Electromagnetic Flowmeter's Signal Processing Based On DSP

Electromagnetic flowmeter's output signal was not only weak, but also contaminated by some noise. Sometimes it was submerged by signal completely. So that the signal's SNR was very low. Therefore, de-noising become an important step to Electromagnetic flowmeter's signal processing. As a new signal processing method, wavelet analysis decomposed different kinds of frequency element to non-overlapped frequency bands, and this method put forwards an effective way to signal filter, signal-noise separating and character picking-up. Especially, it had well de-noising apability. The realization of digital filter was independent of hardware, and only need some software code. It could avoid some design difficulties when the cut-off frequency of analog filter circuit changes, so that filtering operation would become easier and more convenient. Meanwhile, with the rapid development of the DSP technology, as well as with the aid of the powerful data processing capacity and speed of DSP processor, many complex algorithm could be successfully applied in DSP processor to solve many problems in engineering. After designing the signal modulator circuits of electromagnetic flowmeter's weak output signal, this dissertation combined the realization of the digital filtering algorithm and improced wavelet threshold de-noising algorithm on the DSP processor to deploy the study. Its content were as follows:Firstly, electromagnetic flowmeter's noise jamming and the fundamental theories of wavelet analysis were discussed in detail. Based on the traditional wavelet threshold de-noising algorithm, this dissertation analyszed the deficiency of this algorithm, and proposed an improved de-noising algorithm.Secondly, electromagnetic flowmeter signal processing's system block diagram had been given. The signal modulator circuits of electromagnetic flowmeter's weak output signal had been designed.Thirdly, the basic principles andrealization steps of digital filtering algorithm and improved wavelet threshold de-noising algorithm were discussed. In the part of the design of digital filtering algorithm:firstly compared the advantages and disadvantages of two digital filters, and finally employed the FIR filter; secondly, used the conception of convolution to design the filtering algorithm; thirdly, adopted simulated data to verify the function of this algorithm. In the design of improved wavelet threshold de-noising algorithm, design program code of improved algorithm's four steps, which were wavelet's first decomposition; wavelet's second decomposition using high-frequency coefficients of first decomposition; threshold processing of high-frequency coefficients and wavelet reconstruction separately, then used the written code to deal with the noisy signals generated in the Matlab software to compare the result of these two algorithms. The processing result implied that the improved algorithm had a better effect in signal denoising, and retained signal's useful high-frequency information well. Finally, using the above two denoising algorithm to deal with signal modulator circuits's output flux signal, and it could get good denoising effect. After calibration, the accuracy of flow measurement reach±1%.