The 11th Asian International Conference on Fluid Machinery and The 3rd Fluid Power Technology Exhibition

Paper AICFM_TM_015 PaperID: number xxxxxxxxx

November 21-23, 2011, IIT Madras, Chennai, India

Original Paper

Development of Francis Turbine using Computational Fluid Dynamics Kiran Patel1, Jaymin Desai2, Vishal Chauhan2 and Shahil Charnia2, 1

2

Manager, CFD Analysis Center, R & D Hydraulics, Jyoti Ltd., [email protected], [email protected] Executive, CFD Analysis Center, R & D Hydraulics, Jyoti Ltd., Nanubhai Amin Marg, Industrial Area, P.O.Chemical Industries, Vadodara-390003, Gujarat, India. Abstract

High head Francis Turbine design (Ns=80 to 90) is very complex in nature. Initial basic design of all the components of Francis turbine is completed based on basic fluid dynamics turbo machinery principal and from our experience database. 3-D steady state, single-phase CFD analysis is conducted for all the components (Runner + Spiral Casing + Guide vane + Draft tube) of Francis turbine and losses along with fluid flow behavior is checked. Francis Runner design is optimized using CFD analysis to get required performance. Performance of turbine is predicted at BEP as well as at part load and it shows improvement over older design. Cavitation analysis is also conducted to check the performance of design. Model will be manufactured & tested and results are covered. CFD is very useful and reliable tool to reduce product development cycle time and cost. Keywords: Francis Turbine, CFD Analysis of turbine, BEP (Best Efficiency point)

1. Introduction Renewable energy sources are the key area, which brings green revolution in India. It helps India to fulfill its power demand and maintain its economic growth for next decade. Small hydropower is one of the main renewable energy resources and India has good potential in it. There are four basic type of turbine (Francis, Kaplan, Pelton and Turgo) is used for hydropower generation. Computational Fluid Dynamics (CFD) is latest state of art technological tool, which is used by designer for hydraulic design of Turbine. This paper describes how high head Francis turbine is designed using CFD. It starts with basic design of each component (Spiral Casing, Guide vane, Runner, Draft tube) and optimization & performance evolution using CFD. Basic turbine design optimization loop is as shown in Fig. 1.

2. Basic Design of Francis Turbine components Francis turbine consists of four basic components i.e. spiral casing, guide vane, runner and draft tube. Design of Francis turbine starts from Runner design. After completion of the runner design, other components are designed as shown in Fig. 2.

2.1 Runner Runner is most important component of Francis turbine. High-pressure water is entered at runner inlet with partial kinetic energy and flow through area between two blades of runner. Hence, by reaction principle water pressure energy is converted into kinetic energy, which produces pressure difference between two sides of blade. This pressure difference between two sides of blade produces torque and finally mechanical energy in the form of rotation of runner. Basic dimension of runner is calculated from our experience database and using literature [1.]. After freezing basic dimension i.e. diameter, width at inlet and outlet of runner, blade angle is calculated. Accepted for publication. Corresponding author: Kiran Patel, [email protected]

1

Basic Design of Turbine by references, experience and 2D calculation

3D Model Generation

Grid Generation

Problem Definition and solution by Navier Stokes equation

Performance evaluation (Power, Head, Discharge, Efficiency)

Mechanical Design

Design Validation by Model Test

Prototype Design Fig. 1 Turbine design optimization loop

Runner Guide vane

Spiral casing Draft tube Figure 2: Design sequence of Francis Turbine components

Velocity triangles at inlet and outlet are as shown in Fig. 3(a). These basic dimensions are used for making 3D model of runner as shown in Fig. 3(b). This 3D model of runner is used for CFD calculation for optimization and performance evolution purpose.

2

Figure 3: a. Velocity Triangle

b. Runner 3D model

2.2 Guide vane Function of guide vane is to convert partial pressure energy from available total pressure energy into kinetic energy. It develops the swirl, which is required at the inlet of turbine runner. It also distributes equal discharge at inlet of turbine runner and can govern discharge as per load requirement by changing its opening. Positioning of guide vane is based on inlet velocity triangle as shown in Fig. 3(a). Height of guide vane is based on runner height. Inlet and outlet diameter of guide vane housing is based on experience database.

2.3 Spiral Casing Function of spiral casing is to equally distribute discharge at the inlet of guide vane. At the exit of spiral casing, stay vane ring is fitted. The purpose of stay vane is to give mechanical strength to spiral casing as well as provide required flow direction at inlet of guide vane. Two basic design methodologies are used for spiral casing design: first, is constant angular momentum and second is constant velocity. We have designed spiral casing based on constant angular momentum using in house developed code.

2.4 Draft tube Function of draft tube is to convert unused kinetic energy at the exit of runner into pressure energy and hence to increase the effective net head on turbine unit. Draft tube consists of three basic component i.e. cone, elbow and diffuser. Inlet cone converts kinetic energy into pressure energy. Maximum pressure recovery is possible in inlet cone by selecting proper angle of cone. Function of elbow is to guide the flow and change the direction of flow. Due to change in direction of flow and swirling component of velocity, secondary flow is developed which is optimized in small amount by accelerating the flow into elbow. Further pressure recovery is obtained in diffuser part, which is connected to tailrace. Draft tube design is completed based on our experience and in house developed code.

3. Optimization of Turbine components by Computational Fluid Dynamics Computational Fluid Dynamics is science of predicting fluid flow (Turbine, Pump), heat transfer, mass transfer (Perspiration, Dissolution), phase change (Freezing, Boiling, cavitation), chemical reaction(Combustion) and related phenomenon by solving mathematical equations that governs these processes using a numerical algorithm by means of computer based simulation. It can provide detailed information about what is happening in a process where fluid is in motion. The technique is very powerful and spans a wide range of industrial and non-industrial applications. Present day recent advances in computing power, advanced computer graphics, and robust solvers make CFD a cost effective engineering tool. Four basic steps involved in any CFD analysis i.e. modeling, meshing, problem definition and post processing of result. Detail steps carried out for present case is as follows:

3.1 Geometric Modeling Based on 2D initial basic design calculation, we have made 3D fluid flow model for each components. Commercial software's are used for making the entire model. Special purpose software that makes turbo machinery components easily and faster is used for runner and guide vane model. Fig.4 shows assembly of all components.

3.2 Meshing Numerical discretisation or meshing is most important step to carry out CFD analysis. Meshing is nothing but dividing fluid flow volume in small elements. Quality of CFD result depends on meshing of various components. We have used tetra elements along with prism elements at solid boundary for mesh generation. Prism elements are useful to capture boundary layer separation. Fig. 5 shows meshing of various components. Details of no. of nodes in each component are given in Table 1. Quality of mesh generated is as per the standard industrial practice, result is also checked, and they are independent from mesh effect.

3

Figure 4: Francis Turbine components

Prism Layer

Figure 5: Meshing of Turbine components

Sr. No. 1 2 3 4

Table 1: Element generated in each component Components Spiral Casing Guide vane Runner Draft tube

No. of Nodes 517268 675170 1505730 517268

4

3.3 Problem Definition Considering flow inside the various components of Francis turbine is steady in nature, steady state single-phase (only water) CFD analysis is computed. All the components have been used in analysis and complete system CFD analysis accomplished. At inlet of spiral casing total discharge and at outlet atmospheric pressure is specified as boundary condition as shown in Fig. 6. All other parts considered as a wall with no slip condition. Spiral casing, guide vane and draft tube are stationary components & runner is rotating component and solved with Multiple Frame of Reference (MFR) model. The entire component is attached with each other by domain interface. Runner is attached with guide vane and draft tube by frozen rotor interface. Shear Stress Transport (SST) is used as turbulence model with high resolution as advection scheme. Convergence criteria is Max 10-4 is considered with automatic time step. Computation is performed in full operating range from part load to over load (25% to 130%) with various guide vanes opening. Cavitation in Francis turbine is observed at part load. To predict cavitation, separate cavitation analysis with Rayleigh Plesset as cavitation model has been accomplished. In case of cavitation, homogeneous mixture of water and water vapor is considered as fluid.

Figure 6: Problem definition of Francis Turbine components

3.4 Result and Discussion

Efficiency (Nondimensional)

Francis turbine complete system CFD analysis has been accomplished in full operating range from 25% to 130% loading. This is accomplished by changing guide vane opening from partial to full open. Detail discussion on final optimized geometry result is as follows: Fig. 7 Shows performance prediction of Francis turbine from part load to full load. It shows that the performance of the turbine is improving up to BEP and later on reducing with further increase in load. Performance of turbine is improving as moving towards BEP from part load due to reduction of head loss in guide vane, draft tube, and runner while in case of spiral casing head loss increases. Head loss is governed by flow separation and frictional loss in components. Head loss in spiral casing is due to friction loss and as load increases, discharge increases, and hence frictional losses increase as shown in Fig. 7.

1.02 1 0.98 0.96 0.94 0.92 0.9 0.88

0

0.5

1

1.5

Operating Load (%)

5

Head Loss vs Loading

Head Loss (%)

10

casing Draft tube

8

Guide vane Runner

6 4 2

0 0

50

150

100

Operating Load (%) Figure 7: Francis Turbine performance prediction and head loss of components Head loss in guide vane is higher at part load due to shock and frictional loss and it reduces as load increases. Head loss in draft tube is higher at part load as well as at over load, only near to BEP region it is minimum. Head loss in runner is also higher at part load as well as at over load only near to BEP region it is minimum. In case of runner, higher head loss is due to flow separation as shown in Fig. 8. It shows velocity plot in blade-to-blade view of guide vane and runner. At 25% and 55 % loading condition, very high flow separation at runner is observed which is main cause for high head loss in runner in this region and performance of turbine is reduced. At 90% and 100% loading condition, flow separation at runner inlet is reducing significantly hence, performance of machine is improved.

25%

90%

55%

100%

Figure 8: Velocity plots at various loading condition for guide vane and runner

6

Flow separation is also taking place in draft tube because of diffusing section and change in direction of flow. Fig. 9 Shows flow separation region in draft tube at 25% and 100% loading condition. In both the condition flow separation is observed but at 100% flow separation region is less. At 25%, loading condition high head loss is observed in draft tube, which is because of mixing of high & low momentums and flow separation, hence pressure recovery in draft tube is less. At overload condition, circumferential component of velocity is higher which is not converted in to pressure recovery and hence higher head loss is observed. Flow separation in case of draft tube cannot be eliminated, we have to design draft tube such a way that flow separation is at minimum level and pressure recovery is maximum. Runner is receiving water at required velocity and swirl from guide vane. This high-pressure water enters into the runner and because of incidence; flow is being separated at inlet. Optimization of runner design is based on pressure distribution on runner blade. As shown in Fig. 10, two sides of blade have different pressure, which produce tangential force and torque. Pressure difference on runner blade is also known as blade loading.

25% Mixing Region of high and low momentums Flow separation region

100% Mixing region of high and low momentums

Figure 9: Velocity vector plots at various loading condition for draft tube As shown in Fig. 10 pressure difference on two sides of blade increases with loading, hence as the load increases, pressure difference increases. Because of high-pressure difference on both side of blade, high torque is also developing and hence power output increases. Pressure difference on two side of blade is governed by blade angle distribution. Pressure and velocity distribution inside passage of two blades depends on blade angle. Runner is optimized based on pressure and velocity distribution on blade and passage. Pressure distribution should be gradually reduced from inlet of runner to exit of runner as shown in Fig. 10 & 11.

7

Figure 10: Pressure distribution on two sides of runner blade

Figure 11: Pressure distribution on runner blade, hub, and guide vane

At the exit of runner, pressure is below the atmosphere and at part load; it is below the vapor pressure of water. When pressure is below the vapor pressure of water, vaporous bubbles are produced. This phenomenon is called as cavitation. To determine cavitation level, separate cavitation analysis has been carried out using Rayleigh Plesset cavitation model. These bubbles are developing a vapor

8

core in to draft tube as shown in Figure 12. When this bubbles collapse, high-energy pressure waves are generated which produce lots of noise and vibration. Cavitation produces pitting on runner and draft tube. Cavitation analysis also shows that flow inside runner is cavitation free.

Figure 12: Cavitation Zone

4. Conclusion Francis turbine design is accomplished considering CFD analysis as a tool for designer. It helps us to determine trend of hydraulic losses in various components. It helps us to determine various causes of hydraulic losses i.e. flow separation, friction losses but at this stage, it is difficult for us to separate them and find individual contribution. We have also predicted performance of new design in complete operating range. CFD helps to do better and efficient design of Francis Turbine. At present model is under manufacturing and will be tested soon.

Acknowledgement We are grateful to Jyoti ltd. as well as all the persons involved directly or indirectly without the help of which this task of design optimization of geometry of Francis turbine might not be possible.

References [1]. Nechleba, Miroslav, 1957: "Hydraulic Turbine their design and equipment", ARTIA Prague, Czechoslovakia. [2]. Kovalev, N.N, 1965, The National Science Foundation, Washington "Hydro turbine design and constructions".

[3]. Barlit, V.V., 2007,"Hydro turbine volume 1 & 2", Department of civil engineering, MANIT, Bhopal.

9