论文部分内容阅读
提出了一种对稠油热采井下光纤非本征法布里-珀罗干涉型(EFPI)永久压力传感器采集的压力信号进行小波多分辨率降噪的方法,可有效抑制传感器采集压力信号中的非平稳噪声。提出了一种基于信噪比(SNR)提升的小波分解层数确定方法,无需真实压力信号频率范围的先验知识,可通过扫描信噪比提升随小波分解层数的变化估计最优小波分解层数。在新疆某油田稠油热采井的现场试验结果表明,该方法可提高压力信号信噪比约2.6dB,且降噪后压力信号可显著提高对油田稠油热采井原油日产量的预测准确度。
A method for wavelet multiresolution noise reduction of pressure signals collected by extrinsic fiber Fabry-Perot interference (EFPI) permanent pressure sensors in heavy oil thermal recovery wells is proposed, which can effectively restrain the sensor from collecting pressure signals Non-stationary noise. This paper proposes a method for determining the number of layers of wavelet decomposition based on signal to noise ratio (SNR) enhancement. Without prior knowledge of the frequency range of the true pressure signal, the optimal wavelet decomposition can be estimated by scanning the signal to noise ratio (SNR) Layers. The field test of heavy oil thermal recovery wells in a certain oilfield in Xinjiang shows that this method can increase the signal-to-noise ratio of the pressure signal by about 2.6dB and the pressure signal after noise reduction can significantly improve the prediction of daily output of heavy oil thermal recovery wells in oilfields degree.