Research On Drilling Abnormal Condition Monitor Method Based On Rapid Acquisition Of Four Drilling Parameters

Drilling Engineering is a complex process with a strong uncertainty,varieties of drilling abnormal conditions will be often taken place in the drilling process.Due to the low acquisition frequency of the traditional logging and drilling instruments,the serious loss of working condition key information in drilling process makes it impossible to exquisitely describe the drilling abnormal condition in real time and make an accurate warning and targeted decision.The diversified characteristics of four key parameters(hook load,standpipe pressure,rotary torque,wheel speed)in drilling process can not only distinguish the important drilling conditions,but accurately describe most abnormal conditions.Therefore,a monitor scheme based on the drilling four parameters is proposed,including rapid acquisition subsystem,abnormal condition identification and warning subsystem,simulation test subsystem,which can achieve stable data acquisition and accurate abnormal condition identification in real time,verify the reasonableness and effectiveness of method that proposed or used in scheme,provide laboratory testing platform for on-site applications.The abnormal condition identification and warning subsystem contains signal denoising,feature extraction and abnormal condition identification.In the respect of signal denoising,the concept that a sharp impulse noise which generated by drilling mud circulation resides in standpipe pressure signal and a random vibration signal which produced by torsional vibration is mixed in torque signal is presented firstly.Secondly,for the standpipe pressure denoising,in order to filter the sharp impulse noise and reserve step mutation components,a median filter method will be used.For rotary torque denoising,a combination between autocorrelation detecting and empirical mode decomposition method is proposed to save the high-frequency component in the signal when filtering random vibration signal.In the respect of feature extraction,a mutation detection method based on the wavelet transform modulus maxima will be used to achieve feature extraction of mutation and fluctuation,at the same time,a piecewise quadratic polynomial fitting method will be adopted to solve inaccurate description of the complex signals by using the traditional linear fitting and achieve extraction of slowly varying feature.In the respect of abnormal condition identification,for the reason that existing identification methods still have defects on accuracy and real-time,a new gray relative analysis method based on the correspondence relationship between the drilling abnormal conditions and the various feature of the four parameter signals in different working conditions is proposed by comparing the respective characteristics of exsiting drilling abnormal condition identification methods,including the improvement of ABO correlation degree by replacing the A correlation degree by slope correlation degree and increasing the number of the standard abnormal vectors for the sake of giving an identification result with high degree of accuracy,real-time and resolution.In the simulation test subsystem,simulation signal generation module,noise reduction and feature extraction module,abnormal condition and alarm module will be builded by constructing a simulation platform in order to achieve ultimately simulation operation and test of the monitor program of abmormal condition in laboratory.Simulation tests and field measured data analysis indicate that the methods,proposed or used in scheme,are reasonably effective in achieving signal denoising,accurate feature extraction and drilling abnormal condition identification and warning.