Transcript W 1 = 86 dBA

Tools for optimizing
the installation of
warning sounds in
noisy workplaces
Chantal Laroche, Christian Giguère,
Rida Al Osman and Yun Zheng
2010 NHCA
Conference
February 25-27, 2010
Background
 Safety in the workplace:
 Noise is a key hazardous factor and can cause hearing loss
 Acoustic warning signals are crucial to alert workers and reduce the
risk of accidents
 Safety is dependent on alarm recognition and communication ability in
the presence of background noise
 Hearing protectors:
 Minimize the adverse effects of noise in the workplace
… BUT
 Can compromise the audibility of warning signals
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Background
 Current practices for installing warning devices:
 ISO 7731: “Danger signals for public and work areas”
 Devices typically installed on walls or ceiling at a certain distance from
workstations
 Installation is poorly regulated and submitted to intuition
 Factors that must be taken into consideration:
 Audibility in the workplace
 Sound propagation from the device to the various workstations (direct
sound path and reflected sound waves)
 Noise field (level, spectrum, type)
 Warning signal design (frequency components, level)
 Number, location and sound power level of warning devices
 Effects of hearing status (hearing thresholds, frequency selectivity) and
hearing protectors
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
The problem
How many alarm devices needed?
Where?
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Sound power level?
General Framework
WORK ENVIRONMENT
Room layout, Reverberation time,
Workstation coordinates (Xk, Yk, Zk)
WORKERS
Hearing
Frequency
thresholds selectivity
Warning signal
target levels
[TLlow , TLup ]
AlarmLocator
Number Coordinates Power level
(Xi, Yi, Zi)
ND
Lw
WARNING DEVICES
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Detectsound
Noise
Lp
HPD
attenuation
WORKSTATIONS
Detectsound
The outcome of “Detectsound” is a design window for warning sound
levels at each workstation W
110
Window:
dB SPL
100
De sign window
TLup = THR + 25 dB
90
TLlow = THR + 12 dB
80
Background noise
TLmax = 105 dB SPL
70
60
125
200
315
500
800 1250 2000 3150
Frequency (Hz)
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
AlarmLocator
The outcome of “AlarmLocator” is a solution of warning devices (D) to
meet “Detectsound” targets at all workstations (W)
Simulations:
Mirror image method
(early reverberation)
W2
D1
W1
Classical room acoustics
(late reverberation)
Solutions:
W3
Number of devices
Location on walls
Sound Power Level
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
D3
D2
Validation
 Workshop Area (8.77m  14.75m  6.62m) in Building M-37
at NRC (Ottawa).
 Experimental set-up:
– 3 workstations (W1-W3)
– 2 noise sources (N1-N2)
– 2 noise types
(continuous, impact)
– 3 alarm frequencies
(500, 1000, 2000 Hz)
– 5 subjects
– Open ear + HPD
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Validation
 Psychoacoustic validation of “Detectsound”:
– Masked THR prediction error (0.0  1.4 dB)
– Preferred level for a 3-tone alarm (18.3 dB  3.1 dB above THR)
– Detectsound design window (12 to 25 dB above THR).
 Acoustic validation of “AlarmLocator”:
– 3 source positions, 3 workstations, 3 frequency bands (n=27)
– Omnidirectional source B&K 4295 (known power level)
– Workstation SPL prediction error (0.1 dB  0.9 dB)
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Simulation Study
 Goals:
1. Investigate the effects of hearing protectors on the warning sound
design window (TLlow, TLup) for individual workers at specific
workstations.
2. Investigate warning sound design constraints when workers with
different hearing status share a common work area.
Interaction of hearing loss and hearing
protectors on the perception of warning sounds
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Simulation Study
 Work Area:
– Reverberation times: 0.9s (250-1000 Hz), 0.8s (2000-4000 Hz)
– 3 workstations (W1 = 86 dBA, W2 = 91 dBA, W3 = 96 dBA)
– Low-frequency noise (upper spread of masking)
Level (dB SPL)
Noise
Y

X

W1
90
W3

80
70
60
31.5
W2 
63
125 250 500 1000 2000 4000 8000
Frequency (Hz)
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
14.75 m
8.77 m
100
Simulation Study
 Workers (Hearing Status):
Indiv1 (mild HL):
Indiv2 (moderate HL):
Indiv3 (mod. severe HL):
0
Male 40 yr (20 yr @ W1= 86 dBA)
Male 50 yr (30 yr @ W2= 91 dBA)
Male 55 yr (35 yr @ W3= 96 dBA)
HEARING THRESHOLDS
10
20
30
40
50
60
70
80
500
0
Broadening factor
Hearing Level (dB HL)
–
–
–
1000
2000 3000
4000
6000
FREQUENCY SELECTIVITY
1
2
3
4
5
500
Frequency (Hz)
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
1000
2000
3000
Frequency (Hz)
4000
Simulation Study
 Hearing Protectors (CSA Z94.2-02; EN 458-2005):
Selection:
MINIMUM ATTENUATION
Protected levels:
Overprotection:
Acceptable:
Optimal:
Acceptable:
< 70 dBA
70  75 dBA
75  80 dBA
80  85 dBA
Insufficient:
> 85 dBA
0
Attenuation (dB)
Class C:
Leq ≤ 90 dBA
Class B: 90 < Leq ≤ 95 dBA
Class A: 95 < Leq < 105 dBA
5
10
C
15
20
B
25
30
35
125
A
250
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
500
1k
2k
3k
Frequency (Hz)
4k
6.3k
8k
Simulation Study
Common design window for 3 workers at
W1 (86 dBA)
Common design window for 3 workers at
W3 (96 dBA)
110
110
Poste
W3 3
Poste
W 1
100
100
Niveau
SPL)
Level (dB
dB SPL)
Level dB
Niveau
(dBSPL)
SPL)
1
90
80
70
90
80
70
Class C
60
125
Class A
60
200
315
500
800
1250
2000
3150
125
Fréquence
Frequency(Hz)
(Hz)
At high frequencies, warning sounds cannot
simultaneously meet requirements for Indiv1
and Indiv3. No design window above 2500 Hz.
200
315
500
800
1250
2000
3150
Fréquence
Frequency(Hz)
(Hz)
Design window limited by 105 dB SPL
maximum at low frequencies and by the
conflicting requirements for Indiv1 and Indiv3
at high frequencies.
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Simulation Study
 Installation of warning devices:
– Meet common design windows at the 3 workstations
– Four warning signal components (500, 600, 1000, 1600 Hz)
– AlarmLocator (ND = 1)
D1
X
(m)
Y
(m)
Z
(m)
Sound Power
Level Lw (dB)
14.8
5.0
3.0
500 Hz: 116
600 Hz: 114
1000 Hz: 114
1600 Hz: 110
X

W1
W3

D1
W2 
14.75 m
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
8.77 m
Y

Simulation Study
 Results:
– Warning sound design window is fairly insensitive to attenuation of
hearing protectors for workers with normal hearing or mild hearing
loss.
– Design window is highly sensitive to attenuation achieved at high
frequencies (>2000 Hz) for workers with moderate or greater
hearing losses. Accurate warning sound solutions require accurate
estimates of field attenuation.
– Design of warning sounds in a workplace can become a challenge
when workers with different hearing status share a common work
area.
– Warning sounds in the frequency range from 500 to 1600 Hz is
recommended (in agreement with ISO 7731).
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010
Conclusions
 Detectsound provides valid estimates of the optimal design window for
warning sounds based on a psychoacoustical analysis of the relevant
parameters at each workstation.
 AlarmLocator provides possible solutions for the number and placement
of warning devices based on a simulation of the sound propagation in
the work area.
 In general, warning sound frequency components in the range 500-1600
Hz are recommended for workers with hearing loss or wearing hearing
protectors (ISO 7731).
 Care must be taken not to overgeneralize recommendations to special
situations, such as high-frequency noise environments, low-frequency
hearing loss or unusual attenuation profiles.
2010 NHCA Conference – Orlando, Florida
February 25-27, 2010