The Analysis Of Tubing Conveyed Perforation（TCP） Signal Based On Neural Network

The method is very backward that we judge whether perforating bullet explosion by touching the wall of tubing and calculate the number of perforation by counting the number of holes at present.We can not determine whether the perforating bullet explosion because of deep well.There is misjudging situation and it will affect next work even cause casualties at the same time.So we need an effective method to analyze the perforating bullet explosion and calculate the number of perforation.We can determine the frequency domain of the signal through spectrum analysis of TCP signal.We design butterworth bandpass filter to filter to filter the noise. Wavelet transform has the good time-frequency localization characteristic, it can remove the high frequency noise in the effective signal and reconstruct the TCP signal. We can take advantage of the first-order component that comes from the TCP signal after wavelet transform to determine the staring and ending points and calculate the number of perforation. We can get a neural network model through adjusting the network parameters. The neural network can identify the TCP signal after a large number of training sample data.In this thesis, we combine digital filter, wavelet transform and neural network together and put forward the analysis of tubing conveyed perforation signal based on neural network, the method solves the problem that judge whether perforating bullet explosion and calculate the number of perforation effectively, we can use neural network to identify the TCP signal. It provides accurate basis for the oil-field staff to judge the quality of the perforating.