Research on Health Risk Due to Impulsive Noise and Vibrations

Research on Health Risk due to Impulsive Noise and Vibrations Research Results Conducted in Collaboration with NIOSH Professor Jay Kim Students: Xiangdong Zhu, Wonjoon Song University of Cincinnati March 2006 University of Cincinnati Applied Acoustics/Mechanics Lab Presentation Overview    Background Analytic Wavelet Transform as the Basic Signal Analysis Tool for Impulsive Events Hearing Loss Due to Impulsive Sound   Current work Long-term approach Some preliminary results Planned approach  Hand Arm Vibration Syndrome (HAVS)    Other Applications of AWT  Gunshot data/ear protector analysis  AWT based rotating systems analysis University of Cincinnati Applied Acoustics/Mechanics Lab Background: Conducting NIOSH-UC power tool research consortium, lack of method for assessment of exposure risk to impulsive noise and vibrations Impact Wrench 30   Complex noise environment in workplaces  Military noises  Pressure (Pa) Impulsive Noise Induced Hearing Loss (INIHL) 20 10 0 -10 -20 -30 0 0.01 0.02 0.03 0.04  Hand-Arm Vibration Syndrome (HAVS) due to impulsive vibrations Current codes are based on steady-state metrics ignoring temporal variation of spectral characteristics 0.05 0.06 Time (s) 0.07 0.08 0.09 0.1 Measured noise from power wrench University of Cincinnati Applied Acoustics/Mechanics Lab Issues in Risk Assessment of Impulsive Noise and Vibrations  Inherent difficulties of transient events   More parameters are necessary to characterize the event Difficult to formulate metric and relate it to experimental or demographic study results  Characterization technique: time-frequency analysis is necessary   Wavelet analysis should be a choice, but nearly entire existing practices and data are based on Fourier quantities Analytic Wavelet Transform (AWT): a hybrid of wavelet and FFT that works like a superb transient FFT analysis. All Fourier definitions, SPL, frequency spectra, can be defined in transient sense. University of Cincinnati Applied Acoustics/Mechanics Lab Current Status (1/2): Impulsive Noise  Impulsive Noise:  Current standards (OSHA, NIOSH, European standards) are based on equal energy hypothesis (85 dB, 6 dB exchange rule)    Use of dBA is considering spectral information Temporal information is considered in very limited sense through allowable maximum peak SPL Temporal variation of frequency spectrum is not considered AHAAH model by Price and Kalb: time domain simulation of human ear Chinchilla based study on INIHL by Hamernik et. al.    Research efforts to reflect temporal variations:   Expose chinchillas to steady-state and impulsive/complex noise Used Kurtosis as the metric to represent temporal variations University of Cincinnati Applied Acoustics/Mechanics Lab Current Status (2/2): Impulsive Vibrations  Impulsive Vibrations:   Similar to INIHL cases because of the transient nature, but dissimilar because hand and arm do not have spatial frequency sensor as the hearing organ Group of researchers at NIOSH Morgantown    Established frequency weightings for hand-arm vibrations and finger vibrations Developed standard test procedures, numerical models, demographic study and theoretical background Collaboration with UC is embarked in applying AWT and transient analysis technique to HAVS University of Cincinnati Applied Acoustics/Mechanics Lab Analytic Wavelet Transform (AWT): brief background (1/2)  Use variable time-frequency atom: source of the main advantage of wavelet analysis for transient signals  ˆ  u , s ( ) s picks up fast, high-frequency components Problems  /s  s Works in un-familiar terms to engineers and scientists: scale, wavelet intensity, etc. instead of frequency and amplitude  so ˆ  u , s ( ) o o picks up slow, so low-frequency components  / so  u ,s u  u ,s o o uo t University of Cincinnati Applied Acoustics/Mechanics Lab Analytic Wavelet Transform (AWT): brief background (2/2)    Hybrid of wavelet transform and Fourier transform Work in terms of traditional Fourier variables: frequency, amplitude and phase, however all as functions of time A perfect replacement of Short-time Fourier transform (STFT) for transient analysis   Ws (t )  f (u ), t , s   1  u t  f (u ) *t , s dt  u , s (u )     s  s  t2   1 2  (t )  g (t )e jt   2 1/ 4 e 2  (  )   jt e   Our version of AWT is set up so that each AWT provides a time history of 1/3 octave component of center frequency of    s University of Cincinnati Applied Acoustics/Mechanics Lab AWT: application example(1/2) Impact Wrench Inst. 1/3 octave spectrum 30 1/3 octave time history 20 10 AWT Pressure (Pa) 0 -10 -20 -30 0 0.01 0.02 0.03 0.04 0.05 0.06 Time (s) 0.07 0.08 0.09 0.1 Impulsive sound, time domain STFT T-F representation by AWT with cochlea mapping T-F representation by STFT Superiority of AWT compare to STFT is clear University of Cincinnati Applied Acoustics/Mechanics Lab AWT: application example(2/2) 7000 6000 5000 4000 3000 2000 1000 0 -1000 -2000 -3000 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 Airbag sound 100 dBA, 1250 Hz dBA, 125 Hz 100 0 0 0.05 0.1 time, s 0.15 0.2 0 0 0.05 0.1 time, s 0.15 0.2 T-F plot 100 dBA, 1000 Hz 0 0.05 0.1 time, s 0.15 0.2 dBA, 500 Hz 100 0 0 0 0.05 0.1 time, s 0.15 0.2 dBA, 2000 Hz 100 dBA, 4000 Hz 0 0.05 0.1 time, s 0.15 0.2 100 0 0 0 0.05 0.1 time, s 0.15 0.2 1/3 octave time histories University of Cincinnati Applied Acoustics/Mechanics Lab Hearing Loss due to Impulsive Sound (1/3): Current approach in pending NIH proposal Human NIHL model Chinchilla Test Data at SUNY-Plattsburgh Digitized Noise data AWT T-F noise metrics Chinchilla NIHL model Proposed research Statistical correlation study to choose the best metric Various, controlled noise set About 400 Chinchillas Existing data TTS, PTS, IHC and OHC loss data as functions of frequency University of Cincinnati Applied Acoustics/Mechanics Lab Hearing Loss due to Impulsive Sound (2/3): long-term plan Chinchilla ear model Ear simulation model output (basilar membrane displacement) Ear simulation model output Necessary development Human ear model Inter-species scaling law Environmental Noise AWT noise metric Human NIHL model Final form of implementation NIHL risk University of Cincinnati Applied Acoustics/Mechanics Lab Hearing Loss Due to Impulsive Sound: longterm plan (3/3): Develop inter-species scaling law Test Noise Data Simulation ear model for chinchillas Model output (basilar membrane displacement) development use Chinchilla NIHL model Simulation ear model for cats Repeat NIHL model development for cats using cat experiment data Model output Inter-species scaling law Compare to Simulated cat validate Cat NIHL model NIHL data Confirm scaling law University of Cincinnati Applied Acoustics/Mechanics Lab Example of Ear Model: AHAAH model developed by Price and Kalb Outer Ear Inner Ear Middle Ear Diffraction sound field Rdf Ldf 2P Air Plug Rpl Earcanal Concha L1 L2 A1 A2 L3 A3 length area Pe Lh Ue Eardrum conductive part Ldm Cdc Rdc 1:Nt Lds Cds Rds Vestibular Annular Volume ligament Incus Stapes Cochlea Li Ls Cal Ral Lv Uc Cmi Rmi Cis Ris Crw Round window Lo Pc Rc Ro Helicotrema Lpl P Cb Rh Cm Bulla Lever MalteoEardrum independent and area Incudal ratio joint part Incudostapedal joint University of Cincinnati Applied Acoustics/Mechanics Lab Hand Arm Vibration Syndrome (1/3) 200 3000 150 100 2000 x-dir acceleration [m/s 2] x-dir acceleration [m/s 2] 50 0 -50 -100 -150 -200 0 0.01 0.02 0.03 0.04 0.05 0.06 time [sec] 0.07 0.08 0.09 0.1 Time series 1000 0 -1000 -2000 -3000 0 0.01 0.02 0.03 0.04 0.05 0.06 time [sec] 0.07 0.08 0.09 0.1 T-F representation with ISO HA frequency weighting T-F representation with one of frequency weightings proposed for fingers University of Cincinnati Applied Acoustics/Mechanics Lab Hand Arm Vibration Syndrome (2/3) 10 1 Frequency weighted time history: reflects what hand arm feel 0 0.01 0.02 0.03 0.04 0.05 0.06 time [sec] 0.07 0.08 0.09 0.1 10 1 total acceleration [m/s 2] 10 0 total acceleration [m/s 2] 10 0 0 0.01 0.02 0.03 0.04 0.05 0.06 time [sec] 0.07 0.08 0.09 0.1 a f (t )   A ( f , t) i 1 f i nf 2 Sum frequency components at each time point University of Cincinnati Applied Acoustics/Mechanics Lab Hand Arm Vibration Syndrome (3/3) 10 1 Hazard dose curve HAVS threshold acceleration level 1 I1  N 1 Ij  N j 1 N total acceleration [m/s 2]  j 1 N a f (t j )  ath ath a f ( t j )  ath 2 1 N 1 N e I2   I j   N j 1 N j 1 10 0 1 e ath 0 0.01 0.02 0.03 0.04 0.05 0.06 time [sec] 0.07 0.08 0.09 0.1 Non-linear metric function Threshold level Metric based on frequency weighted time history University of Cincinnati Applied Acoustics/Mechanics Lab Gunshot sound analysis Outside of ear protector Inside of ear protector Reduce SPL University of Cincinnati Applied Acoustics/Mechanics Lab Other Interesting Application: Campbell diagram AWT based Fourier transform based Campbell Diagram AWT based Campbell Diagram Rotating system start-up analysis In-situ FRF construction without excitation University of Cincinnati Applied Acoustics/Mechanics Lab Questions/Suggestions? University of Cincinnati Applied Acoustics/Mechanics Lab