Wireless monitoring of a steam boiler - performance measurements in industrial environment

Electromagnetic interferences and other disturbances restrict the coverage area and reliability of wireless sensor networks (WSN) in industrial environment. The need for empirical data to provide evidence of these disturbances has also been addressed. This paper presents results of screening tests performed in the steam boiler process to find out the effect of some factors on jitter variation. The effect of factors is estimated from coefficients of a partial least squares model, which was fitted to on the basis of a statistical quantity calculated from jitter variation data. The quantity used, Pearson mode skewness of histogram frequencies, was selected by comparing different statistical techniques. The selected quantity seems to be able to identify some candidates for more detailed tests, namely distance from nodes to the gateway, mutual interference, sampling interval, number of nodes, frequency converter/pumping installation and the block. The results show that the effect of the single Bluetooth and WLAN transmission is insignificant compared to other factors.