Development of Risk Assessment Method for Complex Noise. Project Period: July 1, 2006 - December 31, 2008.

摘要

Many workplaces are subject to complex noise environments where impulsive noises are embedded in the continuous background noise. Current noise guidelines recommend an exposure limit based on the equal energy hypothesis (EEH), thus overlooking the effect of temporal and spectral variations of the noise. This practice is widely believed inaccurate to assess the risk of complex or impulsive noises. An improved noise risk assessment method is necessary for more effective protection of workers from the noise-induced hearing loss (NIHL), the most common occupational disease. This research is a part of the long-term effort to develop a general noise risk assessment procedure. In this research, an advanced signal processing method and a general noise metric, two basic components of the noise risk assessment, were developed utilizing an existing set of chinchilla noise exposure data. One of the main difficulties in assessing exposure risk to impulsive or complex noise environments is the quantitative characterization of the noise. A highly transient event such as impulsive noise should be characterized in the joint time-frequency (T-F) domain because its time and frequency characteristics are inter-related. An advanced T-F noise characterization method was developed by refining and extending the analytic wavelet transform (AWT) method developed in the PI's previous research. The method obtains TF characteristics of the noise in a set of 1/3 octave time histories, from which the noise metric is calculated. Most noise guidelines currently assess the noise risk based on a single-valued metric, typically the A-weighted overall SPL. A more general noise metric that reflects the T-F characteristics of the noise is necessary to accurately predict hazard of a noise of general type. In this research, 14 new noise metrics that reflect T-F characteristics of the noise in distinctively different ways were designed. The best metric was identified by based on the statistical correlations of these metrics with hearing losses measured in chinchillas.