A Study Of Fracturing Data Analysis Method Based On Water Hammer Theory And Signal Noise Reduction

Hydraulic fracturing is currently the main method of exploiting unconventional oil and gas. Actually, the pump-stop pressure and flow-rate data measured on the ground during hydraulic fracturing process contains lots of important information,such as the formation pressure, permeability and fracture half length. These parameters are significant judgments to decide post-fracturing mining system.During the pump-stop process of fracturing, fluid flow at the wellhead would decrease to zero in a very short period of time, making the pressure at the wellhead a sudden drop. And because of the inertia and compressibility of the flow,this pressure drop will propagate to the bottom of the well in the form of pressure waves, and reflected in the bottom of the well to form a rapid pressure fluctuations.This phenomena is called water hammer in hydraulics. The measured pressure data during construction is actually a superposition of the pressure drop caused by the fracturing pump water hammer and formation seepage. The focus of this thesis is to filter out the noise of water hammer from the construction data, and obtain the pressure drop curve caused by formation seepage, and then further carry out the well test analysisThis thesis presents a new method of pump-stop data-inversion analysis based on noise reduced pump-stop flow-rate and density-changing pressure data of hydraulic fracturing. Firstly, the ground pressure needs to be converted to well bottom pressure according to vertical pipe flow theory. The obtained bottom pressure data, also known as pump-stop pressure data, needs to go through noise reduction process to eliminate the interference of water hammer and other noise signals. At last, the noise reduced data, which contains the information of formation permeability and initial pressure, needs to be processed according to vertical fracture well test analysis. This new method, which has been proved reliable by compared with pressure build-up, makes the real-time evaluation of fracture and formation information possible.In the aspect of signal denoising, this thesis analyzes two methods, namely,digital filtering based on Fourier transform and empirical mode decomposition,which is a new time domain signal analyzing method, and compares the characteristics and advantages of the two methods.