INFLUENCE OF INPUT PARAMETERS ON PREDICTION OF NOISE IN FACTORY WITH BARRIERS AND FITTINGS BY COMPUTER MODELLING

HAL Id: jpa-00230691

https://hal.archives-ouvertes.fr/jpa-00230691

Submitted on 1 Jan 1990

HAL

is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire

destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

INFLUENCE OF INPUT PARAMETERS ON PREDICTION OF NOISE IN FACTORY WITH

BARRIERS AND FITTINGS BY COMPUTER MODELLING

B. Rudno-Rudzinska, A. Jaroch, J. Jurkiewicz

To cite this version:

B. Rudno-Rudzinska, A. Jaroch, J. Jurkiewicz. INFLUENCE OF INPUT PARAMETERS ON PREDICTION OF NOISE IN FACTORY WITH BARRIERS AND FITTINGS BY COM- PUTER MODELLING. Journal de Physique Colloques, 1990, 51 (C2), pp.C2-289-C2-292.

�10.1051/jphyscol:1990270�. �jpa-00230691�

Colloque C2, suppl6ment au n02, Tome 51, Fgvrier 1990 ler Congres Franqais dlAcoustique 1990

INFLUENCE OF INPUT PARAMETERS ON PREDICTION OF NOISE IN FACTORY WITH BARRIERS AND FITTINGS BY COMPUTER MODELLING

B. RUDNO-RUDZINSKA, A. JAROCH and J. JURKIEWICZ*

Institute of Telecommunication and Acoustics, WrocZaw Technical yniversity, u l . ~ . Prusa 53/55, 50-317 WrocZaw, Poland

University of Mining and Metallurgy, A1-Mickiewicza 30, 30-059 Krakow, Pol and

Abstract - The validity of the computer-aided predictive model for the assessment of sound field in industrial halls with barriers,was exami- ned. The range of possible error resulting from input parameters ucer- tainty was analyzed.

1 - INTRODUCTION

Computer-aided predictive model for the assessment of sound field in fitted factory halls with barriers,which is based on new algorithms of mirror images method incorporating diffraction and interferences phenomena, was developed The geometrical/statistical approach was adopted to solve the problem of numerical analysis of sound field. In computation of multiple diffraction the Pierce's solution which is based on geometrical theory of diffraction was adopted [1.21.

For description of steady-state scattered sound field a solution derived in 131 is applied. This solution is based on Kuttruff's formula [4] for probability of the event that sound ray appears in time t in a space element distant of r from the sound source. In numerical procedures the concept of point to point acoustic transfer function T t i , was applied L2.51. Tti was defined as a ratio between acoustic preasure at the end and at the begining of sound ray path.

The room transfer function T is expected to be calculated with limitted accuracy.The influence of input parameters on prediction transfer function T was studied.

2

-

The difference between predicted and measured data can be a measure of model's error (6). The model's error consists of two components: bias error 6s (error of the method) and parameter's error 62. The bias error 61results from limi- ted value of images order N set in computation of transfer function T. When investigating the hall with the barriers for a given value N energy portion neglected in computation is relatively great compared with this energy which is under consideration. So, the increase of bias error 61 is expected.

Sound field in the hall is influenced by such parameters as: geometry of room, barriers and sources, absorption coefficients of walls, floor, ceiling, ba- rriers and fittings, fittings density and sound power levels of sources as wel1.Usually at the stage of planning the values of input parameters can not be accurately estimated. The influence of two main parameters eg. absorption coefficient a and fitting density o on the uncertainty of predicted data, were examined.

An exact description of scattered sound field is not possible. For scattering objects (fittings) which are considerably larger then the sound wavelength the Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1990270

CZ-290 COLLOQUE DE PHYSIQUE

statistical concept is widely used. For such approach scattering properties of fittings are characterized by absorption coefficient as of machine's surfaces and average fitting density u = S/4V, where : S- total surface of scattering objects, V-volume of the room.

In practice, it is difficult to estimate the correct value of o because of irregular distribution of machines and equipment (only layers near the floor and ceiling are occupied) and their complicated shapes, moreover accurate estimation of US is not pesdble

3

-

The predict ions and measurements were carried out for the f 01 lowing rooms :

mechanical workshop hall of dimensions (78.3~14.6~8.3) m with omnidirectional, reference sound source of well known sound power level; large, shalow hall of dimensions (110x78x10)m with industrial noise source (table vibrator) of sound power level determined with technical accuracy; small, reverberant room of dimensions (9x6x3.5)m with reference sound source.

As the measuring signals 1/3 octave band of pink noise with centre frequency from the range 500-8000 Hz, were used. The experiments were made for halls without and with barriers.

Two types of barriers were used:light barrier (B2) of transmissionloss index 10-15 dB and heavy barriers (B1 and B3) of transmissionloss index 26-35 dB in the considered frequency range.

At first step of numerical experiments the bias error 61was tested. The large enough value N for halls without and with barriers was determined. Fig.1 shows the differences ( 6 ) between measured and predicted data for typical values of a and fittings density calculated acc.u=S/4V, as a function of images order N.

When investigating the influence of acoustical input parametres a and a the calculation were made: for N=7 and N=13 for halls without and with barriers respectively. Fig.2. illustrates the selected results of measured and predicted values A=lOlg(T) versus distance d from sound source. The "best fit"

values of parameters a and u providing the smallest error 62 were evaluated.

The "best fit" values of u for the mechanical workshop with (uif and without

(02) barrier for chosen frequencies were depicted in the tab.1.

Theoretically estimated value of u is equal to 0.05.

Tab. 1.

The compatibility of measured and predicted efficiency (E) of one barrier and set of two barriers in industrial halls were examined. The calculation of the barrier efficiency was made for different combinations of previously evaluated "best fit" a and u=0. Selected results are shown in Fig.3.

4

-

When comparing predictions and measurements the agreement always depends on the correctness of input parameters not only on accuracy of the predictive model.This fact cadt be ignored in analysis of the predictive model validity.

When analysing the effect of scattering one should take into consideration its space and frequency dependance. In the frequency range below 500 Hz this effect is negligible. For higher frequency the fittings density values providing best fit with measurements is ussualy less then theoretical.

The presence of acoustical barriers decreases the effect of scattering behind the barrier and as the first approximation a=O can be set.

REFERENCES

/1/ Pierce,A.D., J.Acoust.Soc.Am.. 55 (1974) 941.

/2/ Jar0ch.A.. Jurkiewicz,J., Rudno-Rudzit5ska.B.. Proc.Inter-Noise 88, 1499.

/3/ Janecek,P.,Acustica, 2 , (1986) 285.

/4/ Kuttruff,H., Acustica

18

/5/ Jar0ch.a.. Jurkiewicz,J., Rudno-Rudzit5ska.B.. Proc. 6th Seminar and Exhibition on Noise Control, Pecs-Hungary (1989) 276.

distances d source-observerat the same acoustical parameters. Mechanical workshop: a) without, b) with barriers; 1-d=3m, 2-d=l5m, 3-d=31m, 4-d=61m.

-404 I -404

1 10 Iml 103 I l o lml 103

4

Fig.2.Acoustical transfer function (A=lOlgT) versus distance from the source:

a) mechanical workshop with laboratory source, b) fitted hall with industrial noise source: measured ( 0 ) and predicted ( 6 ) for : (-)-cr=O,

(--- )-u=O.O3, (--- Jw=0.05; (l)-a=0.01. (2)-a=0.3.

C2-292 COLLOQUE DE PHYSIQUE

E [dB1

p *\ 0 2-

\g

'* ^\

.

. . . : - - - . ^i\

2 3 5 10 I m l 2 3 5 l0 Iml 2 3 5

lo

Fig.3. The barriers efficiency (E) as a function of distance from the source:

measured ( 0 ) and predicted (e) for differnt values of or and or, in mechanical workshop with laboratory sources for commonly used a in industrial halls;

(us-with barriers , or-without barriers);

(-)-us-O,a=O ; (---- )-crs=O,a as in tab.1 ; (--- 3 - o r . e as in tab.1