Application of the big data method to search for the main factors controlling the hole collapse in the production process of Shunbei carbonate reservoir

Shunbei Oilfield is a deep carbonate oilfield in the Tarim Basin, whose reservoir coring is difficult and the cores are poorly representative due to their strong heterogeneity. At present, the borehole stability is studied by means of laboratory rock samples or numerical simulation based on a single factor or several confirmed factors. However, the success rate of controlling the hole collapse according to the research results is not high. The idea of big data was introduced to search for the main control factors based on field data, so as to solve the problems of no or few cores and take into consideration the influential factors as much as possible. 850 thousand original data in the operation links (e.g. reservoir drilling, well flushing, testing, acid fracturing and production test) of 17 wells in No. 1 and No. 5 Shunbei belt were sorted out. And in order to ensure the physical significance of the data, the multiple regression fitting of independent variable was carried out by taking the hole enlargement rate as the target parameter and the actual data of drilled reservoirs (including geological, engineering and fluid data) as the independent variables. Then, non-main control factors were removed by means of element reduction method, and the artificially controllable factors were ranked from the top to the bottom, i.e., Φ₆₀₀, drilling time of target layer, Φ₃₀₀, weight on bit while drilling the target layer, Φ₂₀₀, Y-axis, Φ₁₀₀, funnel viscosity, X-axis, Φ₆, drill pump pressure, hole deviation angle and solid content. Finally, it was inferred by using the model of drilling process that Well SHB1-4H whose tubing was blocked due to hole collapse in the process of production had the hole enlargement rate of 3.22%. In addition, the reasons for the blockage in the process of production were interpreted. In conclusion, introducing the idea of big data to search for the main factors controlling the borehole stability in deep layers while no test samples is available for prediction provides a new method for controlling the borehole stability in the development process of deep oil and gas.