Sound emission and noise exposure – an approach to improve the link Wolfgang Probst 1 DataKustik GmbH Dornierstraße 4, 82205 Gilching,Germany

ABSTRACT A lot of work was spent by acousticians the last decades to reduce the noise caused by technical products. Important steps on that way were the description of the sound emission of these products and facilities by unambiguously defined quantities – the A-weighted sound power level and the emission sound pressure level. With a framework of standards for measurement and declaration we supported the buyer of such products in his purchase decision. The emission values of different ma- chine groups were published – in standards and guidelines in anonymous form – and the buyers were encouraged to select from possible alternatives those that offer the lower noise emission in their declaration. But since many years we hear in this buy quiet sessions that the declared emis- sion values are often not very reliable – and they are not understandable for those affected by the noise. This can be improved by establishing a strong link between the emission and the noise im- pact. Some strategies and tools are shown and recommendations for the development of supporting standards are given. Some examples about the advantageous use of emission values are demon- strated. 1. INTRODUCTION

As is well known, it is not easy to describe the noise emission of technical products from small kitchen appliances to complex processing machines in industrial plants in such a way that the ex- pected noise impact at workplaces under any ambient conditions can be clearly derived. The general approach is to apply two emission values: The emission sound pressure level L pA as the information about the A-weighted level at the work- place if room influence or background noise is not relevant. This value describes the lowest possi- ble level at the workplace or operators position because radiation under free field conditions (half space above reflecting floor) is assumed. Further the A-weighted sound power level L WA that describes the total output of sound energy. It must additionally be known to derive the contribution of the source at any position via free field sound propagation and via reflections at room boundaries. It shall only be mentioned that a lot of further information like the spectral directivity would be needed to have an accurate knowledge about the radiation of a noise relevant device, but in many cases the mentioned two emission quantities are sufficient as information to include noise aspects in planning activities and even to predict noise levels at workplaces. This is supported by the legally fixed noise declaration according to the Machine directive. Other details like spectra and directivi- ties are helpful if they are available, but generally it is also possible to link them via machine spe- cific databases or to apply most probable standard settings.

1 Wolfgang.probst@datakustik.de

worm 2022

The link between these emission quantities of all noise-relevant devices in a working environment with the noise impact at all positions where people are staying has been improved the last years with the application of calculation methods with ray or particle tracing techniques based on ener- getic superposition of all relevant sound impacts to predict the total impact at each workplace as A- weighted sound pressure level. All these methods belong to the generic term “geometric acoustics”. As there are some software products available to support this procedure, some activities are planned to organize a sort of quality assurance and to develop “application groups (GA)” with minimum re- quirements that can be checked with some test cases. 2. THE INFLUENCE OF THE TWO EMISSION VALUES ON THE NOISE IMPACT AT WORKPLACES

The simplest way to predict the sound impact caused by a source like a machine at one or more workplaces is discussed with the scenario shown in Figure 1.

worm 2022

Figure 1: Machine M acoustically simulated by a central point source Sc with attached workplace

P 1 and a workplace P 2 in larger distance (left plan view, right 3D-view) The machine – or other sound emitting devices of any size and complexity - is acoustically repre- sented by a central point source Sc described by the single number values L WA and L pA . P 1 and P 2 are two workplaces – P 1 the operators position for machine M. With this example it shall be assumed that the machine supplier declares an emission sound pres- sure level L pA of 82 dB(A) relevant for the workplace. But it is often overseen that the level at the workplace depends on the L pA and on the sound power level L WA . If this machine radiates with an L WA of 85 dB and is installed in a working environment with no extra absorption, we have finally 83 dB(A) at the workplace P 1 and 76 dB(A) at the workplace P 2 , but if the L WA is 95 dB(A) the level at the workplace P 1 is 87 dB(A) and at P 2 about 86 dB(A). Therefore it is not enough to check the declared value of the L pA – even at the operators position the sound power level L WA must be taken into account to evaluate the acoustic quality of the machine. This seems to be trivial but is often the reason for surprising results. A simple example is to reduce the free field sound pressure level L pA by installing a screening device between the operators place and the mainly radiating part, e. g. an infeed device direct in front of the workplace. But such a screen without absorption modifies only the directivity of the radiation – if the machine is installed in a room the additional sound intensity that is added is a function of the L WA and therefore the nice

level-valley produced with free-field conditions is filled up and the effect of the screen optimized with free field conditions cannot be seen with realistic installation conditions. 3. PREDICTION CALCULATIONS ON BASES OF DECLARED EMISSION VALUES

The A-weighted SPL at the workplace P 2 in the example Figure 1 – as well as at all other more dis- tant workplaces - caused by the machine M is calculated on basis of its L WA . With the GA-methods mentioned such calculations are possible for any environment with many other sources and installa- tions and in open spaces or in partially or total closed rooms. But: the accuracy of a calculation of the partial level caused by the machine at its attached work- place P 1 based on its L WA may be poor because the distance is small and the direct sound impact de- pends a lot on the detailed radiation directivity in the near field. Here comes the declared emission sound pressure level L pA into play – it is the best available information about the level caused by direct sound of the machine. Therefore the strategy is to replace the energy-related contribution calculated from the direct radia- tion from Sc to P 1 with inclusion of the reflection on the floor by the sound-energy according to the emission sound pressure level L pA declared by the machine manufacturer. This strategy has an im- portant influence on the calculation procedure – a preprocessing is performed where for each “ma- chine” with a declared value L pA this direct level is calculated (L pA,sim ) by the GA calculation method and using the source radiation with the declared L WA – contributions of other sources and reflections at surfaces other than those part of the machine model and the floor are neglected. In larger environments with many machines this preprocessing is performed stepwise machine for ma- chine. Then the GA – calculation is carried out in which all contributions based on the L WA emis- sion of all sources and up to high reflection orders are summed up at each workplace – the result is the A-weighted sound pressure level L AP,sim at all workplaces. Finally the energy contributions of the calculated direct sound L pa,sim is exchanged against the energy contribution based on the de- clared L pA by performing the calculation

𝐿 𝐴𝑃 = 10𝑙𝑔ሺ10 0,1∙𝐿 𝐴𝑃,𝑠𝑖𝑚 −10 0,1∙𝐿 𝑝𝐴,𝑠𝑖𝑚 + 10 0,1∙𝐿 𝑝𝐴 ሻ 𝑑𝐵 (1) Formula (1) looks quite simple, but it symbolizes more a strategy than an executable equation. This can be shown with the – still simple – machine model shown in Figure 2. Different to the example Figure 1 the model is now not only a point source, but reflects the structure of the machine body with two cuboids and simulates the real distribution of radiating parts with two area sources. The advantage relative to the simple point source simulation is that a more detailed model supports the dialog between the different parties involved in the planning process. The understanding of the noise aspects can better be improved if the noise relevant parts in a plant can be shown with such a more detailed model and even possible measures for noise reduction can better be explained and demonstrated. Figure 3 shows how the measurement of the sound power level with microphone positions distrib- uted on a half-sphere according to ISO 3744 can be simulated using the GA calculation. This allows to check and if necessary to readjust the sound power levels of the different radiating areas till the calculated L WA,sim is the same as the Sound power level L WA declared by the manufacturer. This “calibration” is certainly not necessary with the simple point source model Figure 1 – here the de- clared L WA is directly used as input-parameter for the point source emission. But it is an example

worm 2022

how effective our well known measurement techniques can be transferred to the calculation tech- niques with more detailed machine models and by applying GA calculation methods. Generally there are more noise relevant machines and devices in a workroom – this is “symbolized” by the two machines in Figures 3-5. The above mentioned first step of pre-processing – the calcula- tion of the emission sound pressure level L pA,sim on basis of the sound power level L WA of the mod- els own components – is shown with Figures 3 and 4 for the two machines. This pre-calculation must be obviously be integrated as an automatic procedure if noise predictions for working areas based on declared emission values are to be carried out with as little time as possible.

worm 2022

Figure 2: Model of a machine with two main

Figure 3: Calculation of the level 𝐿 𝑊𝐴,𝑠𝑖𝑚 ap-

xo 6 too a BEARS, Soa

radiating area sources and a workplace

plying the GA simulation

Figure 4: Calculation of the level 𝐿 𝑝𝐴,𝑠𝑖𝑚 apply- ing the GA simulation with machine 2

Figure 3: Calculation of the level 𝐿 𝑝𝐴,𝑠𝑖𝑚 ap- plying the GA simulation with machine 1

Figure 5: Calculation of the level 𝐿 𝐴𝑝,𝑠𝑖𝑚 with all contributions (all sources with reflections sur-

faces) at all workplaces

Finally the calculation of the sound contribution of all sources at all workplaces is shown with Fig- ure 5 (reflections at room surfaces are not shown to keep it simple). The last step in such an auto- mated noise prediction is to apply formula 1 for each workplace that is the operators position of a machine to get the final Table with the A-weighted sound pressure levels L AP for all workplaces. 4. ACTIVITIES NECESSARY TO SUPPORT THE DEMONSTRATED USE OF DE- CLARED EMISSION VALUES IN NOISE PREDICTIONS

It was often claimed that declared emission values are sometimes not reliable and that the effort re- quired to demand and control these values in accordance with the machine directive would often be disproportionate. An improvement is simple – these values should not be seen as a necessary evil to comply with le- gal regulations, but as important information about technical products in commercial transactions. If workrooms with noise relevant installations are planned, it should be a matter of course that for each noise-relevant device or machine with an assigned workplace, both emission values - the sound power level L WA and the emission sound pressure level L pA - are specified and included as a guaranteed value in the purchase contracts. This list of noise relevant technical devices with the two emission values guaranteed by the manufacturers is the starting point of the procedure. Only such contractually guaranteed values should be used as input parameters in order to predict the expected sound levels for normal operation in the workplace. The prediction calculation shows up where acceptable noise levels at work-places may be exceeded. The prediction calculation is part of the planning process to integrate noise reduction measures where it is necessary. But the method can only work if the noise levels at the workplaces are measured after commission- ing and compared with the predicted values and that in case these are exceeded the relevant emis- sion values are controlled. Only this procedure assigns an exceedance of the acceptable noise im- pact at a workplace to the exceedance of guaranteed emissions and thus enables the assignment of responsibility for the improvement. It would further be helpful if the software programs used to perform these calculations could be classified. This can be done on basis of activities similar to those that had been undertaken to sup- port quality assurance in calculations for environmental noise. The core is a set of testcases and the requirement to calculate results in agreed intervals – independent of the detailed calculation meth- ods applied. 5. CONCLUSIONS Experience in the last decades has shown that the responsibility for acceptable noise levels and acoustic conditions at work places, the declaration of noise emission values by manufacturers and the noise-related agreements and assurances in purchase documents are often treated as independent aspects. This should be improved by normative activities and a better information of the parties in- volved. It would save time and money if the procedure to declare the needed emission values, to ap- ply them in for noise prediction at workplaces, to perform the final measurements at the installa- tions when operated and to check the reason if levels are exceeded would be formalized. Even if it may need some more effort short term it will improve the communication between machine suppli- ers and users and will be of benefit for those affected by the noise. 1.

worm 2022