Design and Experimental Study on a New Concept of Preswirl Stator as an Efficient Energy-Saving Device for Slow Speed Full Body Ship

Transcription

1 2004 SNAME Annual Meeting Design and Experimental Study on a New Concept of Preswirl Stator as an Efficient Energy-Saving Device for Slow Speed Full Body Ship M. C. Kim(V), Pusan National University, H. H. Chun (M), Pusan National University, Y. D. Kang(V), Pusan National University Corresponding address: Professor M. C. Kim, Dept. of Naval Architecture & Ocean Engineering, Pusan National University, Busan , Korea, kmcprop@pusan.ac.kr A new type (biased) preswirl stator attached to 300K KVLCC model Abstract Procedures for the design, analysis and model tests of a new type preswirl, namely, the biased preswirl stator system for a 300,000 DWT VLCC are described. It is well-known that the preswirl stator, which recovers the propeller slipstream rotational energy, is selected as the most reliable energy saving device due to its simplicity and reliability. The proposed biased stator, with more blades on the port side than on the starboard side, effectively recovers rotational energy loss by taking the advantage of non-symmetric wake flow characteristics between the port and starboard sides of the propeller disk. To design the biased stators effectively, a design program based on lifting surface theory has been successfully applied to the propeller design and analysis. Resistance and self-propulsion tests are carried out with the designed stator-propeller system to verify its effect. A modified model test analysis method for the preswirl stator-propeller system is used based on the 1978 ITTC standard method. The experimental results show that a 5.7% efficiency gain for the ship with the biased preswirl stator with 4 blades can be obtained compared to the ship without a stator. This gain is almost equivalent to that of the existing symmetric stator with 6 blades, and hence it is considered that this biased preswirl stator has merits in terms of light weight, low cost and small volume etc 1. INTRODUCTION Recently, compound propulsors such as contrarotating propellers, ducted propellers, preswirl stator, etc., have been widely used as energy saving devices since fuel prices have increased. Contra-rotating propellers are reported to be the best among the energy saving devices in terms of the efficiency gain. However, a high cost and

2 difficulty in maintaining the shafting system have restricted their practical applications. For vane wheel propellers, the risk of damage on vane wheel blades is higher due to its larger diameter than the propeller and its free running condition. Many accidents in full-scale ships equipped with vane wheel propellers have been reported. The practical application of ducted propellers dates from the 1930 s when Spita, and more in particular Kort, showed that when the propulsor is heavily loaded quite substantial efficiency gains could be obtained by a ducted propeller. However, as the speed of the ship increases, the duct drag rapidly increases compared to the improvement in ahead thrust. Compared to the energy saving devices mentioned so far, the preswirl stator system has several merits such as a simple shafting system, a relatively less initial cost of installation, considerable efficiency gain, and high reliability. It is well known that a preswirl stator located upstream the propeller can improve the propulsion efficiency by reducing the rotational energy loss in the propeller slip stream. The concept of recovering the energy is not new (see Blaurock 1990). It is regarded and more often claimed that when additional devices for energy savings are used, reliability is an important issue. For example, it was reported that there were severe problems in reliability with the application of CRP or vane wheel, making ship owners reluctant to use energy saving devices although more than 5% efficiency gain can be achieved. As mentioned above, the pre-swirl stator has high reliability because of its fixed appendage and is smaller or the same size compared with the propeller located behind the stator. Some ships have been successfully equipped with preswirl stators, showing some efficiency gain, but its systematic design and theoretical or experimental results are rarely reported. Takekuma(1981) has conducted some basic research. He carried out the calculation of Stokes theorem and simple experiments for a practical blade wheel design which had been applied to a full-scale ship. Several fundamental studies for the preswirl stator were reported by the KRISO group(lee et al. 1991, Kim et al. 1993, Kim et al and Lee et al. 1994). These are basic and fundamental research on the phenomena of the mechanism of an efficiency gain with symmetric stator together with analysis method for model tests. In the present study, a new practical and effective style of preswirl stator, called a biased preswirl stator, is proposed to apply to a full-scale ship. The stator has been designed to have 4 blades; three on the port side and one blade on the starboard side. A study with other numbers of blades is also expected to confirm the bias effect in efficiency, for example two blades on the port side and one on the starboard, or three blades on the port side and none on the starboard side. Instead of such parametric investigations, which will be conducted in near future, a model test using the optimally designed stator, but reversed, has been carried out to evaluate the effectiveness of the present stator. It was observed from the experiment that the proposed stator with 4 blades gave almost the same efficiency as that of the existing one with six blades type (see Kim et al. 1994), indicating that this biased preswirl stator has such merits as light weight, low cost, and small volume. It is anticipated that the cost of the present biased stator is about 2/3 of that of a six blade stator since the cost of propeller and stator is normally proportional to weight. This paper describes the design methodology and the model test results conducted at the towing tank (L=100m, B=8m, D=3.5m) of Pusan National University. Analysis code with lifting surface theory was used for the design of the preswirl stator-propeller system. A non-uniform wake is used instead of a ship hull for the analysis of the propeller-stator system. The loading distribution on the stator blade is represented by an equivalent angle of attack for the determination of optimum stator pitch angle. For the installation of the stator to the aft-body, the end part of the stern of the standard VLCC, designed by KRISO as KVLCC, is slightly modified. The resistance test with and without the biased stator is carried out in the towing tank, and these data are later used for the performance analysis of the self-propelled test. The self-propelled test is also conducted with and without the stator to verify its effectiveness. The ITTC78 method is used for the analysis of model test results except for the analysis of full-scale wake scaling. The model test results are intended to be validated by a fullscale application to be conducted by Daewoo Shipbuilding and Marine Engineering in the future. 2. DESIGN CONCEPT OF A BIASED PRESWIRL STATOR In the case of the full-ship, the proposed biased stator has more blades on the port side since the rotational flow component is different on each side due to upward flows. An upward velocity is normally cancelled by the propeller rotational velocity on the starboard side while the velocity on the port side is doubled. This phenomenon is shown in the typical velocity vector profile just behind the propeller measured by LDV in Fig.1 (Stern et al. 1994). The preswirl stator is normally symmetric and has six blades, as shown in Fig.2; Blades appear to be missing at the top and bottom positions due to blockage effects against an on-coming flow and potential docking problems, respectively (Takekuma 1981). In the present study, the stator blades at top and bottom positions are omitted for the same reasons. The KVLCC, a standard 300,000 dwt VLCC (see Table 1 for principal dimension and Fig. 3 for lines and model) designed by KRISO, is adopted for the model ship. Wake measurements were carried out at the propeller plane by a five-hole pitot tube in KRISO, and tangential velocity components are shown in Fig.4. The phenomenon of unbalanced tangential velocity on the propeller plane is also shown by summing the measured 2

3 tangential velocity (see Fig.4) and the computed propeller induced velocity (see Fig.5) where the unbalanced remaining tangential velocity is clearly seen in Fig. 6. The biased preswirl stator that is optimally designed for the wake system, explained in more detail in the following section, is shown in Fig.7 effectively cancels the remaining tangential velocity. In the fullscale ship, the unbalance of tangential velocity between port and starboard sides is more severe, since the bilge vortex becomes weaker than that for a model due to viscous effects. Therefore, the present biased type of stator would be more effective in full-scale ship. The aftpart of the KVLCC model ship is modified to install the preswirl stator to the hull body as shown in Fig.8. The wake measurement after modification of the aft-part was not conducted in this study because the modification is small, and so the effect of this hull form modification on the wake is regarded as negligible. It is of interest to examine the performance of a reversely installed biased stator system that could be installed simply by turning the optimally designed stator by 180 degrees as seen in Fig.10. The present design stator, which has three blades on the port side and one on the starboard, is called port stator and the reverse case as starboard stator. The port stator is designed optimally for the ship s wake field, and therefore the reversed stator has non-optimum angles as well as nonoptimum blade positions. Fig.1 Typical tangential velocity measured by LDV on propeller plane behind Series 60 ship with propeller working (Stern et al. 1994) Fig.2 Profile of Mitsubishi style reaction fin Fig.3 300K KVLCC Lines and Model 3

4 Table 1 Principal particulars of model ship KRISO 300K VLCC Model Lpp(m) 6.4 B(m) 1.16 D(m) 0.6 C b λ r/r Fig.6 Summed total tangential velocity at 0.5R behind from a propeller plane blade4 blade3 blade r/r Fig.4 Tangential velocity component of 300K KVLCC on propeller plane without propeller working blade2 Fig.7 Proposed biased asymmetric designed stator (looking upstream) r/r Fig.5 Propeller induced tangential velocity without ship at 0.5R behind from a propeller plane Fig.8 Modification of 300K KVLCC Lines for the installation of the stator 4

5 Fig. 9 Flow chart of the design program 5

6 3. DESIGN BY LIFTING SURFACE METHOD The stator model designed with the new concept is shown in Fig.7 where the number of blades on the port side is three with one on the starboard side, and called port stator while the opposite case is called starboard stator as shown in Fig.10. Fig.12 shows the stern view of the model with the stator attached alone and also with the stator-propeller system. It can be easily noticed that this biased stator with 3 blades is compact, light in weight and small in volume compared to a conventional stator with 6 blades (Takekuma et al. 1981, Kim et al. 1994). For reference, the model ship with the propeller and rudder alone is seen in Fig. 11. The stator-propeller system is designed by a developed lifting surface code. The program has been developed for the analysis of asymmetric conditions by considering non-uniform oncoming flow. The program consists of a propeller part and a stator part that compute iteratively by considering the other s induced velocities as shown in Fig.9. The stator wake is assumed to be straight, and the velocity induced on the propeller surface by the stator wake is calculated by interpolation from the mid point of the vortex, where a singularity is avoided. In order to achieve rapid convergence, the steady axisymmetric program is first run two or three times, and finally the unsteady program is run with the converged previous data. The on-coming flow measured on the propeller plane by KRISO(Van et al. 1998) is used for the analysis of preswirl stator-propeller program. The model-scale velocity on the propeller plane is used as an input for the analysis, since the purpose of the study is to show the efficiency gain by the stator through model tests. If fullscale wake scaling is applied, the speed of on-coming flow is a little faster on each side since the flow is contracted. This phenomenon has been calculated by well-known methods such as Sasajima et al.(1966) or Hoekstra (1975). The angle of attack of on-coming flow to the stator blade section might become a little smaller at full-scale if the optimum pitch angle is kept the same as for the model scale. This effect is expected to be considered for a full scale preswirl stator. The effective velocity at the stator plane is also difficult to predict precisely since the interaction of the propeller and stator on the stator plane cannot be deduced easily from the self-propulsion test results. The effective velocity, even on the propeller plane without a stator, is difficult to estimate precisely, since the mean effective velocity is normally only predicted from model test results. As no powerful tool for quantitatively estimating effective wake distribution presently exists, a practical method for the optimum design is to investigate the correlation between model test results and computation by controlling the pitch of the stator. The optimum performance can be found by varying stator pitch angle, which is expected to be studied in the future. In this study, the propeller used in KVLCC is adopted to compare the performance with and without the preswirl stator-propeller system. The stator design has been conducted to have an optimum loading (elliptic loading distribution) for each blade as shown in Fig.14. The computed circulation of radial loading in each blade is converted by the equivalent angle of attack (α equiv ) as introduced in Eq.1. α eauiv. 1 CL = sin 2π 1 2 G (1) = sin u c ( )( ) Vs R where G is non-dimensional circulation as given by Γ G =, c local chord length at each radius, u local 2π RVs inflow velocity, C L lift coefficient and V s : ship velocity at design condition The hydrodynamic angle of attack of the stator is understood by this equivalent angle of attack to have a maximum value not larger than degrees. In previous research (Kim et al. 1994), a controllable pitch stator was used to verify the optimum pitch angle in uniform flow to be between 13 and 14 degrees. These values are the basis of the determination of maximum angle. As shown in the previous work (Kim et al. 1994), the rotational velocity of propeller cannot be completely cancelled by the stator since severe separation occurs on the stator blade if the stator pitch angle is increased more than 14 degrees, which was validated in open water tests in the KRISO cavitation tunnel. The optimum cancellation by a stator was about 50% of propeller rotational velocity components according to Kim et al. (1994). If a counter-rotating propeller is used, 100% cancellation could be achieved. Takekuma et al.(1980) also proposed the optimum angle as 15 degrees based on model tests. The pitch angle of the designed stator is variable along the radii to have an optimum loading. This may pose some difficulties in the manufacture of a fullscale stator. The difference of efficiency with constant and variable pitch angles along the radius was investigated by Kim et al. (1994), which showed that the variable case is about 2% more efficient than the constant case. Although the pitch angles vary along the radial direction, the blade loading shapes do not coincide perfectly with the line shown in Fig.14, considering the smoothness of the radial pitch angle distribution which is related to the manufacture of the stator. The optimum equivalent angle of attack of 14 degrees in uniform flow is also computed for comparison with the designed blades in non-uniform flow. The final propeller-stator 6

7 design profile is shown in Fig. 15. For the validation of the port stator performance, a non-optimally designed stator ( starboard stator ) is made simply by turning the optimal design stator by 180 degrees. The analysis results of equivalent angle of attack with the starboard stator are shown in Fig.16, where a large deviation of loading distribution is found to be far from an optimum elliptic profile. Table 2 Blade geometric angle distributions along radii for port and starboard stators (degrees): blade numbering is given in Fig.10 radius blade blade 1 port starboard blade 2 port starboard blade 3 port starboard blade 4 port starboard α : blade geometric angle a) Side view of stern part without stator b) Oblique view of stern part without stator Fig. 11 Model ship equipped with propeller alone and rudder blade4 blade3 blade1 blade2 port stator a) Stern view with port stator blade1 blade4 blade2 starboard stator blade3 Fig.10 Definition of port and starboard stators and its blade numbering (looking upstream) b) Stern view with port stator and propeller Fig.12 Model ship equipped with Port stator 7

8 ` a) Stern view with starboard stator (a) Looking upstream b) Stern view with starboard stator and propeller Fig.13 Model ship equipped with starboard stator (b) Looking downstream α equiv optimum blade1 blade2 blade3 blade4 Fig.15 Representation of the biased preswirl stator propulsion system by panels (port stator) -8 α equiv optimum blade1-2 blade2 blade3-4 blade r/r Fig.14 Calculated radial α equiv distribution on the stator blades (Port stator) r/r Fig.16 Calculated radial α equiv distribution on the stator blades (Starboard stator) 8

9 4. MODEL TESTS IN TOWING TANK In order to verify the effect of the stator on the selfpropulsion factors of the ship, a series of model tests, consisting of resistance tests with and without stator, open water tests for the propeller alone, were performed in the towing tank. As mentioned above, the series tests with the starboard stator was also conducted to compare performance. A rudder was fitted to estimate the full-scale value more precisely. In addition, the rudder located behind the propeller recovers the rotational velocity to some extent. The rudder is vertical, acting as a post swirl device which matches well with the present preswirl stator since there are no blades in the top and bottom positions. A comparison of resistance test results with and without stator is shown in Fig.17. The increase in resistance from the stator at a design speed of 15.5knots is about 2% larger than the resistance test result in KRISO (see Kim et al. 1994). This may be attributed due to the following: i) the stern profile of the model is slightly modified from the original KVLCC, as mentioned in Section 2; and ii) the scale ratios of the model is 50 while the KRISO model is 36. In addition, this difference in the measured resistance value may come from experimental uncertainty such as bias and random errors. The uncertainty values (root sum of biased and random errors) of the resistance experiment are estimated to be in the overall range of 1% of the measured value at the design speed. Two alternative methods for powering analysis from the model test with stator, basically following the 1978 ITTC method, can be considered. The first one is that a propeller together with the stator is considered as a propulsion system. For the evaluation of model test results by this method, the propeller open-water test results with the stator are necessary. In this study that method is not applied to the analysis of model test results because a Contra-Rotating dynamometer, which is necessary in POW tests, was not available. The second method is to consider the stator as part of ship hull, making the procedure the same as the 1978 ITTC method, except that the wake fraction of the ship is scaled up using the following equation. w = ( t ) s MO C + C + ( w t 0.04) + ( w w C FS A MO MO MS MO) FM where w : wake fraction of a model-ship w/o stator w MO MS : wake fraction of a model-ship w/ stator t : thrust deduction factor w/o stator MO w : wake fraction of a ship S (2) C : Frictional resistance coefficient for ship FS (1958 ITTC Line) C : Frictional resistance coefficient for model FM (1958 ITTC Line) C : Incremental resistance coefficient for model-ship A correction. 1 3 A s s C = [173.2 ( k / L ) 1.631] 10 k m L m 6 s : s : 320 Since the stator is located upstream of the propeller, the angle of attack of the on-coming flow to the propeller blade section usually increases due to the tangential velocity component generated by the stator. That is mainly considered as the potential part since tangential velocity component is rarely influenced by viscosity. The quantity( wms wm O ) is therefore considered to be the same for the model and the ship in this study, leading to further study for full ship verification by sea trials. This method is very similar to the Takekuma et al.(1980) method which is not based on the ITTC method but rather on the JTTC method. The retardation of axial flow that is mainly due to a viscous flow effect is also expected behind the stator, but this component cannot be obtained separately from the present experimental results, so it is ignored here. The predicted full-scale effective wake might be therefore a little larger than the actual sea trial result, which may make the speed of revolution less in a sea trial. There are presently no sea trial data available, but the assumptions above should be verified by sea trials to enhance the practical evaluation of the capability of the present method to correlate model test results with the actual performance of the ship, or if necessary, a new term should be added in equation (2). The analysis results of the model tests are shown in Table 3. The speed of revolution with the stator becomes slower than that of the single propeller system, as expected, and the amount of retardation for the port stator is almost the same as that for the starboard stator. The delivered horsepower at design speed (15.5 knots), with the port and starboard stators, and without the stator, are compared in Fig.18. As shown in the figure, a total of 5.6% savings in delivered power with the port stator is expected. The starboard stator does not work as an energy saving device as expected, showing that the delivered horsepower with the starboard stator is almost the same as that for the single propeller system. This result shows the effectiveness of the optimally designed stator with the given wake field. The efficiency gain of 5.6% is not so small, but it is regarded that there is still some margin to improve the biased stator propeller system if an optimum pitch angle is found using a controllable pitch stator, which will be further investigated. Although the efficiency gain of the stator is almost the same as for the symmetric stator with 6 blades 3 9

10 (see Kim et al. 1994), it can be regarded that the biased stator with 4 blades has some advantages compared with the symmetric stator such as compactness, lighter weight, smaller size, and most importantly lower cost. Although sea trial results are not available, the trend of full-scale value can be roughly estimated from the previous study (see Takekuma et al. 1981) where the efficiency gain in the full ship was 7-8% while the only 5% gain was obtained in the model test. In this study, the speed of revolution with the stator is slower than that without the stator, which may influence efficiency gain by 1-2%. For the application to the fullscale ship, this difference has to be considered assuming that the main engine is not generally changed due to the stator. In the case of CRP, the design concept and main engine selection are different from a conventional propeller system. One of the possible methods of recovering this loss is to reduce the propeller diameter instead of reducing the pitch since the large rotational energy due to large angle of attack can be recovered by the induced counter-rotational velocity of the stator. Research related to the variation of design parameters of the propeller is also expected to be done in the near future. Table 3 Comparison of test results at three different cases knots stator EHP(PS) DHP(PS) RPM DHP(%) without 14,219 21, port 21, ,316 starboard 22, without 16,163 24, port 23, ,279 starboard 24, without 18,175 26, port 25, ,369 starboard 26, without 20,256 29, port 27, ,520 starboard 29, without 22,404 32, port 30, ,700 starboard 31, without 24,621 35, port 33, ,902 starboard 34, without 26,906 38, port 36, ,165 starboard 38, EHP(PS) without stator with stator Ship Speed(Knots) Fig.17 Comparison of resistance test results between with stator and without stator DHP(PS) without stator port stator starboard stator 5.6% decrease Ship Speed(Knots) Fig.18 Comparison of delivered horse power without stator, with Port stator and with Starboard stator 5. Conclusions Instead of the usual symmetric design, a new biased preswirl stator, taking advantage of the non-symmetric wake flow characteristics between port and starboard sides at the propeller disk, is proposed. Procedures for the design, analysis and model test results of this new type preswirl stator system for a 300,000 DWT VLCC (KVLCC) are described. The proposed biased stator, with 3 blades on the port side and one on the starboard side, effectively recovers rotational energy loss by taking advantage of non-symmetric wake flow characteristics behind the ship. The following conclusions can be 10

11 drawn: The program based on lifting surface theory has been successfully applied to the design of the stator. The equivalent angle of attack concept is used for the optimum design of the stator blade angle. However, the present design has room for improvement since the oncoming velocities, being important input for the program, cannot be accurately predicted because of the highly complicated flow interactions of the propeller, stator and hull, including the rudder. The prediction of effective velocity on the propeller plane will be studied in the future. The resistance test and self-propulsion test with and without the designed stator have been conducted to find its effectiveness. The modified analysis method for the model test results of a preswirl stator-propeller system is used based on the 1978 ITTC standard method. The experimental results analyzed by this method show that a 5.6% efficiency gain for the model with the biased stator with 4 blades can be obtained compared to the model without the stator. If the optimum pitch angle is considered, more than 5.6% efficiency gain would be achieved. Investigation about optimum pitch angle in effective wake field by using a controllable pitch stator is considered for future work. This amount of gain is almost equivalent to that of the existing symmetric stator with 6 blades and therefore, it can be regarded that the biased stator with 4 blades has some merits compared with the symmetric stator such as compactness, lighter weight, smaller size, and most importantly lower cost of installation, estimated around 2/3 of the six blade stator since the cost of the propeller and stator is normally proportional to weight. However, as mentioned in the paper, some assumptions made in the analysis are recommended to be verified by sea trials to enhance the practical evaluation of the capability of the present method to correlate model test results with the actual performance of the ship. The adaptation of the new biased stator to a full scale ship is planned by cooperation with Daewoo Shipbuilding & Marine Engineering Co. Ltd, Korea and therefore, these studies can support the practical applicability of the biased stator to the actual ship. Acknowledgement This research is supported by Advanced Ship Engineering Research Center. The support is gratefully acknowledged. BSI D. Kim, M.-C., Lee, J.-T., Suh, J.-C. and Kim, H.-C. (1993), A Study on the Asymmetric Preswirl Stator System, Journal of Society of Naval Architect of Korea, Vol. 30, No. 1, pp Kim, K.-S., Kim, M.-C., Van, S.-H., Suh, J.-C. and Lee, J.-T. (1994), A Preswirl Stator-Propeller System as a Reliable Energy-Saving Device, Proceedings of Propeller/Shafting 94 Symposium, Virginia Beach, pp. 9-1~9-16. Lee, J.-T., Kim, M.-C., Van, S.-H., Kim, K.-S. and Kim, H.-C.(1994), Development of a Preswirl Stator System for 300K VLCC, Journal of Society of Naval Architect of Korea, Vol. 31, No. 1, pp Stern, F., Kim, H.-T., Zhang, D.-H., Toda, Y., Kerwin, J. and Jessup, S. (1994), Computation of Viscous Flow around Propeller-Body Configurations: Series 60 C b =0.6 Ship Model, Journal of Ship Research, Vol. 38, No. 2, pp Takekuma, K., Tsuda, S., Kawamura, A. and Kawaguchi, N. (1981), Development of Reaction Fin as a Device for Improvement of Propulsive Performance of High Block Coefficient Ships, Journal of Society of Naval Architect of Japan, No. 150, pp Takekuma, K. (1980), Evaluation of Various Types of Nozzle Propellers and Reaction Fin as the Device for the improvement of Propulsive Performance of High Block Coefficient Ship, SNAME Shipboard Energy Conservation Symposium, pp Van, S.-H., et al. (1998), Technology of Improvement on Resistance Performance of Ships, KRISO report UCM D. Sasajima, H. and Tanaka, I. (1966), On the estimation of Wake of Ships, 11 th ITTC Report, Appendix, pp Hoekstra, M. (1975), Prediction of Full Scale Wake Characteristics based on Model Wake Survey, International Ship Building Program, No References Blaurock, J. (1990), An Appraisal of Unconventional Aftbody Configuration and Propulsion Devices, Marine Technology, Vol. 27, No. 6, pp Lee, J.-T. et al. (1991), Cooperative Project on the Development of Compound Propulsor System with a Pre-Swirl Stator for High Efficiency, KRISO Report 11