole G Dahlhaug

Figure 12 shows VR at the runner inlet position along the plane of clearance gap. Thus this figure demonstrates the relative distortion of each velocity components at the runner inlet due clearance gap. Velocity components along the circumferential direction is presented. It can be seen that at the clearance gap, the radial component of velocity has increased by more than 3.5 times than that at the mid-span. The other two velocity components have reduced proportionality. Since magnitude of radial component is very small with respect to the tangential component, vector law of velocity components is still valid with such seemingly unproportioned change in VR for individual velocity components. Deviations in velocity components occurring towards the suction side (50-100% PP), matches with the location of crossflow and vortex filament being at the same location (Figure 11). This indicates that velocity components, at the runner inlet, are non-uniform, in both angular direction and span wise direction, due to disturbance of crossflow originating from the clearance gap. Figure 13 shows effects of clearance flow on the velocity conditions at the runner inlet. Based on PIV results, proportionate velocity triangles are drawn at position of mid-height and hub at runner inlet (Figure 7). It is seen that the distorted velocity component at runner hub has significant effects on inlet flow conditions. Close to hub tangential velocity is reduced by 26% and radial velocity is increased more than 3.5 times. This causes relative velocity at the hub to be increased by 3.8 times than that at the middle of inlet height. Such changes on tangential and relative velocity at runner inlet have considerable consequences on turbine performance.

Related Figures (13)

Figure 1. Sediment erosion damage: (a) in GV, (b) in facing plates [4] Sa a I S iiaasieialaaas Chen [1] has done a study on flow field in distributor of a high head Francis turbine and concluded that the exit flow from the guide vane cascade is neither uniform in circumferential direction, nor in span wise direction. Eide [2] has investigated the effects of head cover deflection on flow field and found that clearance gap induces leakage flow and vortices from trailing edge of GV. Brekke [3] has investigated the influence of the guide vane clearance gap on turbine efficiency. He measured a decrease of 0.5 % efficiency at BEP in a prototype with 0.5 mm clearance gap compared to the original runner with 0.3 mm clearance gap. These studies suggest that secondary flow occurs from the GV clearance gap and has significant impact on turbine performance.

Table 1. Reference turbine analytical design values 4.4 LVesIgn and optimization Of Jiow Cascade A symmetric section of reference turbine forms the flow cascade. Some of the researchers have considered a straight channel with single guide vane [14, 15], while others have considered a section of model turbine to investigate the passage flow conditions [1, 16]. Choice of channel profile depends upon objectives and requirements of the measurements. For the present study, three GV with two flow passages out of 24 passages of the GV is considered as a reference case. This configuration has a single GV inside the flow channel. Two outer GV forms as part of walls of cascade and middle GV guides flow in the channel. Thus this layout of the test setup is named as one GV test setup. Table | presents the relevant analytical values for the reference turbine at BEP. These parameters are used to develop the test setup for this study. This power plant also represents a typical case of projects operating under large sediment load in Himalayan basin [8]. Design and drawings of the prototypes are not available. An in-house tool named as “Khoj”, was developed to design a reference turbine with the aim to replace with the existing one [9]. Extensive investigations has been carried out on this reference turbine for optimizing hydraulic design to minimize sediment erosion without compromise in efficiency [10-13]. —eo4+tt ena

REE Ae eS ee: See: ee! Se TN Delos Ne Ce eee design of one GV _ cascade. This configuration re-creates the flow around one GV, inside equivalent two periodic flow channels of a Francis turbine. passage, inside the cascade, is 30 degrees from the turbine center. This be observed. TGV has been designed as an assemb are made of aluminum. The middle part of TGV is at is screwed with the middle part. The end part of TGV hickness to get desired clearance gap with respect to Cc Angular position covered by the flow © Outlet of Cascade y of multiple ransmission of the laser sheet, allowing excess to entire plane inside ached to the plexi part, and the end part of TG Runner inlet Section to optimize velocity omponents at runner inlet Figure 2. Reference design and sections for optimization flow passage is 1/12" size of turbine in angular direction. Figure 3 shows the cross section view of tl est setup along the plane of measurement. For the convenience of manufacturing, assembly and testin he setup is divided into parts shown. The test section contains the tes glass flow channel for optical excess to PIV system. Figure 4 shows section along GV chord. Position of clearance gap and taps for pressure measurements in the gap cz guide vane (TGV) inside a plex the cross section view of the te parts. Plexi part of TGV impar the flow channel. Other two par can be replaced with the similar parts of differe he flow channel wall. The test setup is installed at the facilities of the Waterpower laboratory in NTNU. The setup is mounted in a closed loop system consisting of pump, flow meter, and a pressure tank. Air pressure inside the pressure tank is maintained to atmospheric pressure. All the measurements for this study is conducted with flow rate of 58.2 liter per second. This is 30% of design flow for the same flow area in the prototype at BEP. This is also the maximum flow through the setup for avoiding the cavitation at the

Figure 5. Positions for pressure taps along GV walls PIA! TEsOIUULON CLOss-COMCldUON LOCMMIGUe WIUI SU7/o OVETIdp. Figure 6 shows the region of interest from PIV image along the clearance gap. Circumferential position respective turbine components from SV outlet to runner inlet, inside the flow channel, can be observed. Periodic position (PP) between the cascade walls represents 30° angular position in circumferential direction with respect to turbine center. It can be observed that the PP from 0% to 50% represents the pressure side of flow, and PP from 50% to 100% represents the suction side of flow respectively. Grid points for each interrogation area for PIV processing can also be seen. Unique velocity vector is obtained for each erid point with the time-averaged value for 100 image pairs. Running

eS Re ee er eee ee eee ee eee Pressure measurements are done along the TGV surface, close to the test section wall without clearance gap. 14 pressure taps with 2 mm diameter are inserted in the test section cover plate to measure pressure from both pressure and suction side of TGV (Figure 5). These measurements are done at eight different sections along the chord length. Symmetric positions at pressure and suction side of TGV are taken for the measurements at the respective section. Each pressure tap is connected to piezo-resistive pressure transducer through a plastic hosepipe. All the sensors are pre-calibrated against the dead weight calibrator. Measurement uncertainly was maintained to be below 0.05% at all the measuring points. An average of 2000 samples, for each pressure point, measured at 5 HZ, is considered for the pressure analysis.

mean convergence test and bad vector position analysis has confirmed that 100 number of image pairs are sufficient to keep the standard deviation of time averaged velocity of good vectors below 0.1% of mean velocity.

Figure 9 shows velocity distribution inside the flow cascade obtained from PIV measurements along the plane of GV mid-span. The basic flow phenomenon as stagnation point at the leading edge of GV, development of flow around pressure and suction sides and formation of wake at the trailing edge can be observed in the velocity contour. Lower velocities along the pressure side of GV and higher for the suction side matches the results from pressure measurements. At the trailing edge, flow from the pressure side has a negative pressure gradient and flow from suction side has a positive pressure gradient (Figure 8). This causes an unbalanced mixing of fluids with different pressure gradients at the trailing

Figure 10. Observation of vortex filament from clearance gap I EI IEEE EE Figure 11 shows the results of PIV measurements at the plane middle of 2 mm clearance gap. Contours of total velocity together with the velocity vectors can be observed. Close to the leading edge velocity vectors are close to that of main stream flow both in magnitude and direction. Stagnation of flow by the GV shaft can also be seen. Behind the GV shaft gradual development of crossflow and its mixing with main stream flow can be observed. As also indicated by the pressure measurements, the highest amount of crossflow, both in magnitude and direction, occurs at 60-70% of chord length. Mixing of cross flow with main flow seen in Figure 11 justifies the formation and development of vortex filament as observed in Figure 10.

Figure 11. PIV measurements of flow felid along clearance gap

“igure 13. Velocity conditions at runner inlet: (a) at mid-height, (b) at hub with 2 mm clearance gap

Figure 14. Erosion of Francis Turbine components: (a) Runner hub, (b) Guide vane [19 Several cases of severe erosion at the runner hub have also been reported by past authors [4, 19]. Figure 14 shows a typical case of sediment erosion at runner hub, together with erosion in GV walls at trailing edge [19]. Significant loss in turbine efficiency, due to erosion, has also been reported. On basis of the presented results, these phenomenon can be explained as follows. Symmetric NACA profile has been used to shape the GV geometry. Thus, higher pressure difference towards trailing edge of GV causes higher secondary flows and corner vortex. This causes higher erosion at walls of the GV trailing edge. Erosion in GV trailing edge induces strong crossflow, which reduces tangential velocity and increases relative velocity at runner hub. Losses in turbine efficiency can be correlated with reduction in tangential component and the erosion at runner hub can be correlated with increase in relative velocity.

Figure 10 – uploaded by Ole G Dahlhaug

Related Figures (13)

Connect with 287M+ leading minds in your field