• Establishment of a creep constitutive model for water-bearing mudstone. • The FEA model of "casing-cement-mudstone" system is constructed to reveal the influence mechanism of casing deformation. • Full-scale test for casing service performance evaluation is carried out and the calculation model of string strength is verified. • A prediction model of the service life of casing in mudstone layer is established based on RFR algorithm. Due to the creep behavior of water-bearing mudstone, the casing of mudstone layer is easy to deform under the action of uneven crustal stress, resulting in the decrease of oil and gas well production efficiency and the increase of development cost. Based on the creep test data of water-bearing mudstone in DQ oilfield of China, the creep constitutive model of mudstone is constructed. On this basis, the influence laws of mudstone moisture content, in-situ stress state and service time on typical in-service casing deformation are comprehensively analyzed. Based on the theoretical model and full-scale tests, the burst strength and collapse strength of the deformed casing are obtained, and the safety state of the casing is evaluated by the safety factor method. Finally, a prediction process of casing deformation in mudstone layer is proposed, and a machine learning method is introduced to intelligently predict the remaining life of casing under the influence of various factors. Through optimal comparison, a prediction model of the service life of casing in water-bearing mudstone layer is established based on random forest regression (RFR) algorithm, and the accuracy of the model is 0.901, which provides a potential solution to the prediction of service performance change caused by casing deformation in mudstone layer.
Yang et al. (Fri,) studied this question.