Annular Flow Model And Kick Analysis Method For Manage Pressure Drilling Under HTHP

Deep and ultra-deep layers have the characteristics of numerous fractures and caves and narrow safety density windows,causing problems such as kick and leakage.In order to deal with these problems,managed pressure drilling technology is increasingly used.Since high temperature and high pressure are common in deep and ultra-deep layers,and the density and rheology of oil-based drilling fluids are significantly affected by temperature and pressure,ignoring this effect will lead to a reduction in the accuracy of hydraulic calculations for controlled pressure drilling,thereby increasing the difficulty of pressure control.Aiming at the wellbore flow issues during pressure-controlled drilling in high-temperature and high-pressure formations,this paper conducted the following research:1.A high-temperature and high-pressure drilling fluid PVT experiment was carried out to measure the density of white oil,diesel and actual drilling fluid on site under different temperature and pressure conditions.The measurement temperature range was20°C to 180°C,and the measurement pressure range was 0.1MPa to 180 MPa.There are85 measuring points in total.The calculation error of the drilling fluid density component model was analyzed,the calculation error was predicted using the neural network method,and a modified component model was established.The density prediction accuracy of the modified composition model is high,and the maximum error compared with the actual measurement results is 1.8%.2.A high-temperature and high-pressure drilling fluid rheology experiments was carried out to measure the shear stress of three actual oil-based drilling fluids at different speeds(3rpm,6rpm,100 rpm,200rpm,300 rpm,600rpm).The measurement temperature range is 60℃ to 160℃,and the measurement pressure range is 0.1MPa to150 MPa.There are 56 measuring points in total.Based on the weighted linear regression method,high-temperature and high-pressure drilling fluid shear stress prediction models under different rotational speeds were established.The model was verified using the Holdout cross-validation method.The results showed that the shear stress predicted by the model in this paper was highly accurate,with an average error of4.8%.3.The main factors affecting the cross-section gas content were analyzed,and the input parameters of the convolutional neural network method were determined.A cross-section gas content prediction model based on the convolutional neural network,which improves the accuracy of the cross-section gas content prediction,was established.Compared with the actual measured data,the average error was 8.7%.4.Taking into account the changes in density and rheology of drilling fluids under HTHP(high-temperature and high-pressure),based on the two-phase drift flow model and combined with the cross-sectional gas content prediction model,a two-phase flow hydraulic calculation model for managed pressure drilling under HTHP was established.The model was solved using the difference method.Based on the established model,the well killing process of an overflow well was simulated.The calculation results were in good agreement with the actual vertical pressure and casing pressure curves of the well killing,with an average error of 7.5%,proving that the model calculation accuracy is high.The influence of temperature and pressure,geothermal gradient,drilling fluid rheological changes,etc.on the change of gas cap position during gas invasion in managed pressure drilling was analyzed.5.The surface comprehensive logging time series data of 16 overflow wells were collected.Based on the adversarial generative network method,using 1min,3min,5min,and 10 min as the time length of the input parameters,4 types of overflow identification networks were trained.using integration The algorithm integrates four types of networks and establishes an early overflow identification method.Based on the established model,two overflow wells were monitored.The results showed that the method in this paper can identify the occurrence of overflow in a short time,and the false alarm rate of the model is low.6.A new overflow evaluation index—overflow formation energy—is proposed.Based on the calculation method of fluid flow work,taking into account factors such as the type of overflow fluid(oil,gas,water),energy loss of fluid in formation seepage,bottom hole pressure difference,a calculation method for overflow formation energy was established.The overflow formation energy of six overflow wells was calculated,and the relationship between the overflow formation energy and well killing methods was initially obtained.