The determination of microscopic residual oil distribution is beneficial for exploiting reservoirs to their maximum potential. In order to investigate microscopic residual oil during the carbon dioxide (CO₂) huff-and-puff process in tight oil reservoirs, several CO₂ huff-and-puff tests with tight sandstone cores were conducted at various conditions. Then, nuclear magnetic resonance (NMR) was used to determine the microscopic residual oil distribution of the cores. The experiments showed that the oil recovery factor increased from 27.22% to 52.56% when injection pressure increased from 5 MPa to 13 MPa. The oil recovery was unable to be substantially enhanced as the injection pressure further increased beyond the minimum miscible pressure. The lower limit of pore distribution where the oil was recoverable corresponded to relaxation times of 2.68 ms, 1.29 ms, and 0.74 ms at an injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively. Longer soaking time also increased the lower limit of the oil-recoverable pore distribution. However, more cycles had no obvious effect on expanding the interval of oil-recoverable pore distribution. Therefore, higher injection pressure and longer soaking time convert the residual oil in smaller and blind pores into recoverable oil. This investigation provides some technical ideas for oilfields in design development programs for optimizing the production parameters during the CO₂ huff-and-puff process.

Artificial intelligence methods for oil and gas reservoir development: Current progresses and perspectives

Geared toward the problems of predicting the unsteadily changing single oil well production in water flooding reservoir, a machine learning model based on CNN (convolutional neural network) and LSTM (long short-term memory) is established which realizes precise predictions of monthly single-well production. This study is considering more than 60 dynamic and static factors that affect the changes of oil well production, introduce water injection parameters into data set, select 11 main control factors, and then, build a CNN-LSTM model optimized by Bayesian optimization. The effectiveness of the proposed model is verified in a realistic reservoir. The experiment results show that the prediction accuracy of the proposed model is over 90%, which suggests a penitential application in an extensive range of applications. Production forecasting by the developed model is simple, efficient, and accurate, which can provide a guidance for the dynamic analysis of a water flooding reservoir, and work as a good reference of the development and production of other types of reservoirs.

Heavy oil production methods included the cyclic steam stimulation (CSS), steam flooding and steam assistant gravity drainage (SAGD). It is well known, when water invasion occurs in heavy oil reservoirs with edge/bottom water, for a long time, how much of the produced water comes from the steam injection process as oppose to edge/bottom water invasion? Reservoir engineers do not clearly understand the process, because after the steam condenses a part of the condensation comes to the surface and another part is stored within the reservoir itself. This is the problem that still gives reservoir engineers trouble. Through studying the production data of thermal recovery at the Liaohe oilfield, in China, for the first time, this paper reveals a new theory and calculation method that shows how to calculate water invasion from the bottom/edge water reservoir, therefore, a new concept of conversion rate was presented and a new parameter of steam-water ratio was defined, as k. new equations of calculating produced water invasion and water invasion of the remaining formation with bottom/edge water were derived according to the new parameter k. In addition, an estimation method of breakthrough time of bottom/edge is presented in this paper. The methods have been used to calculate and analyze the production performance of heavy oil with steam injection in Liaohe oilfield. The calculation results accord well with actual situation of heat-recovery reservoirs. Therefore, this method will be recommended to optimize production performance parameters of thermal recovery. IntroductionCurrently, there are huge reserves of heavy oil in the world; however the physical properties of crude oil are very poor. Therefore a thermal recovery must be implemented in order to meet heavy oil production due to its high viscosity. CSS has been applied to commercial production of heavy oil reservoirs since the 1960s. During the 20th century, annual oil production reached 20 million tons, and thermal recovery technologies were greatly improved. CSS has become the main development technology for heavy oil development in the world. Performance parameters of CSS surface operation consist of two parts: reservoir parameters and operation parameters. In order to efficiently improve reservoir management, reservoir engineers must understand performance parameters. For many years, reservoir engineers have studied parameters, such as: oil-steam ratio [1-5], production-injection ratio, back water production rate, and water storage rate [6-8] of heavy oilfields using CSS. However, formation water production, cumulative volume of formation water invasion, and formation pressure drop rate are still problems looking for better solutions. This research focuses on how to calculate the parameters in theory. This paper also presents the relationship among these parameters, and provides effective analytic method for performance analysis of heavy oil reservoirs with the aim of improving heavy oil...