International Journal of Fluid Machinery and Systems Vol. 10, No. 1, January-March 2017
DOI: http://dx.doi.org/10.5293/IJFMS.2017.10.1.001 ISSN (Online): 1882-9554
Original Paper
Computational Design of Bifurcation: A Case Study of Darundi Khola Hydropower Project Ravi Koirala, Sailesh Chitrakar, Hari Prasad Neopane, Balendra Chhetri, Bhola Thapa
Turbine Testing Lab, Kathmandu University Dhulikhel-7, Kavre, Nepal,
[email protected], s
[email protected],
[email protected] [email protected],,
[email protected],
[email protected],
[email protected], bhola@ku.
[email protected]. edu.np np Abstract Bifurcation refers to wye division of penstock to divide the flow symmetrically or unsymmetrically into two units of turbine for maintaining economical, technical and geological substrates. Particularly, water shows irrelevant behavior when there is a sudden change in flow direction, which results into the transition of the static and dynamic behavior of the flow. Hence, special care and design considerations are required both hydraulically and structurally. structurally. The transition induced losses and extra stres ses are major features to be examined. The research on design and analysis of bifurcation is one of the oldest topics related to R&D of hydro-mechanical components for hydropower plants. As far as the earlier approaches are concerned, the hydraulic designs were performed based on graphical data sheet, head loss considerations and the mechanical analysis through simpli fied beam approach. In this paper, the multi prospect approach for design of Bifurcation, incorporating the modern day’s tools and technology is identified. The hydraulic design of bifurcation is a major function of dynamic characteristics of the flow, which is performed with CFD analysis for minimum losses and better hydraulic performances. Additionally, for the mechanical design, a simplified conventional design method as pre-estimation and Finite Element Method for a relevant result projections were used. Keywords: Hydraulic Design, Mechanical Design, CFD
1. Introduction Bifurcation of Penstock in the Hydropower plant is used to divide the flow into the two units for power generation. This division can be either symmetric or unsymmetric. The symmetric refers to the equal division of flow and unsymmetric refers to the unequal division of flow. The angle of bifurcation, losses due to it and structural support in the intense diverted velocity directions are some of the critical approach in its design. They are usually installed near the powerhouse where the maximum possible pressure prevails pr evails i.e. the static pressure and the t he surge pressure, which have an extra ordinary hydraulic and mechanical behavior related to vibration, power swings, propagation of pressure hence hence both the considerations are important. [1] In this study, an under construction site, Daraudi Khola Hydropower Project in Gorkha district of Nepal has been taken as the design case. The detail specification regarding the site is mentioned in the table Table 1 Specification of Daraudi Khola Project Specification Head Main Flow Each Flow Penstock diameter Length
Description 63.8 11.32 1.32 5.66 2.1
Unit m m /s 3 m /s m
298.77 + 16
m
Received June 9 2015; revised July 2 2015; accepted for publication October 17 2016: Review conducted by Yoshinobu Tsujimoto. (Paper number O15026J) Corresponding author: Ravi Koirala,
[email protected]
1
Designs of the water convey system usually comprises of the hydraulic and mechanical geometries. Most the small hydropower plants during their design phase prioritized the mechanical strength design in the case of branching. However, both mechanical and hydraulic issues are equally important in ensuring the hydraulic function and mechanical rigidity of the system. A series of design process for a hydraulically efficient and mechanically rigid bifurcation system was designed in this study. Penstock bifurcation requires special design that is unique from proj ect to project. [2]
Fig 1 Bifurcation of Penstock [3]
The nature of flow and the changes after the division is one of the major hydraulic behaviors in the conduit. The vibration problems caused due to unnecessary eddies developed can cause huge head losses. These losses will decrease the potentiality of the plant. Malik et. al. in 2009 explained the importance with Computational analysis of hydraulic behavior in Trifurcation for Madi Khola Hydropower Project, Kaski Nepal probably for the first ti me in Nepal. [4] The stress magnitude in a pipeline is usually 3–9 times greater than in regular pipeline shells. For this reason special reinforcements are provided in order to decrease the stress concentration in crucial spots. The penstocks of hydropower plants built in the first half of the twentieth century are rarely equipped in such kind of reinforcement. The lack of reinforcement can result penstock failure, especially under sudden pressure rise conditions. The rupture of the penstock at Lapino Power Plant, Poland is an example of the stress induced failure caused by improper design. [5]
2. Need of Bifurcation in Penstock Particularly, there can be many important reasons behind the need and use of bifurcation in the hydropower plant [2, 6]. These reasons are discussed below: i. Flow variation There can be significant amount of flow variation in the water bodies in an Annual Seasonal Cycle [ASC]. Since the part flow and part load operation of turbine results into efficiency deterioration along with the maintenance issues on continuous operation in off design condition, a number of units are installed. Hence, instead of operating a turbine in 20% flow with single unit, one of the two units will be closed and more efficiency can be attained along with longer machinery life. ii. Economical perspective of penstock An alternative for installation of two units without bifurcation is using two penstocks but this doubles the cost of penstock installation, hence bifurcation is preferred. iii. Maintenance perspective If bifurcation is included in a plant, then cyclic operation of plant based on priority can be done without completely shutting down the entire system for maintenance activities. iv. Geographical perspective The topography of the area and the gradient in the site may not allow the installation of larger penstock or multiple penstocks hence a bifurcation would be required to install multiple units. v. Technical specifications In case of reaction and impulse turbine, the scenario may not be similar. In some rare cases, the capacity of the turbine and the potentiality of the site may be different. In such cases, installation of multiple units may be assisted by it. In other cases, there may be effect of the size of penstock or material used in the penstock to select the branching in them.
3. Hydraulic design of Bifurcation Wyes branches must be designed for smooth hydraulic flow to avoid excessive head loss, vibration, and cavitation. They must be geometrically detailed for proportional flow distribution, eliminate acceleration or deceleration of flow in the adjoining branches, and thus minimize head loss. Head loss in the penstock, including losses in wyes and branches, contribute to inefficiencies in the
2
power generation system a nd may result in lost generating revenue or in the case of pumped storage projects, additional pumping costs. Angle of bifurcation, ratio of cross sectional area, type and shape of bifurcation, flow, velocity and Reynolds’s number are some of the major factors governing head losses. Approximation of these parameters using set of equations at two dimensions may not be relevant to determine the effectiveness. So far the practices are concerned, often hydraulic design (angle) of bifurcation are prepared based on the flow ratio referencing the graphs resulted from various researches. In some cases the graph may give a valid bifurcation angle (but it’s rare the cases match) but many others were designed on larger hydraulic losses. Based upon the series of experiments, graphical representation of the losses has been prepared. Graphical representation from the Miller experiments and Munich test are some of the major representations. Fig 2 shows the representation from Miller’s experiment.
Fig 2 Miller's Plot for Head Loss Coefficient in Symmetric Bifurcation [7]
3.1 Geometry development, Computational Model and Computational analysis The branching sections were only considered during the modeling and the mesh was prepared. Primarily, Mesh Independent Test with the predefined convergence criteria of 1% on the design was performed. Rest of the analysis was done on that size i.e. 2,018,417 nodes. Fig 3 represents the result of Mesh Independent Test. An important consideration during the analysis should be maintained during the selection of the length for computational analysis to obtain steady developed flow. Hence, an additional length of 15000 mm has been considered.
Fig 3 Mesh Independent Test Result o
o
After the predetermination of the size of computational model, mesh on the geometry of each angle from 45 to 70 was performed to determine the point of minimum head loss. Definition of the domain, boundary conditions and solver parameters are defined in Table 2.
3
Table 2 Boundary conditions and fluid features for CFD analysis
Domain Definition Fluid Water Density 1000 kg/m Morphology Continuous fluid Domain motion Stationary Turbulence Model Shear Stress Transport [8] Boundary Conditions Inlet Pressure 624000.366 Pa Outlet 1 Mass flow rate 5682.9 kg/sec Outlet 2 Mass flow rate 5682.9 kg/sec Wall Features Mass & Momentum No Slip wall Wall roughness Smooth wall Analysis type Steady State Analysis Convergence Control Min Iteration 1 Max Iteration 200 Convergence Criteria Residual type RMS Residual target 10 e-5
The head loss was determined by the Equation 1 for different angles. [9]
P V Pout i V hli = in + in − + outi ρ × g 2 × g ρ × g 2 × g
(1)
o
From the analysis at varying angle of bifurcation, it has been found that the minimum head loss is at 51 and other head losses were in irregular pattern. Angle of bifurcation is dependent on site and has no empirical relationship for least loss. The hydraulic behavior of water in a flow conduit system is dependent on static and dynamic conditions of flow along with the geometry , the deviation may result into the formation of vortices and turbulence in the wye. Hence, at the particular condition of 63.8 m head 3 o and 11.32 m /s flow, water exhibits best possible behavior at 51 . The selection process also largely depends on the geometry and the allowable length from turbine unit to the point of manifold based on civil constraints. Fig 4 is the pressure and velocity plot in o the bifurcation with angle 51 . The velocity at the inlet to the turbine is around 2.5 m/s, which eventually reaches to 6 m/s at the inlet of the spiral casing. The pressure generated from Fig 4 can be used in the structural analysis, but the actual design is performed on maximum head considering surge.
Fig 4 Pressure and velocity distribution in the flo w domain
Fig 5 represents the flow in the mid plane of the bifurcation. It depicts the operational scenario and location for the need of 5 special considerations during structural design. The point of division has wall and hence, faces the higher pressure of 6.24×10 Pascal in the region. Similarly, all points where the cross section gradually changes has variations in the flow property. The free stream of flow can also be seen in the conto ur plot.
4
Fig 5 Pressure distribution in the central plane and the point of maximum pressure 3.2 Single unit off condition Performance of the single unit in the case of any circumstances, like maintenance has been a major concern in many power plants. In this study, an estimation of the pressure before and after closure has been analyzed to eliminate the possible consequences.
Fig 6 Location of points of observation
Fig 7 Pressure distribution at the plane before and after closure of Outlet 2
This analysis was performed noting the velocity and pr essure conditions at outlet 1, outlet 2 and plane at the region of divisio n. During the analysis, outlet 2 was closed and the observations were made. It was found that there has been significant change at the point of division, which was slowly balanced by the length of the flow as it reached to a distance of 5000 mm from point of division. Fig 6 shows the flow property distribution and Fig 7 shows the pressure variation before and after closure of single unit. The head loss in case of single unit operation is 9.21 mm which is comparatively higher due to the effect of the accumulated water at the point of division.
4. Mechanical Design of Bifurcation 4.1 Identification of basic mechanical features The backgrounds like material properties and fluid properties were identified prior to the analysis. The Indian Standard guidelines for the selection of the material, its allowable stress and corrosion allowance have been allocated. Table 3 shows the background for the analysis.
5
Table 3 Structural Properties for analysis Particular Dimensions
Hydraulic Parameters
Internal Pressure External Pressure
Material Properties
Description ID Angle Design Head Flow Units Max. Design Head Max. Surge Head Total Head Buried Pressure [8m buried] Mild Steel of the standard Density Young’s Modulus Poisson’s Ratio Ultimate Tensile Strength Yield Strength
Allowable Stress [10]
σ allowable
Thickness [11]
Minimum shell thickness Corrosion Allowance Minimum standard thickness
Unit 2.1 m 51 Degree 63.8 m 3 11.32 m /s 2 63.8 m 26.2 m 90 m 435.8465 MPa MS IS-2062 3 7850 Kg/m 210 GPa 0.3 410 MPa 250 MPa 102.5 MPa 18.03 mm 1.5 mm 20 mm
4.2 Design of structural members The structural members were designed incorporating both the analytical method and the finite element method. Analytical calculations were used for pre-estimation while the finite element was used for optimized solution. 4.2.1
Simplified Curves Beam Method [12]
A penstock wye or branch connection usually has several stiffening beams to resist the loads applied by the shell of the pipe. The method incorporated rib shortening, shear deformation of the stiffener beams, and variable flange width. Although valid, this systematic design process is not efficient considering available modern computing methods. However, a spreadsheet can be used to drastically cut down the time involved with this design method. To analyze the wye and branch connections using beams, many simplifications and approximations are used [6]. The localized effect of structural discontinuities, restraints of the stiffening beams, foundation support, and dead load of the water filled pipe are neglected. End load effects and conicity of the outlet pipes are also neglected and considered to be small in comparison to the vertical load on the beams. Here a finite certain sections are considered and computation is prepared based on this approach. This analysis also includes this approach of approximation. The prime design processes include the deflection of the members AO and BO at point O is equal from Fig 8. This computation was performed considering the uniformly varying load on symmetric section as shown in Fig 9.
Fig 8 Application of stiffeners in the wye segment
6
Fig 9 Simplified beam method for computation of the stiffener
4.2.2
Finite Element Analysis
The finite element analysis design method provides a more complete representation of the penstock shell – stiffener system. It is becoming the standard design tool for penstock wye design in the hydropower industry [2,4]. The minimum shell thickness was determined using the analytical methodology for pre-setting the computational limit. Based upon which 3D CAD was developed to perform the unstiffened structural analysis for locating the critical locations with maximum stress and deflection. The analysis was performed locating the maximum internal and external pressure at the inlet and outlet. Although it is Fluid Structural system, the resultant force caused by surge exceeds the designed phenomenon hence design for maximum is important but the results and analysis in terms of stress and deformation are the similar phenomena. The computational models were applied with the load at different stiffened conditions and resultant Von misses stress and deformation were observed. [13] The primary dimensions were then modified to counter act on the resultant deflections and stresses. The result was obtained with iterative computation of the stiffer and girdle rings. The main constraint for the selection of the size of the support was the site consideration. Figure 10 shows the result of the finite element analysis. After an iterative computation for the sizes, 20 mm thickness at the point of division, 26 mm girdle thickness and 16 mm pipe thickness in the branched segments were identified. Both the manufacturability and the performance were considered during the analysis. Finally maximum stress of 100.97 MPa was noted which is below the allo wable stress derived in Tables 3.
Fig 10 Stress distribution in the geometry
During the analysis, simplified beam approach and analytical design methodology were used for assumption of the external reinforcement geometry and determining the minimum thickness. Numerical simulation aided the reliability of the design by visually expressing the critical locations and ascertaining the strength of the multiple reinforcements in the system, which is not possible with single method of design. No comparative study has been performed; rather an integrated approach selecting the reliable processes excluding major approximations was implemented.
7
4.3 Manufacturing and Installation The design was manufactured by Machhapuchhre Metal & Machinery Works, Pokhara. Fig 10 are some of the photos taken during manufacturing and installation.
Fig 11 Manufacturing and Installation of Bifurcation
5
Conclusion
Improvising the design practices for realistic results and overcoming the possible circumstances is the present day need. Hydropower has been a reliable green energy source, in which, a large sum of investment, lots of time and intellectual minds are invested. Hence, an optimum economical design for the reliability of plants is important. Practices changes with time, but past references are always required. Present day computational facility eases the design process excluding the tedious iterations involved in it. Selection of angle of bifurcation has always been a major issue which is solved by this computational facility additionally the conventional calculation of the support in two dimension increases the risk. Hence the major conventional design procedure with the modern day’s tools and technology will surely aid in strengthening the reliability of design.
Acknowledgments The authors acknowledge Nirvana Tech Engineering Solution Pvt. Ltd., Kathmandu and Turbine Design Services Pvt. Ltd., Lalitpur. We also extend our gratefulness to all the members of Turbine Testing Lab and Darundi Khola Hydropower Project for the help, suggestion and guidance during the activities.
Nomenclature hli V in V outi Pin
Head loss at branch i Average velocity at inlet Average Velocity at outlet in branch i Average pressure at inlet
Pouti g ρ σ all
Average pressure at outlet i n branch i Acceleration due to gravity Fluid Density Allowable stress
References [1] F. Flemming, J. Foust, J. Koutnik, R.K. Fisher, 2009, “Overload Surge Investigation using CFD Data ,” International Journal of Fluid Machinery and Systems, Vol. 2, No. 4, pp 315-323. [2] J. H. Bambei, 2012, “Steel Penstock,” Ed. 1 ASCE Publications. [3] K.S. Ltd.,2012, “Technical Highlights Vol. 5” Retrived 2014, from KOBE: http://www.kobelco-welding.jp/education-center/technical-highlight/vol05.html [4] R.K. Malik, P. Paudel, 2009, “Flow Modeling of the first trifurcation made in Nepal, Hydro Nepal: Journal of Water, Energy and Environment, Kathmandu, Nepal,” Vol. 5, pp. [5] A. Adamkowski, 2001, “Case Study: Lapino Power Plant Penstock Failure,” J ournal of Hydraulic Engineering, Vol. 7, pp. 547555. [6] A.W.W.A. Staff, “Steel Pipe: A Guide for Design and Installation” Ed: 4, American Water Works Association. [7] D.S. Miller, 1990, “Internal Flow Systems 2,” Cranfield: BHRA. [8] S. Chitrakar, M. Cervantes, B.S. Thapa, 2014, “Fully coupled FSI analysis of Francis turb ines exposed to sediment erosion,” International Journal of Fluid Machinery and Systems, Vol. 7, No. 3. pp. 101-109. [9] C.A. Aguirre, R. G. R. Camacho, “Head losses analysis in symmetrical tr ifurcations of Penstocks – High Pressure Pipeline Systems CFD,” AGHEM. [10] Bureau of Indian Standards, 1995, “Pressure Vessels”. IS 11639-2. [11] Bureau of Indian Standards, 1969, “Code for unfired pressure vessels,” IS 2825. [12] American Iron and Steel Institute, 1998, “Buried Steel Penstock”. [13] X. Huan, L. Houlin, T. Minggao, C. Jianbao, 20 13, “Fluid-Structure Interaction Study on Diffuser Pump With a T wo-Way Coupling Method” International Journal of Fluid Machinery and Systems, Vol. 6, No. 2, pp. 87-93.
8
