Development and calibration of a particulate mattermeasurement device with wireless sensor network function

Development and calibration of a particulate mattermeasurement device with wireless sensor network function
Park et al (2013)
A Zigbee-based ubiquitous sensor network (USN) has
many industrial applications and provides flexible
measuring environments. In particular, the USN syst
em can replace existing measuring devices in harsh
environments such as subway stations. To monitor the intensities of v
arious pollutants and air qualities in subway tunne
ls, this study applied the USN technique. A novel wireless sensor module,
PMX, was designed and manufactured to simultaneousl
y detect PM10 and PM2.5. Measurements were conducted at a subway station in Seoul. The PM concentrations using PMX were measured,analyzed, and compared with those obtained using an
established commercial dust spectrometer (Grimm Aerosol Technik,1.109). The measurements were performed from 24 March 2010 to 9 April 2010. PMX and the dust spectrometer measured PM10 levels of 98.3 and 40.7㎍/㎥, respectively, and PM2.5 concentrations of 86.5 and 16.6㎍/㎥
, respectively. Themonitored PM levels were investigated in a bimodal form during the sampling period. The PM10 and PM2.5 average correlations between PMX and the dust spectrometer were r2=0.81 and r2= 0.97, respectively. The two systems showed a similar time series trend, even though the measured
values differed. A simple correlation analysis of the two data groups
showed coefficients of determination of 0.7 for PM10 and 0.9 for PM2.5.
The PMX data were mostly concentrated around the trend curve. Therefore, calibration of PMX data
was required prior to use in the field. For the calibration, simple linear
regression and nonlinear regression were used. The resulting correlation coefficients of simple linear
regressions were 0.8 for PM 10 and 0.9 for PM 2.5 , whereas those for nonlinear regressions were 0.7
for PM 10 and 0.9 for PM 2.5.
The higher correlation coefficient for PM 10 by the nonlinear regression indicates that it is the better method for calibrating the system developed in this study.
PM2.5, PM10
No file