Particulate surface contamination is of concern in production industries such as food processing, aerospace, electronics and semiconductor manufacturing. There is also an increased awareness that surface contamination should be monitored in industrial hygiene surveys. A conceptual and theoretical framework for designing sampling strategies is thus developed. The distribution and spatial correlation of surface contamination can be characterized using concepts from geostatistical science, where spatial applications of statistics is most developed. The theory is summarized and particulate surface contamination, sampled from small areas on a table, have been used to illustrate the method. First, the spatial correlation is modelled and the parameters estimated from the data. Next, it is shown how the contamination at positions not measured can be estimated with kriging, a minimum mean square error method using the global information. Then methods for choosing a proper sampling area for a single sample of dust on a table are given. The global contamination of an object is determined by a maximum likelihood estimator. Finally, it is shown how specified experimental goals can be included to determine a suitable number of samples. The importance of including the spatial correlation into the calculations is demonstrated. Studies of the spatial correlation of particulate surface contamination will contribute to the understanding of particulate surface contamination processes.
|Journal||Journal of Aerosol Science|
|Publication status||Published - 1990|