Comparison of three side-by-side real-time dust monitors in a duct using average and peak display dust levels as parameters of performance evaluation

In South Africa, the monitoring of dust in the mines is a requirement in terms of Section 12.2 and 12.3 of the Mine Health and Safety Act (MHSA) of 1996. In order to ascertain the magnitude and range of dust levels and to react when an unhealthy dust exposure occurs, a real-time personal monitoring instrument for mineworkers is undoubtedly required. This paper discusses a comparative study of the three real-time (PDR units) monitors in a duct using coal and sandstone dust. The Higgins-Dewell (HD) and Dorr-Oliver (DO) type cyclone operated in accordance with the international size-selective curve were used as 'true samplers. 'The average and peak display levels recorded by the three PDR units positioned randomly side by side, in the duct were analyzed using statistical techniques. The results of the study have showed that the dust levels measured with the three PDR units were not significantly different to the HD sampler data. Interestingly, the results showed significant differences in measured dust levels between HD and DO cyclones positioned side-by-side. The implication of this finding is that the majority of real-time monitors (e.g., Tapered Element Oscillating Microbalance (TEOM)) use these as a 'reference sampler. 'This means that, based on measured differences found between the two cyclones, the introduction of TEOM for legal monitoring purposes may create ambiguity in its current state, i.e., agreement on the use of 'true cyclone. 'The study demonstrated that, if the DO cyclone were used in the TEOM, it would measure significantly lower dust levels than the HD cyclone. Therefore, consensus on a 'true sampler for use in real-time monitors' must be established in the mining industry. Pair-wise t-test analyses were performed to compare the three PDR units using the average and peak recorded level. The study indicated that when peak value is used to evaluate the performance between instruments, resulted in different inferences on the recorded levels when compared with the average value. The implication of this is that in practice, the random selection and use of a real-time monitor for engineering dust control application may be in favor or against the seriousness of the dust problem. Although the recorded levels show the differences in dust levels, ANOVA results showed the contrary: dust type, monitoring units or position were not the sources of variation in the measured average and peak dust levels between the three PDR units. Light scattering monitors depend solely on air movement to move the dust particles into the sensing zone. It is unknown, if the particle charges have any specific effect in terms of their movement towards the sensing chamber that could have contributed to the recorded differences. It is proposed that, for real-time monitor evaluation, the use of 'peak display' level may ascertain the probable sources of variations. The intention of this paper is not to suggest that the peak levels should be used in place of average levels for exposure monitoring, rather an evaluation parameter in understanding of variations experienced by researchers.