Accurate identification of lithology in petroleum engineering is very important for oil and gas reservoir evaluation, drilling decisions, and petroleum geological exploration. Using a cross-plot to identify lithology only considers two logging parameters, causing the accuracy of lithology identification to be insufficient. With the continuous development of artificial intelligence technology, machine learning has become an important means to identify lithology. In this study, the cutting logging data of the Junggar Basin were collected as lithologic samples, and the identification of argillaceous siltstone, mudstone, gravel mudstone, silty mudstone, and siltstone was established by logging and logging parameters at corresponding depths. Aiming at the non-equilibrium problem of lithologic data, this paper proposes using equilibrium accuracy to evaluate the model. In this study, manifold learning is used to reduce logging and logging parameters to three dimensions. Based on balance accuracy, four dimensionality reductions including isometric feature mapping (ISOMAP), principal component (PCA), independent component (ICA), and non-negative matrix factorization (NMF) are compared. It is found that ISOMAP improves the balance accuracy of the LightGBM model to 0.829, which can effectively deal with unbalanced lithologic data. In addition, the particle swarm optimization (PSO) algorithm is used to automatically optimize the super-parameters of the lightweight gradient hoist (LightGBM) model, which effectively improves the balance accuracy and generalization ability of the lithology identification model and provides strong support for fast and accurate lithology identification.

During the drilling of natural gas hydrate reservoirs, gas invasion may occur, which has an adverse impact on the upward return process of cuttings and the control of bottom-hole pressure. A multiphase fluid CFD simulation method based on the Eulerian model was optimized. With the help of Fluent 23R1 software, simulations of the three-phase flow of drilling fluid–cuttings–invaded gas in the vertical section, build-up section, and horizontal section of the complex annulus multiphase flow in the hydrate drilling wellbore based on particle dynamics were carried out. The characteristics of multiphase flow in the wellbore after gas invasion were revealed, and the impacts of gas invasion on the deposition and migration of cuttings and the control of bottom-hole pressure were analyzed. The research shows that after gas invades the wellbore, it has little impact on the upward return of cuttings in the build-up section. However, in the vertical and horizontal sections, with the increase in the flow rate of the invaded gas, the slip and upward return of cuttings are promoted. The gas invasion velocity has a more significant impact on the cutting migration in different well sections compared to the gas invasion volume fraction. A faster invasion velocity accelerates the upward return of cuttings in the wellbore annulus. When gas invasion occurs, it has little impact on the pressure in the build-up section, decreases the pressure gradient in the vertical section, and increases the pressure gradient in the horizontal section. At the same time, there is a high linear correlation between the pressure gradient in each well section and the density of water-based drilling fluid. By reasonably controlling the density of drilling fluid according to the gas volume fraction detected at the wellhead, the aggravation of gas invasion can be significantly prevented. The research methods and calculation results are helpful in providing a reference for the subsequent drilling process of hydrate reservoirs.

The diagnosis of overflow risk has always been an important area of research in drilling operations in the field of oil and gas engineering. In the face of the limitations and lag of traditional overflow diagnosis methods, the practical application effect of existing models and methods is not obvious, and there is no integration of the physical model and the intelligent algorithm model for overflow diagnosis, this paper proposes a method of adaptive weight fusion of physical model and intelligent algorithm model diagnosis results. Based on the fusion of the physical model and the intelligent algorithm model, an overflow diagnosis model of managed pressure drilling is established. The research results show that the fusion model in this paper can combine the accuracy weight of the physical model and the intelligent algorithm model for the intelligent diagnosis of overflow risk, which improves the mechanization and interpretability of the model and diagnosis results while ensuring accuracy and efficiency. And the intelligent algorithm model used in the fusion model is superior to other algorithm models. The overflow diagnosis accuracy of the fusion model on the test set samples reaches more than 98%, and the accuracy of the validation set is 94.25%. The content of this study provides guidance for drilling overflow diagnosis and model fusion methods.