An Improved Radial Impulse Turbine for OWC

lable at ScienceDirect

Renewable Energy 36 (2011) 1477e1484

Contents lists avai

Renewable Energy

journal homepage: www.elsevier .com/locate/renene

An improved radial impulse turbine for OWC

Bruno Pereiras a,*, Francisco Castro a, Abdelatif el Marjani b, Miguel A. Rodríguez a

a Energy Engineering and Fluid Mechanics Department, University of Valladolid, Paseo del cauce 59, 47011, Valladolid, Spainb Labo. de Turbomachines, Ecole Mohammadia d’Ingénieurs (EMI), University of Mohammed V Agdal. Av Ibn Sina, B.P. 765 Agdal Rabat, Morocco

a r t i c l e i n f o

Article history:Received 15 April 2010Accepted 17 October 2010Available online 18 November 2010

Keywords:Wave energyOWCRadial impulse turbineCFDFlow analysis

* Corresponding author. Tel.: þ34983184536; fax: þE-mail addresses: [email protected] (B.

(F. Castro), [email protected] (A.el Marjani), miguel

0960-1481/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.renene.2010.10.013

a b s t r a c t

Traditionally, wells turbines have been widely used in OWC plants. However, an alternative has beenstudied over recent years: a self-rectifying turbine known as an impulse turbine. We are interested in theradial version of the impulse turbine, which was initially proposed by M. McCormick. Previous researchwas carried out using CFD (FLUENT�), which aimed to improve knowledge of the local flow behavior andthe prediction of the performance for this kind of turbine. This previous work was developed witha geometry taken from the literature, but now our goal is to develop a new geometry design with a betterperformance. To achieve this, we have redesigned the blade and vane profiles and improved the inter-action between them by means of a new relation between their setting angles. Under sinusoidal flowconditions the new design improves the turbine efficiency by up to 5% more than the geometry proposedby Professor Setoguchi, in 2002. In this paper, the design criteria we have used is described, and the flowbehavior and the performance of this new design are compared with the previous one.

� 2010 Elsevier Ltd. All rights reserved.

1. Current status of air turbines

Wave energy power plants based on oscillating water columns(OWC) convertwaveenergy into low-pressurepneumatic power. AnOWCplant consistsmainly of a submerged air chamber connected tothe atmosphere through a duct where a turbine is installed. Thesuccessive sea water waves come into contact with the chamber,compressing and decompressing the air in it by the periodic motionof the oscillating seawater free surface. This periodicmotion createsa bidirectional periodic flow through the turbine. Under theseparticular operational conditions, and although both self-rectifyingturbines and conventional ones are characterized by unidirectionalrotation, the first ones show a better behavior. Here, it must be saidthat there have been proposals on wave energy devices usinga system of non-return valves for rectifying the air flow, togetherwith conventional turbines [1], but they are complex and difficult tomaintain.

The development of self-rectifying turbines has been problem-atic for two reasons: the geometrical design itself is complicated,and the designer must find the best possible solution to achieve thehighest efficiency of the entire system. The global efficiency of thewave energy plant depends not only on the turbine efficiency butalso on the performance of the chamber. This means that the

34983423363.Pereiras), [email protected]@eis.uva.es (M.A. Rodríguez).

All rights reserved.

turbine must provide the optimal pneumatic “damping” (pressuredifference across the turbine) so that the capture efficiency of theOWC chamber is maximized. In this study this point is notconsidered, the objective is only the turbine performance.

Different types of self-rectifying turbines have been proposedfor use in OWC plants, the wells turbine being the first one in 1976.Subsequently, impulse turbines were suggested as an alternative.There are two kinds of impulse turbine: axial and radial (Fig. 1).

The performance of the Wells turbine has been described inmany articles and reports [2e4]. All the research agreed about themain disadvantages of this turbine: narrow range of flow rates withgood efficiencies, poor starting characteristics, high speed opera-tion, high noise level and high periodical axial thrust. In order toovercome the drawbacks of theWells turbine certain modificationshave been tested: self-pitch-controlled guide vanes [5], variable-pitch angle blades [6], contra-rotating rotors [7,8], using differentchord blades [9] and geometry ratios [10]. One of the alternatives tothe Wells turbine is the axial impulse turbine with self-pitchinglinked guide vanes proposed in [11]. The site trials confirmed thesuperiority of this type of turbine over the Wells turbine [12].However, the moving guide vanes lead to maintenance and oper-ating life problems [13]. Therefore, an axial impulse turbine withfixed guide vanes was also studied.

There are reports which compare the Wells and axial impulseturbines with fixed guide vanes. A comparison between turbineperformances under irregular wave conditions is made in [14e16],and they show that axial impulse turbines are superior in runningand starting characteristics under irregular flow conditions.

Nomenclature

AR Characteristic areab Rotor blade heightCA ¼ DP=12rðv2R þ u2RÞ Input coefficient.

CT ¼ TO=12rðv2R þ u2RÞARrR Torque coefficient

PDIN ¼ 0:5rv2element Characteristic dynamic pressureDP Total pressure dropDPo Total pressure drop between settling chamber-

atmosphereq Flow raterR Mean radiusT PeriodTo Output mechanical torque

UR¼urR Circumferential velocity at rRvelement Characteristic element velocityvR¼ q/2prRb Mean radial velocitya Absolute flow angleb Relative flow angleh Efficiencyhrotor Rotor efficiencyh ¼ 1

T

R T0 T0u dt=1T

R T0 DPq dt Mean efficiency

z ¼ DP0=PDIN Loss coefficientr Air densityu Rotational speed4¼ vR/uR Flow coefficientF Flow coefficient amplitude; 4¼F sin(2pt/T)

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e14841478

In [17] an experimental work of the impulse radial turbine,proposed by [18], is made. The study shows that this kind of turbinehas an acceptable efficiency. The main advantages of the radialturbines are, according to [19], their low manufacturing cost, thehigh torque obtained due to the radial configuration and theirruggedness. Another advantage with regard to axial turbines forOWC is the lack of bidirectional axial trust which reduces thefatigue loads on the bearings. However, the radial turbine causesa high damping on the OWC.

In order to develop a high performance radial turbine, in [20]a turbine with pitch-controlled guide vanes was proposed, whichin terms of efficiency is better, but its manufacturing and mainte-nance is quite expensive.

The aim of this work is to improve the performance of a radialimpulse turbine with fixed guide vanes for OWC using CFD. Apreviously validated numerical model [21] and improved in [22,23],allows us to study in detail the flow through the machine and itssources of energy loss. Using these data, the blades and the guidevane profile have been modified by means using the one-dimen-sional analysis and the Euler’s equation for turbomachines. A newrelation between the setting angle of the guide vanes and the rotorhas been introduced to reduce incidence losses, a new blade profilehas been designed to reduce the flow instabilities and to increasethe torque obtained by means of increasing the flow deflection.This new geometry achieves a remarkable advance in turbineperformance.

Fig. 1. Radial impulse turbine.

2. Numerical model

In order to validate the numerical model, the turbine geometryused was that considered as Case 1 in [17]; hence forward, thisturbine geometry will be denominated M8 turbine, Fig. 2a. This isequipped with a single rotor of symmetrical blades (R), one row ofouter guide vanes (OGV), one row of inner guide vanes (IGV) and anelbow (Fig. 3). The main geometrical characteristics are brieflyindicated in Table 1. Details of the turbine geometry characteristicsand dimensions can be found in Ref. [17]. The guide vane profileused in this M8 model consists of a straight line and circular arc,Fig. 2a. The blade profile consists of a circular arc on the pressureside and an ellipse on the suction side.

The flow simulation is solved with FLUENT v6.3�, which uses thefinite volume numerical method for solving the Navier-Stokes equa-tions. Since the computational volume includes rotating components(u¼ 234 rpm), the sliding mesh technique was used in order tomanage the relativemovementbetween the rotorand the statorof theturbine. A hexaedrical unstructured grid of 500.000 cells is used.

The flow model solves the incompressible fluid conservationequations by using a segregated solver. The realizable kee turbu-lence model was used with the standard wall function. The timedependent term is approximated with a second order implicitscheme. The pressureevelocity coupling was recreated through theSIMPLE algorithm. The highest order MUSCL scheme has been usedfor convection terms discretization and the classical centraldifferences approximations for diffusion terms. The flow charac-teristics description is conducted by solving equations in the three-dimensional turbine geometry. However, in order to reduce vari-able storage and to improve numerical accuracy, we have reducedthe 3D calculation domain to a small angular sector with periodicboundaries (Fig. 4).

3. Global performances

3.1. M8 turbine analysis

Turbine performance under steady flow conditions wasobtained by CFD andwas evaluated as in [17]. Fig. 5 shows the CTe4and CAe4 characteristics for theM8 turbine. Fig. 6 shows the steadyefficiency. The curves show a big difference between the perfor-mances in exhalation and inhalation.

A detailed flow analysis inside the M8 turbine is made in [22,23]and taking into account these data we can state that:

Inner guide vanes (IGV): the IGV guide the flow towards therotor adequately during the exhalation. However, during the

Fig. 2. Radial turbine geometry. (a) M8 geometry [17]. (b) M16 geometry.

Fig. 3. Turbine sketch and reference surfaces.

Fig. 4. Periodic calculation domain and boundary conditions.

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e1484 1479

inhalation the IGV do not guide the flow towards the elbow verywell, which increases the secondary flow loss in that element.The guiding could be improved with longer vanes.Rotor: in exhalation, the blades do not guide the flow in a correctway because of the divergent passage between them. Moreover,the torque produced during exhalation is not as high as inha-lation due to the high incidence losses and because inflow is thenatural flow direction for a radial turbine, which corresponds toinhalation. In inhalation, the inter-blade passage is convergentand the guiding is good.Outer Guide Vanes (OGV): during exhalation, the flow approachvelocity vectors are mismatched with the leading edge angle ofthe OGV vanes. This mismatch causes additional energy losswhich is referred to as incidence or incidence losses. During theinhalation phase there are no problems.

Table 1Main geometrical characteristics.

M8, Case 1 [17]

Blade number Chord length Solidity Setting angle

IGV 52 50 mm 2.29 25�

Rotor 51 54 mm 2.02 19.8�/35.8�

OGV 73 50 mm 2.28 25�

It canbe concluded that important energy loss exists in the elbowand in the guide vanes, mainly in exhalation. Therefore, the modi-ficationsmade in theM8 turbine geometry aim to diminish this lossand increase the rotor torque. In the new geometry (henceforward,M16 turbine, Fig. 2b) the rotor blade and guide vane profiles, bothinner and outer, have been modified. The IGV length has also beenincreased. These changes are based in results given by a one-dimensionalmodel whichwas used to look for a better alignment ofthe flow with the blades and vanes. The deviation angle of the flowin the rotor was changed to improve the rotor performance inexhalation, as a result the inter-blade passage is almost uniform.

M16 (improved geometry)

Blade number Chord length Solidity Setting angle

34 71 2.54 20�

51 47 1.78 20�/25�

85 45 2.42 20�

Fig. 5. (a)Torque coefficient CT and (b) input coefficient CA.

Fig. 6. Steady efficiency.

Fig. 7. Flow angles (b, a) and

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e14841480

3.2. M16 turbine performance

Fig. 2 shows the differences between both geometries. Thenumber and geometry of IGV have changed. The number of IGV hasbeen reduced to avoid loss, mainly during inhalation. The new innerguide vanes profile used in geometry M16 consists of straight linesand circular arcs, Fig. 2b, the vanes are longer and the outer angle is20� in order to increase the rotor torque during exhalation. Theblade profile consists of circular arcs on the pressure side and ellipsearcs on the suction side. Furthermore, the inner and outer angles aredifferent. The geometryof theOGV is very similar to that of theM8. Itconsists of a straight line and a circular arc, but the inner angle is 20�

so that the rotor performance improves during inhalation. Thenumber of OGV is increased to improve the guidance.

Figs. 5 and 6 show the effect of the geometry on turbine perfor-mances under steady flowconditions. The rotor torquehas increased,as expected, due to the new angles of the elements, mainly duringexhalation. The input coefficient, CA, has also increased but the effi-ciency attained byM16 is higher in exhalation and it does not drop ininhalation. This way, turbine performance between the inhalationand the exhalation is more balanced. This fact is very importantbecause the turbine works alternatively between exhalation andinhalation.

geometry angles (b*, a*).

Fig. 8. Flow angles in the sections E, D, C and B.

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e1484 1481

Fig. 9. Loss coefficient, z. (a) Elbow, (b) IGV, (c) OGV.

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e14841482

4. Flow analysis

The next topic studied is the flow pattern inside theM8 andM16turbines, whichwill be performed via the analysis of the flowanglesin sections B, C, D and E (Fig. 3). The flowanglewith the geometricalangle of the turbine elements, vanes and blades, will be compared.In sections which involve fixed elements (sections B and E) the flowangle under study is the absolute one (a). In sections delimiting therotor (C and D) the relative flow angle (b) is also considered. Anglesa* and b* are the geometrical angles of vanes and blades, respec-tively (Fig. 7).

4.1. Inhalation

Section E: in this section the vanes are radial, so they do not haveany influence.

Section D: Fig. 8a shows that the OGV guiding is better in theM16 turbine than in the M8 as expected, due to the higher numberof vanes. Moreover, the geometrical angle is lower for the M16where a*D¼ 20�. Fig. 8b shows that the relative flow going into therotor is well adapted to the setting angle at high flow coefficient.

Section C: Fig. 8c shows that the flow going into the inner guidevanes from the rotor shares the same angle in both geometries.However, as the geometrical angle of the inner guide vanes isdifferent, incidence losses is more important in the M16.

Section B: it is important to point out that, in turbine M16,section B is out of the inner guide vanes, too. Therefore, Section B inM16 and M8 is in different places, although in both cases it isequivalent. The guidance made by the IGV of the M16 in inhalationis far better than the one made by the M8. This can be verified inFig. 8d, since for the M16 there is a difference of 5� between theflow angle and the setting angle. In the M8 the difference is 20�,which corroborates that the IGV guidance was poor.

Fig. 10. Rotor efficiency.

4.2. Exhalation

Section B: in this section the vanes are radial, so they do not haveany influence.

Section C: the IGV guide the flow in an adequate way in bothturbines, though efficiency is higher in the M16 than in the M8,Fig. 8h. This is because in the M16 the flow direction is betteradapted to the leading edge angle of the rotor blade, Fig. 8g, and itcauses smaller incidence losses and an increase in the rotorefficiency.

Section D: In Fig. 8f it appears that the absolute rotor outlet flowangle and the setting angle of the downstream outer guide vanesare similar in the M16 geometry. However, for the M8 there arelarger differences (25�) which means a rise in incidence losses.

Section E: the OGV guide the flow more efficiently in the M16than in the M8, Fig. 8e.

5. Loss analysis

In order to study the energy loss in fixed elements we have useduse the loss coefficient z. This coefficient is the relationshipbetween the total pressure drop in an element and the represen-tative dynamic pressure in it.

The loss coefficient (z) of the elbow can be seen in Fig. 9a, theyare very similar. The flow is better guided by the IGV in the M16than in the M8. However, as the number of guide vanes is lower inM16, there are important secondary flows.

In Fig. 9b there appears the loss coefficient of the IGV. In theM16the guiding vanes were designed with the main goal of benefitingto the rotor. During the exhalation phase, the IGV work appropri-ately because of the low number of guide vanes and the convergentevolution of the transversal section. However, during the inhalationthere are several drawbacks, the main one being incidence losses inits entrance.

From the analyses of Fig. 9c it can be said that the loss in OGV aremore important in the M16 than in the M8. This is the result of thedifferent number of guide vanes, since in the M16 there are 85,whereas in the M8 there are only 73. This increases both frictionand incidence losses, since increasing the number of OGV does notavoid the flow detachment at the leading edge of the OGV.

Efficiency of the energy exchange in the rotor is calculated byusing the relation:

hrotor ¼ rðUvuÞInlet�ðUvuÞOutletðP0ÞInlet�ðP0ÞOutlet

Rotor efficiency is depicted in Fig. 10. It can be observed that rotorefficiency drops during inhalation due to the stronger tip andsecondary flows. However, the turbine efficiency is maintained

Fig. 11. Mean efficiency and energy per cycle.

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e1484 1483

(Fig. 6). The main point is the efficiency increase during exhalationcaused by the improved angles relation.

6. Turbine performances under sinusoidal flow conditions

In order to clarify the turbine suitable for wave energy conver-sion, it is necessary to evaluate the turbine performance inconnection with an OWC under irregular flow conditions [16]. Inorder to evaluate the behavior under unsteady flow we use themean efficiency, h, [14]. The running characteristics of the turbineunder periodic flow conditions can be simulated by using thesteady flow characteristics of a turbine at a constant rotationalspeed by assuming quasi-steady flow conditions, [24]. Fig. 11 showsthe mean efficiency and the energy per cycle when the turbineworks under sinusoidal flow (period T¼ 10 s). It can be observedthat in the M16 mean efficiency is better than in the M8. For F¼ 1where efficiency is highest, it is 5% better. The energy per cycle ishigher in M16.

Energy per cicle ¼ZT

0

T0u dt

The better performance in the M16 is caused by the well balancedbehavior between inhalation and exhalation. This causes that theefficiency along a periodical and bidirectional flow was higher inthe M16.

7. Conclusions

In this paper, we have analyzed the optimization of the turbine’sdesign by means of a numerical model. This model, which waspreviously validated, allows us to study in depth the flow throughthe machine and its sources of energy loss. The new proposedgeometry means a remarkable advance in turbine performance.

The guide vanes are an important source of energy loss as muchin inhalation as in exhalation (mainly the inner ones). When theguide vanes are upstream of the rotor and they guide the flowtowards the rotor, the loss is mainly associated with the curvatureof the vane and the variation of the channel’s passage section. Thevane curvature is related to the intensity of the tip flow, whereasthe variation of the passage section is important in order to avoidflow disattachment in the inner part of the guide vanes. These twoaspects have more relevance in the inner guide vanes because flowvelocity is higher there.

Above all, the loss is important when the vanes are downstream.The curvature and variation of the channel passage section are alsoimportant in this case. Nevertheless, the determining factor of theloss, in this case, is the flow detachment which takes place in theleading edge of the vanes. This problem has been reduced by fixinga new relation between the setting angle of the guide vanes and therotor angles.

Increasing the length of the IGV in the radial direction hasproved to be quite positive, since it reduces the loss in the elbow.This reduction is based on the fact that the flow circulates in a moreordered way, and consequently the loss associated with secondaryflows is reduced.

With these results, we can deduce that from the point of view ofthe global behavior of the turbine, it is better to improve the rotorperformance than that of the guide vanes. Due to this, in the guidevanes of the proposed geometry, the loss increases with respect tothe initial geometry.

One of the main problems of the initial geometry was theinequality of its performance between inhalation and exhalation,the difference in efficiency reaches 10%. This is because the torqueobtained by the rotor in both modes is very different. The proposedblade profile is superior to the previous one because the torqueduring inhalation stays the same and increases notably in exhala-tion. Thanks to this, the efficiency in exhalation increases by up to9%, whereas in inhalation it remains approximately equal. Conse-quently, the mean efficiency under unsteady flow increases sensi-tively in the new geometry by around 4e5%.

Acknowledgements

Herewewant tomention that these researches are conducted aspart of the common project between the Fluid Mechanic andTurbomachinery research teams of both the University of Valla-dolid (Spain) and the University of Mohammed V-Agdal (Morocco)for the development of a project based on an OWC converter plant.This project is an AI of the Agencia Española de CooperaciónInternacional. This team would like to thank to group GR57 for thesupport provided.

References

[1] Ueki K, Ishizawa K, Nakagawa H. Output of electrical power from pneumaticwave power generation system with water valve rectifier. Proceedings of 10thISOPE conference, Seattle, vol. I, pp. 339e404; 2000.

[2] Raghunathan S. The wells turbine for wave energy conversion. Prog AerospaceSci 1995;31:335e86.

[3] Govardhan M. Numerical studies on performance improvement of a self-rectifying air turbine. Eng Appl Comput Fluids Mech 2007;1(1):57e70.

[4] Inoue M, Kaneko K, Setoguchi T, Raghunathan S. Simulation of startingcharacteristics of the wells turbine. ASME 4th fluids mechanics, plasmadynamics and lasers conference, Atlanta; 1986.

[5] Kim T, Setoguchi T, Kaneko K. The optimization of blade pitch settings of an airturbine using self-pitch-controlled blades for wave power conversion. Journalof Solar Engineering 2001;123:382e6.

[6] Setoguchi T, Santhakumar S, Takao M, Kim T, Kaneko K. A modified Wellsturbine for wave energy conversion. Renewable Energy 2003;28:79e91.

[7] Raghunathan S, Eves A, Whittaker T, Long A. The biplane Wells turbine.Proceedings of the OMAE conference, Houston; 1987. p. 475e9.

[8] Gato L, Curran R. Performance of the biplane Wells turbine. Journal of OffshoreMechanics and Arctic Engineering (OMAE) 1996;118:210e5.

[9] Thakker A, Abdulhadi R. Effect of blade profile on the performance of the wellsturbine under unidirectional sinusoidal and real sea flow conditions. Inter-national Journal of Rotating Machinery; 2007; doi:10.1155/2007/51598.

[10] Torresi M. Experimental and numerical investigation on the performance ofa Wells turbine prototype. In: Proceedings of the seventh European Wave andTidal (EWTEC), Oporto; 2007.

[11] Setoguchi T, Kaneko K, Taniyama H, Maeda M, Inoue M. Impulse turbine withself-pitch-controlled guide vanes for wave power conversion: guide vanesconnected by links. International Journal of Offshore and Polar Engineering1996;6:76e80.

[12] Santhakumar S, Jayashankar V, Atmanand M, Pathak A, Ravindram R, Seto-guchi T, Takao M, Kaneko K. Performance of an impulse turbine based wave

B. Pereiras et al. / Renewable Energy 36 (2011) 1477e14841484

energy plant. In: Proceedings of eighth international offshore ad polar engi-neering, 1, Montreal; 1998.

[13] Maeda H, Santhakumar S, Setoguchi T, Takao M, Kinoue Y, Kaneko K.Performance of an impulse turbine with fixed guide vanes for wave energyconversion. Renewable Energy; 1999:533e47.

[14] Setoguchi T, Shanthakumar S,MaedaH, TakaoM, Kaneko K. A reviewof impulseturbines for wave energy conversion. Renewable Energy; 2000:261e92.

[15] Kim T, Takao M, Setoguchi T, Kaneko K, Inoue M. Performance comparison ofturbines for wave energy conversion. International Journal of Thermal Science2001;40:681e9.

[16] Setoguchi T, Takao M. Current status of self rectifying air turbines for waveenergy conversion. Energy Conversion and Management; 2006:2382e96.

[17] Setoguchi T, Santhakumar S, Takao M, Kaneko K. A performance study ofa radial impulse turbine for wave energy conversion. Journal of Power andEnergy 2002;216(A1):15e22.

[18] McCormick M, Rehak J, Williams B. An experimental study of a bidirectionalradial turbine for pneumatic wave energy conversion. Mastering the Oceansthrough Technology 1992;Vol. 2:866e70.

[19] Dixon SL. In: BUTTERWORTH-HEINEMANN, editor. Fluid mechanics andthermodynamics of turbomachinery. 5th ed. Elsevier, ISBN 978-0-7506-7870-4; 2005.

[20] Takao M, Fujioka Y, Setoguchi T. Effect of pitch-controlled guide vanes on theperformance of a radial turbine for wave energy conversion. Ocean Engi-neering; 2005:2079e87.

[21] Marjani AE, Castro F, Rodríguez M, Parra M. Numerical modelling in waveenergy conversion systems. Energy 2008;33:1246e53.

[22] Pereiras B, Castro F, Rodríguez M. Tip clearance effect on the flow pattern ofa radial impulse turbine for wave energy conversion. ISOPE’09. Osaka, Japan;2009. ISSN 1098-6189.

[23] Pereiras B, Castro C, Marjani AE, Rodríguez M. Radial impulse turbine for waveenergy conversion, A new geometry. OMAE’08. Portugal: Estoril, ISBN 978-0-7918-4823-4; June 15e20, 2008.

[24] Inoue M, Kaneko K, Setoguchi T, Saruwatari T. Studies on the Wellsturbine for wave energy generation (turbine characteristics and designparameter for irregular wave). JSME International Journal 1988;31:676e82.