The Research On Adaptive Filtering Methods For Pressure Signal Of Pipeline

Long transportation Pipeline has become a major transportation way because of it's high effective, convention, environmental protection.However, because ageing,erode,beening mangled by human, the leakage events is often happen.those events cause bigness economy losts and environment pollution. Therefore pipeline leak detection is an important issue in the pipeline transportation industry in the world wide range. Hardware-based leak detection can't operate continuously, the characters of real time,reliability,precision,sensitivity of software-based leak detection are unsatisfactory. During the far-ranging use of SCADA systems in pipeline transportation industry, the complex software-based leak detection based on the information process has become a focus and tidal current and get much attentions by academe.The original pressure signal contain amount of noise. Effective denoising is improtant for pipeline leakage detection. On the base of previous works, following works been done in this paper.Firstly, using the adaptive median filtering algorithm preprocess the original pressure signal. The simulations indicate that the method can effectively improve the impulse noise.Secondly, due to that the hard thresholding method based on wavelet transform is discontinuous on the thresholding value, and soft thresholding method also has disadvantages, a new adaptive wavelet thresholding method is proposed in the paper. The simulations indicate that the method provide a higher de-noising accuracy than hard and soft thresholding.Thirdly, in the wavelet thresholding de-noising methods, decomposition order is important. On the basis of the analysis, a self-adaptive method to determine a proper decomposition order is proposed according to the white noise verification. The method increase white noise remarkable verification of decomposition level. The simulations indicate that the method can effectively improve the de-noising efficacy.Fourthly, due to the immense effect that decomposition scale inflicted on wavelet decomposition, a maximum distance rule is applied as object function, on which the selective search is carried out and the optimum decomposition scale is obtained by adopting PSO algorithm. Based on this the adaptive selection of scale is realized. Simulation proves that adaptive scale is more suitable for the characteristics of signal; hence it is necessary for improving the effect of wavelet transformation.