An Improved Method for the Processing of Signals Contaminated with Strong Common-Mode Periodic Noise in Correlation Velocity Measurement

This is the latest version of this item.

Electrostatic sensors have been successfully used for the velocity measurement of pneumatically conveyed particles and the rotational speed measurement. However, the signal from an electrostatic sensor is usually vulnerable and susceptible to contamination in a hostile environment. The acquired original signal may be contaminated by different types of noise that can be within or outside the frequency range of the signal. This paper presents a novel correlation signal processing method to minimise the impact of noise in the signal through a de-noising process and hence improve the performance of correlation-based measurements in general. The method is applied to the rotational speed measurement based on electrostatic sensors in particular. The de-noising process is an essential task in digital signal processing to improve the signal-to-noise ratio before implementing the measurement algorithm. A hybrid de-noising method is proposed to combine a cut-off frequency method to remove the noise components outside the signal bandwidth and a median filter to smooth the signal. Subsequently, the signal is de-noised in the time domain by employing an advanced digital filtering method based on correlation techniques to suppress the noise frequency components mixed with the original signal. The rotational speed measurement system with the proposed technique has proven to be effective in de-noising signals that are buried in noise with which they are correlated. Moreover, the technique is capable of producing more accurate and repeatable measurements with a wider measurement range than the existing system. Experimental results suggest that the relative error of the improved system is mostly within ±0.1% over the speed range of 300 rpm - 3000 rpm and within ±0.2% over the speed range of 40-300 rpm.