Research on the optimization method of selecting hearing protectors in power station

3.1. Preliminary screening of hearing protectors

We tested 112 kinds of hearing protectors insertion loss, they are all selected randomly from the market. During the test, we referred to the following standards: ISO 4869-3:2007 “Acoustics-Hearing protectors – Part 3: Measurement of insertion loss of ear-muff type protectors using an acoustic test fixture” and ISO 4859-2:1994 “Acoustics-Hearing protectors – Part 2: Estimation of effective A-weighted sound pressure levels when hearing protectors are worn” [5, 6].

We tested using the instruments by B&K 4128C Head and Torso Simulators, using the pink noise as the noise source. The sound pressure levels were tested with [1] the hearing protector and without the hearing protector, respectively. The results of errors are insertion loss, that is:

where $L_{l,Aeq}$ is the equivalent continuous A-weighted sound pressure level after wearing hearing protectors when workplaces are tested. $L_{p,i}$ is the actual test of each frequency band sound level in different workplaces. $L_{e,i}$ is the insertion loss of each frequency band sound level of different kinds of hearing protectors, $i$ is 33 frequency bands.

Fig. 1The spectrum of 1/3 octave of noise in 10 working places

The steps of testing are:

(1) The sound pressure level at the microphone is tested using PLUSE 7758 when Head and Torso Simulators are not inserted the hearing protector.

(2) Place the hearing protectors in the Head and Torso Simulators’ ear canal in sequence, and ensure the hearing protectors is located in the center of the microphone (as shown in Fig. 2).

(3) After about 30 s, the sound pressure level is tested again using PLUSE 7758 when Head and Torso Simulators are inserted the hearing protector.

(4) Repeat steps (1)-(3) three times with the same hearing protectors, the average value of three times is used to calculate the insertion loss.

Fig. 2The contrast of before and after hearing protectors insertion

The equivalent continuous A-weighted sound pressure level can be obtained by weighting the insertion loss of each frequency band. Based on GB/T 23466-2009 “Guidelines for selection of hearing protectors”, the effective value of the weighted sound pressure level is between 75 dB(A)-80 dB(A) after the operator is protected by the hearing protector. While less than 70 dB(A) indicates that the hearing protectors are overprotective, and the more than 80 dB(A) are lack of protection [4]. The Eq. (1) is used to calculate the equivalent continuous A-weighted sound pressure level with hearing protectors in each test place. There are 24 kinds of hearing protectors, the equivalent continuous A-weighted sound pressure level are between 75 dB(A)-80 dB(A) after wore, as shown in Table 2.

3.2. Statistical analysis

Principal component analysis (PCA) is a multivariate statistical method that transforms multiple indexes into a few comprehensive indexes based on the idea of dimensionality and the principle of minimizing the loss of data information [7].

In this paper, the insertion loss of each central frequency band of 24 kinds of hearing protectors is analyzed by PCA method, as shown in Table 3 and Table 4.

According to Table 3, it can be obtained that the principal component coefficients extracted in each frequency band are all more than 0.5, which indicates that the extracted principal components have a higher degree of interpretation of each variable. According to Table 4, two principal components were extracted in this calculation analysis. The characteristic root of the first principal component is 6.990, the variance contribution rate is 69.899 %. And the characteristic root of the second principal component is 1.349, the variance contribution rate is 13.492 %. The cumulative variance contribution rate of the two principal components is 83.391 %. It is further indicated that the extraction of the two principal components is appropriate.

As shown in Table 5, the correlation coefficient between the first principal component and frequencies from 63 Hz to 8000 Hz are all close to 1, this indicate that there are more hearing protectors with noise reduction range from 63 Hz to 8000 Hz in the tested hearing protectors. Similarly, there are more hearing protectors with noise reduction range with frequency bands of 31.5 Hz and 16000 Hz in the second principal component.

Each principal component can be weighted and summed according to the characteristic root of each principal component:

where $F_1$ and $F_2$ are two principal component functions respectively, $u_1$ and $u_2$ the characteristic roots of the two principal components respectively. According to Eq. (2), the ranking of principal component score and comprehensive score of 24 kinds of hearing protectors can be determined, as shown in Table 6.

We can see from the Table 6 that in the comparison of each brand of hearing protection in this study, the hearing protection of Xing Gong brand is more representative among 63-8000 Hz, and Pluggerz-sleep brand’s hearing protection is more representative in 31.5 Hz and 16000 Hz.

Combined with the above principal component analysis, it can be concluded that among the 112 kinds of hearing protection devices tested, most of the hearing protection devices are more suitable for the 7 sites studied: #1 generator layer, #1 outdoor water turbine, #1 indoor water turbine, #2 outdoor draft tube cone, #2 indoor draft tube cone, #3 outdoor draft tube cone and #3 indoor draft tube cone.