Investigation of seal technology for Francis turbine
MetadataShow full item record
Leakage loss and disk friction loss caused by the clearance gap flow at the back of a runner have a major impact on the efficiency of hydraulic turbines. Accordingly, it is extremely important to develop and improve the seal technology by investigating the gap flow. Generally, there are two types of the gap flow: axial gap flow between a rotating disk and a stator (e.g. the flow at the back of a runner) and the annular gap flow (e.g. the flow at an annular seal). Firstly, the overview of previous researches on labyrinth seal, rotating disk flow and Taylor-Couette flow are summarized. Labyrinth seals are the primary type of seals for turbo machinery. However, most researchers studied it for compressible flow only. It is also found that the enclosed rotating disk flow with through-flow can be studied instead of the gap flow in a hydraulic machine. Furthermore, the above mentioned annular gap flow is similar to the Taylor-Couette flow. The Taylor-vortices are formed in the annular gap due to the rotation of disk, which could be used as a resistance of flow. Therefore, in the present work, three parts are investigated. The first part is the investigation of the labyrinth seal for Francis turbines. The second part is the investigation of the gap flow between two stationary walls. The third part is investigation of the annular gap flow between one stationary and one rotating wall, based on the theory of Taylor-Couette flow. Afterwards, the theoretical formulas for leakage flow of a traditional labyrinth seal used in high head Francis turbine is derived and is verified to give acceptable results. The theoretical model is useful to predict the leakage flow by the measurements of Francis turbine at the Åbjøra Power plant. For straight-through labyrinth seal, the effects of cavity dimensions, numbers and locations on the leakage flow are investigated numerically. Smaller cavity depth, longer cavity length and fewer cavity numbers are required in order to reduce leakage loss. Furthermore, the straight-through labyrinth seal with the bilateral cavity produce less leakage loss than with the unilateral cavity given identical cavity dimensions. However, the cavity location has only slight influences on the leakage flow. In addition, it is proven via experiments that the leakage flow rate is proportional to the clearance gap and pressure difference. It is also found numerically that vortices are formed when the flow enters a bigger cross section. These vortices consume the kinetic energy to a large extent, thereby reducing the leakage flow. Lastly, the experimental set-up of the rotor-stator system is built up as a simplification of a Francis turbine runner in the Waterpower laboratory at NTNU for two purposes. One is to investigate how the rotational speed of smooth disk affects the leakage flow rate, and the other is to study what are the effects of the geometries of rotating disk on the leakage flow rate and disk friction torque. With dimensional analysis, the function of flow coefficient and friction coefficient for the gap flow between one stationary and one rotating wall is derived. From the measurements for the smooth disk (SD), it is found that the leakage loss can be reduced by increasing the rotational speed due to the dominance of flow coefficient. While the leakage flow rises up due to the dominated pressure difference between the above and below the rotating disk. Comparing the measurements of the smooth disk (SD) and the disk with straight grooves (DSG), it is revealed that the geometry of the rotating disk has an impact on the leakage flow, and the performance of DSG is better than SD. Stage I only exists during the range of the current testing for DSG. However, the geometry of the disk has only slight effects on the disk friction torque.