Automatic well test interpretation based on convolutional neural network for a radial composite reservoir

An automatic well test interpretation method for radial composite reservoirs based on convolutional neural network(CNN) is proposed, and its effectiveness and accuracy are verified by actual field data. In this paper, based on the data transformed by logarithm function and the loss function of mean square error(MSE), the optimal CNN is obtained by reducing the loss function to optimize the network with "dropout" method to avoid over fitting. The trained optimal network can be directly used to interpret the buildup or drawdown pressure data of the well in the radial composite reservoir, that is, the log-log plot of the given measured pressure variation and its derivative data are input into the network, the outputs are corresponding reservoir parameters(mobility ratio, storativity ratio, dimensionless composite radius, and dimensionless group characterizing well storage and skin effects), which realizes the automatic initial fitting of well test interpretation parameters. The method is verified with field measured data of Daqing Oilfield. The research shows that the method has high interpretation accuracy, and it is superior to the analytical method and the least square method.

Study on characteristics of well-test type curves for composite reservoir with sealing faults

The pressure response for the composite reservoirs with a sealing fault locating in inner and outer region is different,which neglected by previous researchers,would cause significant errors during well-test interpretations.Based on seepage theory,a well-test model of two-region radial composite reservoir with infinite outer boundary has been built in this study considering wellbore storage and skin effects.The solutions for this model and characteristics of the type curves have been analyzed by applying the method of mirror image,Laplace transformation and superposition principle,including a straight fault,a perpendicular fault and parallel faults cases.The study shows that the dimensionless pressure derivative curves would be obviously different in two cases:the well to fault distance is larger,and smaller than the half length of the inner-region radius.Therefore,type curves are presented with reasonable parameters to analyze the distance effect on the dynamic pressure response.The results in this study are of great significance。

Application of Heterogeneous Composite Model with Consideration of Stress Sensitive in Low Permeability Gas Reservoir

Stress sensitivity is a key factor affecting the productivity of single wells in low permeability gas reservoirs. A well test model for heterogeneous composite gas reservoirs under the influence of stress-sensitive effects was established. Based on the theoretical model, the well test was designed by gradually increasing the pressure difference. The relationship between abnormal high pressure and reservoir stress sensitivity was analyzed. Theoretical research shows that stress sensitivity will cause permeability damage during the production process, and the pressure drop test curve shows that the physical properties of the reservoir have gradually deteriorated. The pressure recovery test curve shows that the physical properties of the reservoir are getting better. Field practice shows that stress sensitivity is related to the formation of abnormally high pressure in the formation without considering the micro-cracks in the formation. Stress-sensitive reservoirs are generally unbalanced and compacted due to overpressure, for fluid expansion/conduction overpressure in Ledong Area. For these reservoirs, there is almost no stress sensitivity. The research results have significance for guiding the design and data interpretation of stress-sensitive reservoir.

Investigation of the influences of asphaltene deposition on oilfield development using reservoir simulation

This paper investigates the deposition of asphaltenes in the porous medium of the studied field in Russia and predicts production profiles based on uncertainty evaluation. This problem can be solved by dynamic modeling, during which production profiles are estimated in two scenarios: with and without the activation of the asphaltene option. Calculations are carried out for two development scenarios: field operation under natural depletion and water injection into the aquifer as a reservoir pressure maintenance system. A full-scale compositional reservoir simulation model of the Russian oilfield was created. Within a dynamic simulation, the asphaltene option was activated and the asphaltene behavior in oil and porous medium was tuned according to our own special laboratory experiments. The model was also matched to production historical data, and a pattern model was prepared using the full-scale simulation model. Technological and the asphaltene option parameters were used in sensitivity and an uncertainty evaluation. Furthermore, probable production profiles within a forecast period were estimated. The sensitivity analysis of the pattern model to input parameters of the asphaltene option allowed determining the following heavy-hitters on the objective function: the molar weight of dissolved asphaltenes as a function of pressure, the asphaltene dissociation rate, the asphaltene adsorption coefficient and the critical velocity of oil movement in the reservoir. Under the natural depletion scenario, our simulations show a significant decrease in reservoir pressure and the formation of drawdown cones leading to asphaltene deposition in the bottom-hole area of production wells, decreasing their productivity. Water injection generally allows us to significantly reduce the volume of asphaltene phase transitions and has a positive effect on cumulative oil production. Injecting water into aquifer can keep the formation pressure long above the pressure for asphaltene precipitation, preventing the asphaltene deposition resulted from interaction of oil and water, so this way has higher oil production.

Interpretation of Temperature Profiles during Soak Periods in Steam-stimulated Wells

To address the problems existing in testing steam injection profiles in a steam-stimulated well during steam injection and production periods, this paper proposes that the temperature profile in the completion interval could be tested during the soak period. A mathematical model for calculating the vertical distribution of temperature in a single layer reservoir is established based on the temperature characteristics of steam stimulated reservoirs, and the vertical distribution of temperature in a single layer reservoir could be obtained and heat loss could be calculated. The temperature, which is disturbed by thermal conduction in a multilayer reservoir, and heat loss could be derived based on the superposition principle of temperature potential. This paper establishes a multilayer testing temperature profile interpretation method and interprets the actual test temperature profile of Well Gao 3-7-66. The results indicate that the temperature profile in the soak period can reflect the thermal absorption conditions in various reservoir beds.