Modification of the trailing edges of the large stay vanes and their influence on dynamic stresses

Transcription

1 IOP Conference Series: Earth and Environmental Science Modification of the trailing edges of the large stay vanes and their influence on dynamic stresses To cite this article: A Gaji et al 21 IOP Conf. Ser.: Earth Environ. Sci View the article online for updates and enhancements. Related content - Static and dynamic stress analyses of the prototype high head Francis runner based on site measurement X Huang, C Oram and M Sick - Fatigue analyses of the prototype Francis runners based on site measurements and simulations X Huang, J Chamberland-Lauzon, C Oram et al. - Design of large Francis turbine using optimal methods E Flores, L Bornard, L Tomas et al. This content was downloaded from IP address on 21/9/218 at 23:4

2 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 Modification of the trailing edges of the large stay vanes and their influence on dynamic stresses 1. Introduction 1.1 Basic data A Gajić 1, P Manzalović 2 and Z Predić 3 1 Faculty of Mechanical Engineering, University of Belgrade Kraljice Marije 16, Belgrade 1112, Serbia 2 HPP Iron Gate, Kladovo, Kralja Petra 2, Serbia 3 A.P. Company doo, Tausanova 2, Belgrade 11, Serbia agajic@mas.bg.ac.rs Abstract. In the purpose of discovering causes of high vibrations and long cracks on stay vanes of the turbines at HPP Iron Gate I measuring of stresses in the upper zone and in the lower zone along the stay ring are completed. Investigations performed some 2 years ago, in order to find out the excitation of vibrations, included measurements of the pressure along the trailing edge and stresses on stay vanes. The main frequency was about 36 Hz, which is very close to the natural frequencies of some stay vanes. It was concluded that there were two excitations, one belonging to Karman vortexes and the other to attack flow of the guide vanes, both with very close frequencies, causing beating oscillations. Modification of the stay vanes was made by grinding of the convex part of the trailing edges until half-circular concave of the same radius of 2 mm was reached. On site measurements performed at one unit, with original and modified trailing edge, with the same measuring devices, at the same measuring points and at the nearly same operating regimes. Influence of the modification on the static and dynamic stresses of the stay vanes, as well as their frequencies of the oscillations, is shown in the paper. Turbines HPP ''Iron Gate I'' are of the Kaplan type, diameter 9. m. As shown in Fig.1., the water flows from the spiral case to the runner between 12 stay vanes of different dimensions and shapes, then between 32 guide vanes and it turns around turbine runner made of 6 blades with the speed of rotation of min -1. During 197 and 1971, when the machines were set off, they were working with the designed characteristics: - installed flow rate 72 m 3 /s - nominal turbine power output 178 MW - max. net head 34.6 m - min. net head 17. m - frequency of rotation f o =1.19 Hz - frequency of runner blades f rk =6*1.19=7.1 - frequency of the turbine stay vanes f st =12*1.19=14.3 Hz - frequency of the turbine guide vanes f sa =32*1.19=38.13 Hz Since the set off, the machines went through the different phases of exploitation, with the constant growth of the flow and the power output. The growth was made by enlargement of the head, but also by enlargement of the installed flow on the basis of the research from It was not until 1997, when the power was increased even to 199 MW, than the units had been working with the maximum power of 194 MW. c 21 Ltd 1

3 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 Fig. 1 Disposition of the spiral case and the stay vanes 1.2 Cracks on the large stay vanes of the turbine and the research After the increase of the power and the flow through the turbines, the phenomenon of the cracks on the stay vanes was detected. The cracks were first noticed in the After the repairs which followed, the cracks were recorded on all 6 turbines, but not all vanes were equally attacked, which is the consequence of the difference in the dimensions, shapes and positions in the turbine. The height of the vanes is 38 mm. The vane No. 1, which is simultaneously the tongue of the spiral, is of the triangular cross section. The vanes No. 2 and No.3 are of the same curved profile, length 11 mm. The vane No. 4 has the symmetrical enhanced profile of the length of 139 mm, and all the other vanes (from to 12) have the same asymmetrical profile of the length of 139 mm. The trailing edge of all the vanes had the same half - circular shape, with the radius of 2 mm. All the damaged places, discovered during the repairs, were cut and welded again. Simultaneously, the analyses were made, with the goal of finding the solutions. The first analyses were made at the end of 1986 [1, ] on the vanes No. 3, No. 4, No. and No. 9, as the representatives of the different types, and No. 9 was taken because of the position. The vane No. 1 didn't have the cracks due to the special construction. Then the pulsations of the pressures on the trailing edges and the natural frequencies of the vanes in the air were measured. Although not completely reliable, since there was no data on response of the vanes, that is, on the dynamic stresses, it was concluded that the resonance in the field of the frequency from 36 to 38 Hz was the possible cause of the cracks. The two dominant frequencies of the dynamic pressures were - the first around 7 Hz, and the second (especially expressed with the power larger than 16 MW) was between 36 and 38 Hz. The frequency between 36 and 38 Hz was especially interesting, because the same was very close to the frequency of the conductive machine, close to the estimated natural frequency of the vanes in the water, and probably the consequence of the shape and the frequency of cutting off the vortex on the trailing edge of the vane. Because the cracks always appeared on or very close to the stay rings, and the doubts that it may be the same with the residual stresses, the analyses of the residual stresses were organized in the The same, however, didn't imply that the residual stresses could have been the cause of the craking of the vanes. The further analyses were continued with the purpose of researching the resonance phenomena on the vanes of the turbine stator. The new program of the analyses made in 1989 [2, ], included the research of the natural frequencies of the vanes in the water and in the air. Excitation loads were performed by special pneumatic impact hammer. Intensity of its impulse exciting force was controlled by pressure, while the duration of the impulse was a function of the hardness of the removable impact head [9]. The static and dynamic condition of the stay vanes of the turbine and the pulsation of the pressures on the contour of the vanes. Fig. 2 shows the pressure 2

4 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 oscillations related to the discharge. The measured natural frequencies of the vanes in the air (NF A ) and in the water (NF W ) and Strouhal number of Karman vortexes on stay vane No. is shown in Fig. 3. Fig. 2 Pressure oscillations related to discharge Fig. 3 Measured natural frequencies and Sh-number Fig. 4 Distribution of the amplitudes of main harmonics 21 Hz and 36 Hz along the stay vane No. Spectral analysis (PSD) of the measured data shows that the main harmonics are 36 Hz, 21 Hz, 7 Hz etc. correspond to frequency of rotation, of runner blades, of the turbine stay vanes, of the turbine guide vanes, but also of Karman vortexes at the trailing edge [6]. Distribution of the pressure amplitudes harmonics 21 Hz and 36 Hz are presented in Figure 4. The results clearly showed that there were conditions for the phenomenon of the resonance in the field of the larger power with the frequency between 36 and 38 Hz. However, but due to the economical situation, nothing was done for the solving of this problem. The cracks were repaired from the one to the other repair, when they appeared. 3

5 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 With the new growth of the power output after the 1997, the action of finding the solutions for the stay vanes was restarted, now also in the light of the next revitalization and the further growth of the power. The new program of the analysis, which came and was set off on the Unit 1 in 1999, referred only to the static and dynamic stress of the vanes, at the increased maximum power output of 2 MW. With this program the unit was transformed. The previous analyses were made on the Unit No. 3, and then the Unit No.1 was chosen. Also, the measurements were made on the vanes No. 2 and No.12. The results have completely confirmed the previous conclusions and secured the reference values for the evaluation of the solution which will be applied. Turbine manufacturer decided finally to reconstruct the shape of the stay vanes, by changing the trailing edges, according to their own experiences applied several times from 1963 until nowadays [9]. Fig.. Modification of the vanes the trailing edge from the convex transformed into the concave shape The adopted solution, shown in Fig.., reminds on the tail of the swallow, and was made by grinding the half-circle of the radius of 2 mm on the trailing edge and in the vane the half-circle with the radius of 2 mm was concaved. 2. Effects of the modification of the shapes of the trailing edges 2.1 Analyses of the modified stay vanes The analyses, after the modification of the trailing edges of the stay vanes, were made on the turbine 3 in the December 23. Since it was about the check of the effects of the modification, the results of the analyses of the vanes No. 2 and No. 12 on the turbine 1 were taken as the referents, and the mere analyses were made on the vanes No. 2 and No. 12, as on the unit.1, and with the same methodology, with the same position of the points for measurements, with the three strain gauges on each side, in the lower and the upper zone of the vane on 8 mm from the stay ring. Considering the relatively short period between the measurements before and after the modification of the trailing edges, the same measuring system with the same type of the strain gauges was used for the measurements. The processing and the presentation of the data were made in the same way as before the modification. In this way, the contrastive analysis was brought to the contrasting the same diagrams and the tables before and after the modification. The analyses were then made with nearly the same regular regimes to 2 MW, but also above this level, even to the power of 28 MW. The flow was about 1 m 3 /s, which compared to the designed flow represents, the growth of around 38 %. For the needs of the measuring the static component, the complete load rejection was made by lifting of the rotor, like in Comparing the results It is obvious that the accomplished modification didn't significantly affect the natural frequencies of the vanes, because neither the speed of the rotation nor the number of the vanes had changed. The only possible modification was the frequency of the Karman vortexes. It would be important for the science to answer how much was the change in the pulsations and in which direction, but unfortunately, the dynamic pressures in the turbine with the modified vanes were not measured. According to the results of the measurements of the dynamic stresses, there are indications that the frequency of the Karman vortexes was changed on the values over 4 Hz, with the relatively small amplitudes, as shown in Fig.6. It is possible that the power of these vortexes wasn't larger, or much larger, before the modification, but now is changed beyond the boundaries of the natural frequencies of the stay vanes, and there is no resonance. 4

6 th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 TURBINE 1, STAY VANE 12 ϕ= +1 ο 3 2 MT FLOW DIRECTION 2. MT BEFORE TURBINE 3, STAY VANE 12 MODIFIED, ϕ=+1 ο 3 2 MT FLOW DIRECTION 1. MT Fig. 6 Dynamic stresses for the three measuring points before and after the modification AFTER The contrastive analysis for only three measuring points M13, M14 and M1 presented in Fig. 6. is given for the illustration of the character and the size of the dynamic component of the stress. The summarized comparison is given in Figs. 7 and 8. Fig. 7 shows that the dynamic stress state after the applied modification of the trailing edges of the vanes, measured with the standard deviation of the fluctuations, is 3. times more convenient. Measured with the double amplitude of the stress fluctuation (see Fig.6), the improvement is even larger. What is evident here is

7 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 that the increase in the intensity of the dynamic stresses was connected with the power range between 16 and 19 MW. With the larger outputs, the frequencies of the Karman vortexes went out of the range of 36 to 38 Hz, because the standard deviations of the fluctuations fell with the larger power. ST. DEVIJATION (MPa) MT1 MT2 MT3 MT4 MT MT6 MT7 MT8 MT9 MT1 MT11 MT12 MT1 MT16 MT17 MT18 MT19 MT2 MT21 MT22 MT23 MT POWER OUTPUT (MW) BEFORE 1.6 ST. DEVIJATION (MPa) MT1 MT2 MT3 MT4 MT MT6 MT7 MT8 MT9 MT1 MT11 MT12 MT1 MT16 MT17 MT18 MT19 MT2 MT POWER OUTPUT (MW) AFTER Fig. 7 Fluctuations of the dynamic stresses of the vane No. 12 before and after the modification It is typical for the modification of the vane that the parallels between the frequencies are avoided and that the dynamic component has no sudden increase, but gradually raises with the flow. Dynamic component now has its maximum on the maximum power. Beside that according to the value this maximum is much lower than the previous maximum; it appears with the regime that is little used in the exploitation. Regarding the static (middle) component of the stress Fig. 8 shows that there were no changes, because if we compare the data for 2 MW, we will see that they are the same. This, obviously, wasn't expected, but it confirmed to the analysts and to the customer that the measurements were correct. 6

8 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 MID STRESSES ON STAY VANE 2 BEFORE MODIFICATION MT1 MT2 MT3 MT4 MT MT6 - Poly. (MT3) Poly. (MT6) POWER OUTPUT (MW) MID STRESSES ON STAY VANE 2 AFTER MODIFICATION MT8 MT MT1 2 - MT11 MT12 Poly. (MT6) -1 Poly. (MT3) -1 MT1 MT2 MT3 MT4 MT MT6 MT POWER OUTPUT P (MW) 3. Conclusion Fig. 8 The middle stresses of the stay vane No.2 before and after the modification The modification made on the trailing edges of the stay vanes of the turbine brought the change in the conditions of the flow on the trailing edges and the decrease of the intensity of the dynamic component of the stresses. The previous peak at the power output between 16 and 19 MW didn't show up, which at the power to 2 MW and at the flow to 84 m 3 /s secured the smooth working of the units and the larger resistance of the stay vanes on the material fatigue. Nomenclature A l P Cross-section area [m 2 ] Length of the profile [mm] Power output [ MW] h f σ head [m] Frequency [Hz] Stress [Pa] 7

9 2th IAHR Symposium on Hydraulic Machinery and Systems IOP Conf. Series: Earth and Environmental Science 12 (21) 12 doi:1.188/17-131/12/1/12 References [1] Muškatirović J and Predić Z 1987 Hydr. measurements on stay vanes at HPP Iron Gate I Institut Jaroslav Černi (Beograd, Serbia) (in serbian) [2] Predić Z 1991 On site measurements of the stay vanes at turbine 3 at HPP Iron Gate I Institut Jaroslav Černi (Beograd, Serbia) (in serbian) [3] Predić Z at al 22 Report of dynamic stresses measurements on stay vanes in HPP Iron Gate I A P Company doo (Beograd, Serbia) (in serbian) [4] Predić Z et al 24 Report of the complex measurements at unit 3 of HPP Iron Gate I vol Stress analysis of the stay vanes A P Company doo ( Beograd, Serbia) (in serbian) [] Muškatirović J and Predić Z 1989 Analysis of Hydrodinamic Pressures Acting on Stay Vanes of Kaplan Turbines Proc. Int. Symp. on Large Hydr. Mach. and Assciated Equipments (Beijing, China) [6] Gajic A, Predic Z, Muškatirović J, Pejović S and Manzalović S 199 Investigation of Cracks on the Kaplan Turbines Stay Vanes Proc. 1 th IAHR Symp. (Belgrade, Serbia) vol 1 p A4 pp A4.1-A4.12 [7] Pulpitel L and Robert P 199 Discussion on 6 Proc. 1 th IAHR Symp. (Belgrade, Serbia) vol 3 pp [8] Gajic A, Oba R, Ikohagi T, Ignjatovic B and Muskatirovic J 1994 Flow Induced Vibrations and Craks on Stay Vanes of a large Hydraulic Turbine Proc. 17 th IAHR Symp., Int. Research Centre of Hydr. Mach. (Beijing, China) vol 3 pp [9] Albijanic R, Ignjatovic B, Aleksic M and Boskovic V 1994 Identification of Vital Hydrounit Component- Dynamic Parameters in Water Enviroment Proc. 17 th IAHR Symp., Int. Research Centre of Hydr. Mach.(Beijing, China) vol 3 pp [1] Gummer J H and Hensman P C 1992 A review of stayvane crecking in hydraulic turbines Water Power and Dam Construction [11] Aronson A J, Zabelkin V M and Pilev I M 1983 The investigations of the Cracks on Stay Vanes on Mixedflow Turbines (in Russian) [12] Ulith P 198 Flow Induced Cracks on Stay Vanes IAHR/IUTAP Symp. Practical Experiences with Flow- Induced Vibrations (Springer-Verlag) [13] Kirchner H 198 Dynamic Stresses on Stay Vanes of Vertical-Shaft Large-Size Francis Turbines IAHR/IUTAP Symp. Practical Experiences with Flow-Induced Vibrations (Springer-Verlag) [14] Grein H and Staehle M 1978 Fatigue Cracking in Stay vanes of Large Francis Turbines Escher Wyss News 1 [1] Grein H 198 Vibration Phenomena in Francis Turbines- Their Causes and Preventation Proc 1 th IAHR Symp. (Tokyo, Japan) vol 1 pp [16] Casacci S, Lourdeaux B and Wegner M 1982 Comportment Dynamique des Avant-Distributeurs de Grandes Turbines Francis Proc. 11 th IAHR Symp. (Amsterdam, Netherland) pp [17] Abernathy F H 1962 Flow over an Inclined Plate Trans. of ASME, J. of Basic Eng [18] Francis T B and Katz J 1988 Observation on the Development of a Tip Vortex on a Rectangular Hydrofoil, Trans. of ASME, J. of Fluids Eng