Ocupational noise exposure measurement: searching for the best strategy

Arezes, Pedro M.

Production and Systems Department (Human Engineering Group) / School of Engineering of the University of Minho / 4800-058 Guimares, Portugal/ parezes@dps.uminho.ptBernardo, CarlosInstitute for Polymers and Composites / I3N / School of Engineering of the University of Minho / 4800-058 Guimares, Portugal/Mateus, OlgaInstitute for Polymers and Composites / I3N / School of Engineering of the University of Minho / 4800-058 Guimares, Portugal

ABSTRACT

Assessing workers’ occupational noise exposure can be a challenging task. Different strategies can be used for specific types of exposures profiles. However, these strategies can impact a lot on the time allotted for the noise assessment. Several industrial workplaces were selected, based on the variability of the exposure time and the tasks duration. Noise exposure assessment was made using the strategies proposed in ISO 9612:2009. The results obtained showed that the strategies recommended by ISO standard for each specific case represented the lowest time- consuming ones, both considering the preparation and measurement times and, with few exceptions, also led to lower values for the associated expanded uncertainty. Ultimately, the present work is a contribution towards improving the measurement procedure for the assessment of occupational daily noise exposure.

Keywords

Noise, occupational, measurement, uncertainty, stategy.

INTRODUCTION

Due to the magnitude of the noise exposure problem among European workers, European legislation has been published, namely a European Directive that establishes exposure limits for workers protection according to their exposure profile [5].

To assess whether a worker is exposed or not to a specific sound pressure level and to compare it with the established legal limits, it is necessary to carry out anadequate and reliable risk assessment. However, there are some difficulties to obtain a reliable assessment in occupational settings, due to several factors, such as the temporal and spatial variability of noise levels, the specific acoustic characteristics ofthe noise sources, the variability of exposure times, and the need to limit the number and duration of noise measurements, due to the inherent costs and production interference [4], as well as the type of equipment and procedures used in the measurement [8].

Taking this into consideration, the decision on which is the most appropriate measurement strategy will be an important issue when dealing with varied andcomplex exposure situations, such as those occurring in some occupational settings. The adopted measurement strategy or methodology should be adequate to measureand register, as accurately as possible, the “real” exposure situation and the corresponding noise exposure levels.

Focusing on the measurement process and the corresponding uncertaintyestimate the International Organization for Standardization (ISO) has published a revised standard on this domain, the ISO 9612: 2009 [7], which proposes three different strategies for assessing the occupational daily noise exposure, as well as the associated uncertainty. This revision was carried out due to the inadequacy of the estimate proposed in the previous version, in particular when dealing with small samples and high variability measurements [6]. Concurrently, several studies discussed the motivations and the basic principles in which the revised standard should be based [12][11][13][14].

Due to the relatively recent publication of the revised document, there are only few studies reporting its application [9][3] and there is still no detailed information onthe comparison between the three proposed measurement strategies when applied inreal work contexts [2].

The quality of the results obtained is even more important, knowing that small differences in exposure levels may imply different scenarios, considering the legal obligation for adopting preventive and protection measures. For example, the presentlegislation imposes the adoption of collective prevention measures and of hearing protection devices if daily exposure noise levels are above the defined action value. On the one hand, the adoption of hearing protection will imply an additional cost for thecompany and it can also represent an extra “cost” for the employee, due to the potential discomfort of its use [1]. On the other hand, a wrong decision in the adoption of hearing protection will imply that workers may develop hearing losses, and these losses will represent a cost for both, company and worker.

It is then clear that the adopted measurement strategy should be as reliable as possible, leading to the satisfaction of all the involved stakeholders.

Aiming at reducing the time and effort involved in the measurements, this studyexplores the application of the proposed strategies, by comparing the corresponding results, namely the uncertainty associated with field measurements of daily exposure levels.

METHODOLOGY

The current study describes a comparison between the proposed strategies to measure occupational noise exposure, as defined in ISO 9612 [7]. These strategies are the Task Based Measurement (TBM), the Job Based Measurement (JBM) and the Full Day Measurement (FDM). For that, different jobs/workplaces were selected according to their characteristics, allowing for a comprehensive comparison between distinct noise measurement outputs, namely the daily personal exposure level (LEX,8h), the associated expanded uncertainty (U), and the time spent in both the preparation and implementation of each measurement strategy.

Selection of the workplaces

This stage of the methodology consisted in the selection of the workplaces to be compared/analysed. This was done considering the requirements of each workplace in terms of its exposure characteristics. As stated in ISO 9612: 2009, the selection of the most appropriate strategy (called “recommended” strategy in the standard) to measure noise exposure will depend on the job characteristics, namely the work type and pattern, including the mobility of the workers and the complexity (number and predictability) of the task(s) carried out. ISO 9612: 2009 indicates that each strategy can be used in several types of workplaces, but for each type of workplace just one of the three strategies is recommended.

Table 1 presents the different types of work places and patterns and the corresponding recommended measurement techniques/strategies. For the purpose ofthis study, all these different possibilities were classified as different cases, in a total of seven cases. From the table it is possible to see that each case has a specific recommended strategy, but the other strategies can also be applied. Exception occurs for cases no. 6 and 7, in which the TBM strategy was not considered due to theunpredictable nature of the work, i.e. work with no tasks assigned or with no specific duration. These job characteristics did not allow the application of TBM, as it was not possible to define the requirements for its specific application.

Table 1 – Different cases considered in the study according to the type or pattern of work (adapted from ISO 9612, 2009).

*R: Recommended; O: Others

Several companies were contacted in order to select a range of work contexts comprising the different characteristics’ combinations mentioned in table 1.

The sample considered for this study included 23 industrial companies, majority (74%) being SMEs. In terms of activity, the companies chosen included, for example, the textile, metalworking, construction, car repair and inspection, mining, electricalcables, tyre and shoes manufacturing sectors. From each company, some workplaces/work stations were selected and included in the sample analysed, having in mind the need to cover all the seven cases defined in table 1. According to the selectedsample, 43 workstations and 236 workers were involved for each of the types of work considered.

Noise exposure level assessment

Once all the workstations, or workplaces, were selected, the second stage consisted on the application of the three (or two, for cases no. 6 and 7) proposed measurement strategies for assessing noise exposure. As referred before, in the selection of the measurement strategy, according to ISO 9612: 2009, it is proposed that only one shall be selected. However, the aim of the present study was to compare the results obtained when applying the different strategies, regardless of their reported suitability to the characteristics of the workplace. Jointly with the noise exposure measurements, the associated uncertainty was also estimated, according to the procedures defined in ISO 9612: 2009.

Even though the measurement procedures are not explained in detail in this paper, the application of the strategies implied the adoption of some specific actions, including a previous work analysis that was carried out in order to provide sufficient information about the work and the workers under consideration.

For all the measurements, the basic measurement quantity considered was the Lp,A,eqT, and the corresponding LEX,8h. Finally, the associated handling error and uncertainty (or uncertainties) were computed for each measurement.

As suggested in Pinto & Stacchini [10], the influence of the technician was not taken into account, considering that all the measurements were carried out by the same practitioner and in accordance to a pre-defined measurement protocol. The application of each strategy was also carried out considering the steps foreseen in the standard. Depending on the selected strategy, different steps were adopted, which are briefly presented in the following paragraphs.

TBM strategy

For the application of this strategy, both class 1 and 2 integrating-averaging sound level meters were used, as well as the corresponding calibrators. The sound level meters as well as the calibrators had been verified/calibrated by a certified entity.

The general steps adopted for the application of this strategy were the division of the workers’ nominal day into tasks, the determination of the tasks duration, the measurement of Lp,A,eqT,m for each task, and the calculation of the contribution of eachtask to the daily noise exposure level. Finally, the uncertainty was also estimated,considering that each measure should always take longer than five minutes. In applying this strategy it was assumed that for tasks with randomly fluctuating noise the duration of the measure should be representative of the whole task.

JBM strategy

As recommended in ISO 9612:2009, the application of this strategy implied preferentially the use of a noise dosimeter. In this study, several class 2 dosimeterswere used, as well as the corresponding calibrators. The experimental procedure also included the required steps, namely the computation of the daily noise exposure levels for workers in a homogenous exposure group and the corresponding uncertainty estimate.

FDM strategy

Similarly to the previous described strategy, the application of the FDM strategy implied the compulsory use of noise dosimeters. The experimental procedures weresimpler in this case, because they only included the determination of daily noiseexposure level by means of one, or more, full-day dosimeter measurements. As suggested by ISO 9612:2009, whenever the differences between the three initialmeasurements differed by more than 3 dB(A), two additional full-day measurements were taken.

Uncertainty and time estimate

The estimate of the uncertainty of the measurements was done transversally for all the three strategies, according to the procedures proposed by ISO 9612:2009. Despite that, and depending on the applied strategy, different sources of uncertainty were considered. Uncertainties can be caused both by errors or natural variations in the work environment. Accordingly, the main sources of uncertainty considered were(i) variations in the daily work, operating conditions, uncertainty in sampling, etc.; (ii) instrumentation and calibration; (iii) microphone position; (iv) false contributions, forinstance from wind, airflows or impacts on the microphone and the rubbing of themicrophone on clothing; (v) lacking or faulty work analysis; and (vi) contributions from non-typical noise sources, such as speech, music (radio), alarm signals and non-typicalbehaviour.

As suggested in ISO 9612: 2009, when the uncertainty contribution due to sampling (the parameter c1u1 in the uncertainty calculation) was greater than 3.5 dB,changes were made to reduce this value, either by modifying the composition of the homogeneous group or by increasing the number of measurements.

Considering that the selection of the adequate measurement strategy seems to be a balance between the magnitude of the expanded uncertainty and the time needed(and easiness) to apply a specific strategy, it is important to assess both parameters. Accordingly, all the practitioners involved in the measurements were requested to estimate the time they spent preparing the application of each measurementstrategy/method (for example, carrying out the work analysis, interviewing workers and supervisor, identifying the different sectors and special noise events, selecting the strategy, defining the measurement plan, etc.), which was called “preparation time”. They were also asked to assess the time spent on the measurement itself, byestimating the cumulative periods of the measurements with the sound level meter and/or the cumulative periods in which the dosimeters had been applied, expressed in hours.

As the majority, if not all, of the measurements involved more than one worker and in some cases a large group of workers, both the preparation time and the measurement time were divided by the number of workers in each measurement. Thusboth variables are expressed in hours per worker.

RESULTS AND DISCUSSION

After the application of the different strategies to determine noise exposure, the final main outputs were the LEX,8h for each workplace and the corresponding expanded uncertainty (U), as well as an estimate of the time devoted to the preparation of the measurements and to the corresponding implementation (measuring process).

Table 2 presents the overall results for the time estimate, for each case studied and for each type of strategy, i.e., for the “R-recommended” strategy in ISO 9612:2009 and for the “O-other” strategy(ies).

Table 2 – Preparation and measurement time (in hours per worker) for each of the cases studied.

* R - Recommended; O - Other.

From table 2 it is possible to verify that, in the majority of the cases, the “recommended” strategy for each analysed case represents the lowest time-consuming strategies, both considering the preparation and measurement times. An exception occurred for case no. 5, where the total average time per worker for the other strategies (3.20 h/worker) is lower than that for the recommend strategy (1.94 h/worker). For this specific case, it is possible to see that the result is somehow influenced by the measurement time, as the preparation time is one of the lowest of all the recommended strategies (0.15 h/worker). In this case, the work pattern is unpredictable, thus the preparation time is low but the measurement duration should be longer, as it is necessary to cover a significant part of the work period.

As LEX,8h for each of the considered cases were determined by the application of the three strategies (two for the last two cases, see table 1), 120 exposure levels were determined, ranging from 83.3 dB(A) and 87.7 dB(A).

Table 3 presents the values for the expanded uncertainty (U) for each of the cases studied (in terms of work pattern and type) and for each type of strategy, corresponding to an overall uncertainty average of 3.82 dB (±0.77). The expanded uncertainty was, for all cases, obtained from the computation of the combined standard uncertainty (u) multiplied by k = 1.65, which corresponds to a one-sided coverage probability of 95%.

Table 3 - Expanded uncertainty (in dB) and difference for the combination: cases studied/

* differences are significant at the 0.05 level ** differences are significant at the 0.01 level

The differences for the expanded uncertainty were analysed by applying an independent sample test using the t-test for testing the equality of means. Previous to this test, the Levene’s test for the equality of variances was also applied and considered for a 0.05 level.

Hence, the t-test was performed by assuming, or not, the equality of the variances for both means. Results of the computed differences and the corresponding p-value for the statistical test are also presented in table 3. According to the average values for the uncertainty presented in the table, it is possible to confirm that, with the exception of case 7, the recommended strategies present lower values for the expanded uncertainty and that the differences between the values obtained are statistically significant, with the exception of cases number 1 and 6, which are still positive but not statistically significant.

The last column of table 3 presents the difference between the uncertainty of the recommended values (R) and the others (O). A negative value means that the recommended strategy has a lower value and, inversely, a positive value means a higher uncertainty average.

Only in the case no. 7 a positive difference was observed, which was also statistically significant at a 0.01 level. For this case, the “other” strategy is the FDM strategy and its application resulted in an uncertainty value lower than that of the JBM. For this particular case, where no tasks are assigned, the variability in terms of noise exposure can be very diverse and the application of FDM may result in lower uncertainty values, but it will also imply a higher time per worker for preparation and measurement.

The same test for the equality of means was also applied for the entire data (N= 120), considering the expanded uncertainty for the recommended methods (3.53 dB±0.60) and for the other methods (4.00 dB±0.82), as suggested in ISO 9612: 2009.

For this test, the equal variances were assumed based on the Levene’s test (p=0.015), and the difference of -0.47 dB was found to be statistically significant at a 0.05 level. These results seem to reinforce the idea that the application of the recommended strategies will results in a lower value for the uncertainty.

The adequate selection of a specific strategy needs to consider both the resulting uncertainty and the difficulties in the application of each specific measurement method, considering the associated preparation and measurement times. Therefore, for analysing the association between time and uncertainty, the Spearman’s rho (N=120) correlation coefficients were computed and analysed. According to the results obtained, neither the correlation between the uncertainty and the preparation time per worker (r=0.156, p=0.097), nor the measurement time per worker (r=0.136, p=0.150), were statistically significant.

Accordingly, it seems that there is no association between the time required for the preparation and measurement and the corresponding uncertainty, which in a certain way leads to the conclusion that the selection of a specific method can, at least in principle, be made taking into account the total time involved, and that will notaffect the uncertainty.

From the results obtained it is possible to make a comparison between strategies and their corresponding results. However, for this comparison it is necessary to include only the values obtained when the strategy is considered to be a “recommended”, otherwise results will be biased by the application of a specific strategy for a situation where it is not appropriate. For example applying FDM in a steady noise workplace will represent a long, but obviously dispensable, time of measurement. According to this, values obtained for recommended strategies were compared and are presented in figure 1.

Figure 1 – Average total time spent per worker and the corresponding average uncertainty for each (recommended) strategy.

When recommended techniques are compared, it is possible to verify that there is only a small variation on uncertainty, with values ranging from 3.72 dB (FDM) to4.94 dB (TBM). For its turn, the differences in the parameter “time spent by workers” are higher, as they range from 1.75 h (TBM) to 2.87 h (FDM).

It is also possible to verify that JBM and FDM have, in the majority of cases, very close or even equal uncertainty results, with a slight tendency to higher values in the case of JBM. It was expected that the values resulting from these two strategies would be closer, as both use the noise dosimeter and entail long sample durations. Accordingly, and considering the results obtained, it is questionable whether there is room for the application of the JBM strategy, or if it is more straightforward to opt always for the FDM strategy, as the experimental procedures are simpler (even if the measuring time is longer).

As mentioned in ISO 9612: 2009, the TBM and JBM strategies are not mutually exclusive, as they can be applied for almost the same situations, excepting when no tasks are assigned or their duration cannot be identified. Comparing the results of figure 1 for these two techniques it is possible to confirm that the uncertainty average values obtained are relatively close.

Another important characteristic of JBM is that this technique can be applied for all the analysed cases, even if, as for example in case no. 1, its application may result in an unnecessary long measurement.

CONCLUSIONS

According to the obtained results it seems that the selection of a specific strategy can have an important effect in the uncertainty estimate, hence in the interpretation of the results obtained.

In the majority of the studied cases, the strategies recommended by ISO 9612: 2009 for each specific case represented the lowest time-consuming ones, both considering the preparation and measurement times and, with few exceptions, also led to lower values for the associated expanded uncertainty. Moreover, the difference for the uncertainty values between the recommended and the other strategies was demonstrated to be statistically significant.

For all cases, when the time required for preparation and measurement is compared with the corresponding uncertainty, no statistically significant associationseems to be found. In a certain way, this may lead to the conclusion that the selectionof a specific strategy can be made taking only into account the time needed to apply the technique.

If the selection is exclusively based on uncertainty values, it can be questionable whether there is room for the application of the JBM strategy, or if it ismore straightforward to opt always for the FDM strategy, as the experimentalprocedures are simpler (even if the measuring time is longer). However, it is important to recall that JBM is the only strategy that can be applied for all the analysed cases.

Regardless of the previous conclusions, and although the sample of workplaces of the present study is quite comprehensive, it is important to acknowledge that it wasrestricted to industrial environments, which is an important limitation. The possible distinct workplaces characteristics in non-industrial environments may affect some of the conclusions presented herein.

ACKNOWLEDGMENTS

The authors would like to thank to all the companies involved in the present study and the respective workers, who voluntarily accepted to take part of it.

REFERENCES

• 1. Arezes, P. M. & Miguel, A. S. (2002). Hearing protectors acceptability in noisy environments. Annals of Occupational Hygiene, 46, 531536.
• 2. Arezes, P.M., Bernardo, C. & Mateus, O. (2011). Measurement strategies for occupational noise exposure assessment: a comparison study in different industrial environments. Submitted to publication.
• 3. Asselineau, M. & Gaulupeau, A. (2010). Noise exposure assessment: Some arguments regarding the simplification of ISO 9612 methods. INCE Conference Proceedings, 221, 254 259.
• 4. Cagno, E., Di Giulio, A. & Trucco, P. (2005). Statistical evaluation of occupational noise exposure. Applied Acoustics, 66, 297318.
• 5. European Parliament EP (2003). In: EC (ed.) European Directive 2003/10/EC on the

minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (noise).

6. Grzebyk, M. & Thiery, L. (2003). Confidence intervals for the mean of sound exposure levels. AIHA Journal, 64, 640-645.

• 7. ISO (2009). ISO 9612. Acoustics Determination of occupational noise exposure Engineering method. Geneve.
• 8. Maue, J. (2006). Determination of Noise Exposure by means of Noise Dosimeters Measurement Method and Uncertainty. Paper presented at the Euronoise 2006, Tampere, Finland.
• 9. Maue, J. (2008). Measurement according to the new ISO DIS 9612. The Journal of the

Acoustical Society of America, 123, 3526.

• 10. Pinto, I. & Stacchini, N. (2006). Uncertainty in the evaluation of occupational exposure to wholebody vibration. Journal of Sound and Vibration, 298, 556562.
• 11. Thiery, L. & Ognedal, T. (2008). Note about the statistical background of the methods used in ISO/DIS 9612 to estimate the uncertainty of occupational noise exposure measurements. Acta Acustica United with Acustica, 94, 331334.
• 12. Thiery,  L.  (2007).  Relationships  between  noise  sampling  design  and  uncertainties  in occupational noise exposure measurement. Noise Control Engineering Journal, 55, 511.
• 13. TurunenRise, I., & Ognedal, T. (2002). Work place noise: Nordic requirements, standards and measurement methods. Paper presented at the NAS meeting 2002, Bergen, NO.
• 14. TurunenRise, I., Ognedal, T., Nielsen, M., Nilsson, P. Å. & Laitinen, H. (2003). Improved measurement methods of occupational noise exposure. Acta Acustica (Stuttgart), 89, S49 S50.