A study of cabin inside noise of Hanjin highspeed

Transcription

1 A study of cabin inside noise of Hanjin highspeed passenger coaches Y.J. Lee HHIC Technical Research Institute, Hanjin Heavy Industries Ltd, 118, 2-Ka, Nadaemun-Ro, Chung Gu, Seoul, Korea Abstract Hanjin Heavy Industries Co., Ltd. has developed and constructed high speed passenger coaches running from Guangzhou to Hengyang in China with the running speeds up to 200 km/h. Ministry of Railway in China requested that the cabin inside noise of the passenger coaches be less than 65 db(a) at maximum speed. Various tests and calculations on the cabin inside noise are performed to design the structures of car bodies under the assumption that both the wheel/rail noise and the aerodynamic noise have the value of 100 db(a) and 95 db(a), respectively. 1. Introduction Hanjin Heavy Industries Co., Ltd. has developed and constructed five types high speed passenger coaches, first class car with compartments, second class car, second class car with broadcasting room, dining car and power generating car. These coaches will run from Guangzhou to Hengyang in China with the maximum speed of 200 km/h. Ministry of Railway in China requests that the inside noise level be less than 65 db(a) at the central area and at 1.5 ± 0.1 m above floors, except dining car and power generating car. The inside noise of dining car should be less than 68 db(a) at the center of the bar and the dining rooms. And the noise level of cabin attendant room and control room in power generating car should be less than 75 db(a) and 80 db(a), respectively. The general drawings of passenger coaches are shown in Figure 1 with measuring positions of microphone. Since there are no given design criterion or measured data on outside noise sources, it is assumed that the wheel/rail and the aerodynamic noise are

2 34 Computers in Railways of 100 db(a) and 95 db(a) at maximum running speed, respectively. These values on noise sources are about 15 db(a) higher than that of ICE in Germany, which has about 82 db(a) of wheel/rail noise and 77 db(a) of aerodynamic noise. Overall noise transmission loss of multi-layered structure is in general used to determine the best sound absorption and insulation materials and to estimate cabin inside noise from the outside noise sources [1]. Transmission loss values on each part of the high speed passenger coaches can be determined through the tests in anechoic and reverberation chamber. Computations on three dimensional models are carried out to estimate the inside noise level from the noise sources including the noise generated by air conditioning units which affect cabin inside noise. The results indicate that the cabin inside noise level at the test area of passenger coaches may be less than that of specified requirements. Note that the structure born noise from the interactions between wheel and rail is not considered in this paper, since the bogies and wheels are developed and produced by Changchun Railway Car Works in China. 2. Noise transmission loss tests The car body structures of passenger coaches are composed of four main parts, side, end, roof and underframe, and hence, four specimens are constructed to test noise transmission loss. Each panel has the same width and length of 1200 x 850 mm with different thickness. Figure 2 shows the drawing of the specimen of side panel. Noise transmission loss tests are performed by changing various insulation materials, fiber glass, polyethylene form, etc. Noise transmission loss is calculated from the average sound pressure levels in source room and receiving room by adjusting the value with correction factor determined by the volume of testing room and the area of partition [2]. The results are shown in Figure 3. The overall noise transmission loss can be defined by the weighted transmission loss, R^, in DIN PART 4. When fiber glass is selected as an insulation material, the resulted weighted transmission loss values are summarized in Table 1. Table 1. Weighted transmission losses of car body structures RW Underframe 41 db Side panel 38 db End panel 39 db Roof panel Side window 38 db 31 db The transmission loss TL of body side structures can be calculated from the ones of side panel and window through the equation [2] as follows

3 Computers in Railways 35 XL = 10 log (1) where ys. %. -^ T = ^ and TJ = 10 1 (2) Both T and S denote the transmission coefficient and area, respectively. The resulted noise transmission loss of body sides is 36 db. By using these transmission losses, cabin inside noise level can be easily estimated from the outside noise sources, wheel/rail and aerodynamic noise. 3. Estimate of cabin inside noise level Since there are no measured data on the outside noise sources or no given criterion from Ministry of Railway in China, it is assumed that wheel/rail noise level is 100 db(a) and aerodynamic noise 95 db(a). In order to calculate the cabin inside noise level, the information on both exterior noise levels of car bodies and noise reduction by sound absorption of interior equipment has to be known through experiments. But, at an earlier design stage of passenger coaches, it is difficult to estimate both quantities exactly, since there is no possible way to measure them. Therefore, it is assumed that the exterior noise level of underframe is the same as wheel/rail noise level, 100 db(a), and the other exterior noise levels are the same as aerodynamic noise level, 95 db(a). And it is also assumed that the sound reduction by absorption materials is 2 db(a), which may be generally accepted factor for a coach running in underground [1]. In the following section, these quantities are calculated by using conical beam method of RAYNOISE developed by NIT. By using these assumptions, cabin inside noise level can be calculated through the following equations : inside = 10 log (2 x 10^ + 2 x 10 "» + 10^ + 10^) - 2dB(A) (3) where L%, L%, L^ and L% denote the interior noise of side panel, end panel, underframe and roof, respectively. The interior noise of each panel is defined as the difference between exterior noise level and transmission loss. The calculation results show that the estimated level is 63.7 db(a) at the maximum running speed 200 km/h, which is much lower than the required one 65 db(a). This calculation process is summarized in Table 2. Note that the wheel/rail noise and aerodynamic noise are only considered as noise sources of the passenger coaches, and hence, the other noise sources are excluded in these calculations.

4 36 Computers in Railways Noise Source Exterior Noise (EN) Transmission Loss (TL) Interior Noise (IN) Cabin Inside Noise Table 2. Calculations of cabin inside noise level Wheel/Rail Noise 100 Underframe 100 Underframe 41 Underframe 59 Body Sides 95 Body Sides 36 Body Sides Design of car body structures Aerodynamic Noise 95 Body Ends 95 Body Ends 39 Body Ends 56 Roof 95 Roof 38 Roof 57 unit: Using Fiber Glass db(a) IN = EN - TL Calculation Results by eqn(3) Whole bodies are made of stainless steel, SUS 304, by using spot welding, which cannot keep airtightness of the body structures. To keep this, sealing materials are inserted between panels before processing spot welding and damping materials are coated on the inside of the structures. After construction of the coaches, the airtightness of the car body structure is confirmed through tests on the car by giving pressure, 4000 Pa, to cabin inside and by checking the time duration of the pressure drop from 4000 Pa to atmospheric pressure. To reduce noise generated from air-conditioning units, both parallel baffles and sound resonators are designed and placed in ducts. By placing two thick panels of polyethylene form in the flexible ducts, the noise level measured at the outlet of flexible ducts reduces from 84.6 db(a) to 79.3 db(a) and the noise measured at 1 m apart from the outlet reduces from 71.3 db(a) to 68.7 db(a). By placing parallel baffles and by attaching polyethylene form, the noise level reduces from 59.6 db(a) to 55.5 db(a) when it measured at the center of the cabin. In these tests, the body structure is not completely constructed, that is, it is composed of stainless steel panels, body frames, airconditioning units and duct panels. Later, the polyethylene forms attached to the duct wall are removed in complete body structures since it may disturb the air flow and reduce the capability of air-conditioning. One of main noise sources of power generating car is two power generating engines in engine room as shown in Figure 1. To insulate this noise, thick wall having 200 mm is placed between engine room and control room. Thick double window having R^= 46 db(a) is also placed between two rooms. Two locks are attached to doors in order to prevent noise through the gap between door and door frame. Figure 4 shows the frequency spectrum of cabin inside noise measured at

5 Computers in Railways 37 the center of second class car at standing in open fields, after the structure of the car is constructed completely. The resulted overall noise levels measured at specified positions as shown in Fig. 1 are summarized in Table 3. Note that the measuring instruments used in tests are B& K 3550 and Larson & Davis Table 3. Resulted cabin inside noise level at standing First class car Second clatss car with broac casting room Dining car Dining room Power generating car Type of car Bar Control room Crew's room Noise 1< 2vel 53.7/54, 0dB(A) db(a) 56. 7dB(A) 60. 9dB(A) 68. 6dB(A) 62. 6dB(A) In final running tests in China, noise measurements, SONY 8 ch. DAT recorder and B&K 3550 will be utilized according to test procedure [3]. 5. Computational analysis of high speed passenger coaches Acoustic and structural modes of second class car are computed with SYSNOISE developed by NIT [4]. And vibro-acoustic analysis is also performed on two dimensional section of the cabin interior, in which the wheel/rail noise is represented by a cylindrical source and aerodynamic noise by a plane wave source. Figure 5 shows the 3-D model composed of 5616 linear brick elements for the analysis of acoustic modes. The first acoustic mode occurs at 6.65 Hz which corresponds to a half wavelength along the cabin length A, 2 x 25.5 m (3) For a long tube-like domain, the number of mode is large and increases quickly with increasing frequency, and hence, the 3-D acoustic analysis using modal superposition by finite element methods would require enormous number elements since the analysis requires a lot of acoustic modes. The 3-D models of both second class car and dining car are constructed for the analysis of cabin inside noise level by using the conical beam methods of RAYNOISE [5]. Note that the effects of the sound resonators and parallel baffles are not considered in these models. The 3-D model of second class car

6 38 Computers in Railways is shown in Figure 6. The geometry is modeled by using CATIA and the grids are generated by using PATRAN III. Wheel/rail noise sources are represented by point sources and aerodynamic noise sources by linear sources positioned at 100 m apart from the passenger coach. The frequency spectrums of exterior noise of the dining car can be computed from these noise sources as shown in Figure 7. The noise source by air-conditioning units is represented by a point source having 90 db(a). The computation results indicate that the cabin inside noise is largely determined by the noise generated by air conditioning units, since this source level is very high compared to the other noise sources. The resulted exterior noise of the underframe is from 93 db(a) to 96 db(a) and the other exterior noise levels are from 92 db(a) to 96 db(a). Figure 8 and Figure 9 show the computed cabin inside noise by the interior noise of panels and air-conditioning unit. The resulted noise level of second class car is 66.5 db(a). And the cabin inside noise at center of dining room is 67.0 db(a). If the sound resonators designed through tests are included in the models, the results are much lower than these values. 6. Conclusions Various tests and calculations on the cabin inside noise are carried out to design the structures of car bodies under the assumption that both the wheel/rail noise and the aerodynamic noise have the value of 100 db(a) and 95 db(a), respectively. The calculation and computation results may prove that the cabin inside noise levels are lower than the required ones by Ministry of Railway in China. The static test results show that the noise generated by electrical equipment may not have a contribution to the cabin inside noise when the high speed passenger coaches runs with its maximum speed 200 km/h. Since the final running tests on the high speed passenger coaches are postponed from February in 1996 to April in 1996, the results of running tests are not included in this paper.

7 References Computers in Railways Breda Construzioni Ferroviarie, Noise level estimate, Technical Report, SC^rD-OD&L 306, Pistoia, Italy, Kinsler, L. E., Frey, A. R., Coppens, A. B. & Sanders, J. V. Fundamentals ofacoustics, John Wiley & Sons, New York, HHIC Technical Research Institute, Type test procedure for measuring the interior noise level of passenger cars for Hengyang - Guangzhou railway project, Test Procedure, RETEST-003, summitted to MOR (Ministry Of Railway in China), Hanjin Heavy Industries Co., Ltd., Seoul, Korea, Segaert, P. & Migeot, J. L. Simulation of high speed train cabin noise, Technical Report, Sysnoise Project and Support Group, Numerical Integration Technologies, Leuven, Belgium, RAYNOISE User's manual. Rev. 2.1, Numerical Integration Technologies, Leuven, Belgium, 1995.

8 40 Computers in Railways (a) First class car (h) Second class car with broadcasting rooms (c) Dining car "-1 II on I'igurc 1: Drawings of high speed passenger coaches with the arrangements of microphone.

9 Computers in Railways 41 itl SECTION A-A \ } 1200 SECTION B-B! L? FRAME FRA1* frmf FRAME FRAME SIDE POST SIDE POST SIDE POST "W* DESCRIPTION I* ERI o rr PZ PZ PZ t RE 40 JO to JO 40 fto JO 40 JO 40 JO to JO wax Figure 2: Drawing of specimen of side panel for the test of transmission loss. CQ Td & JV n._... ^ ^AI _...._.. / r - 7 ' -- ^- -^_r"^ -~ -- ' ' /^ - ' '^T^-i±± ^ ^~- Fiber Glass 50t Polyester Form 40t Polyethylene Form 80 t LD50ALSI 50t Frequency (liz) Figure 3: Test results of transmission loss of side panel.

10 42 Computers in Railways k 2.50k 6.30k 16.0k 1'reciucnc)- (Hz) 4: Cal'in iii.siilc noise level otsecond ckiss ear \\'iili hroadeasiing mom ai staitding Figure 5: Acoustic finite element meshes of cabin interior (envelope of half-model).

11 Computers in Railways 43 Figure 6: Three dimensional model of the second class car with broadcasting room for the analysis of the cabin inside noise. (a) Top view (b) Bottom view (c) Front view (d) Side view Figure 7: Acoustic finite clement meshes of cabin interior (envelope of half-model).

12 44 Computers in Railways SPL dba wide band r k. Y Figure 8: Cabin inside noise of the second class car with broadcasting room by using conical beam method of RA YNOISE. SPL dba wide band Figure 9: Cabin inside noise of the dining car by using conical beam method of RAYNOISE.