Petroleum Chemistry最新文献

Reservoir fluid properties are important data in the calculation of many aspects of production and reservoir engineering. These properties are critical for efficient reservoir management throughout the life of the reservoir, from discovery to abandonment. Basically, the sequence followed in fluid modeling begins with collecting the samples from the reservoir, analyzing the samples and then developing the mathematical models that describe the thermodynamic behavior of the fluid. This study is interested in estimating the physical properties and prediction of the phase behavior of a reservoir fluid. The prediction was done by using Winprop of CMG® software which has a regression technique to tune the pressure-volume-temperature (PVT) data and phase behavior of the reservoir fluid. Two types of cubic equation of state (EOS) were used in this study. The results of this study showed that Peng‒Robinson (PR) equation was more accurate than Soave‒Redlich‒Kwong (SRK) equation in predicting the phase behavior of the fluid although the calibration process by SRK model was better in most of the experiments than PR model by observing the extent of convergence between the real (experimental) data and the data obtained from the simulation.

Supervised machine learning is one of the important tools that has helped solve many problems facing humanity, especially problems that cannot be solved by humans. Building a successful and high-accuracy model depends on several things, such as the collected data, choosing the appropriate model, the method of data splitting to be used in training and evaluating the model, and choosing the appropriate hyperparameters. Data splitting is one of the most important things to do to obtain a high-accuracy model and to avoid overfitting, which produces a model with high training accuracy but fails in testing and prediction. This paper investigates the impact of different data splitting strategies such as hold-out with different testing sizes, K-Fold, and shuffle split on the effectiveness of a supervised machine learning model for prediction drilling rate in Rumaila oil field in southern Iraq and selecting the optimal data splitting strategy. The highest testing accuracy obtained was 0.827 when the shuffle split strategy was used.

Machine learning and artificial intelligence are recently used in many engineering sectors. Artificial neural network (ANN) has been widely used in oil and gas to predict many important parameters. This work uses ANN to predict the required surface pump pressure at the surface, considering the impact of different drilling parameters. These parameters are: depth, rate of penetration (ROP), weight on bit (WOB), rotation per minute (RPM), stroke per minute (SPM), mud weight, and mud flow rate. ANN models were built using two layers, and both hyperbolic Tanh and Log sigmoid transfer functions were used to predict the model’s validity. Around 2020 data values were used to test, train and validate model prediction. Sensitivity analysis used 2, 4, 8, and 10 neurons for each transfer function (Log sigmoid and hyperbolic Tanh). Results indicated that the prediction for the eight nodes Tanh model best matches the overall data available for the test. For instance, a 99.67% R for training, 99.45% test, 98.57% validation, and 99.47% overall data set were obtained. On the other hand, using a Log model with ten nodes offered the best data set matching for the same data tested above. Results show that test data converged 99.58 with the model prediction method, while 99.52 and 98.95 were obtained for training and validation, respectively. Therefore, we suggest a new model based on the Log model to predict surface pump pressure. This model would be beneficial for predicting the required number and size of pumps at any drilling site.

Lost circulation is a major issue that increases the cost of petroleum exploration operations. During the well planning period, consideration of the degree of severity of mud loss may lead to significant technical and financial benefits. This will assist in the prevention of losses by putting preventative measures in place before running into lost circulation region. This study aimed to predict the amount of mud loss rate (MLR) by using new models with artificial neural networks (ANNs). This model was built in order to obtain a knowledge of the relationship between the amount of loss and the drilling parameters that can be controlled, such as (the rate of penetration (ROP), flow rate (FLW), standpipe pressure (SPP), weight on bit (WOB), nozzle area (TFA), rotation per minute (RPM), and torque (TRQ)), the drilling fluid properties and geomechanical properties like unconfined compressive strength (UCS). Gaining information about UCS along the wellbore is essential for dealing with drilling problems like lost circulation. The new model was developed using a dataset of 209 loss events that were collected from 21 oil wells in the Rumaila oil field’s Dammam and Hartha formations that encountered loss circulation events. Apart from other controllable drilling parameters, it was demonstrated that the rate of losses was also sensitive to UCS values. The amount of mud losses rate constantly rises with increasing UCS. The suggested artificial neural networks (ANN) model was employed to forecast the rate of losses for 21 wells. A comparison plot depicting the actual rate of lost circulation versus the predicted rate was generated as a function of depth. The results indicate that the new model is able to precisely forecast the lost circulation function of controllable drilling variables, drilling mud properties, and UCS with a correlation coefficient of 0.995.

This paper presents an integrated approach to investigate the impact of well spacing on the performance of tight oil and gas reservoirs depleted by advanced multi-stage hydraulic fracture in horizontal wellbores distributed parallelly in rectangular drainage areas. Thus, an analytical models has been developed considering different reservoir configurations and the associated flow rate, cumulative production, ultimate recovery has been recorded. Also, the pressure behavior for early stages of production conditions and the flow regime dominated by boundary effects has been analyzed. The results indicate that well spacing significantly impacts reservoir performance, particularly at late production stages and that well interference and an ee of well spacing optimization in tight oil and gas reservoirs, offering valuable insights for the strategic planning and development of these vital resources.

Produced water re-injection (PWRI) is one of the most important management methods to dispose fluid associated with oil and natural gas production because it is economic and environmentally friendly method. However, several formation damage mechanisms are associated with the re-injection, and the most important damage is clogging of pore throats phenomenon due to the suspended particles, which will reduce the permeability of formation. Therefore, it is necessary to build mathematical model to predict the growth and extend of the formation damage in reservoir to help surveillance PWRI operations. Empirical correlation is one of the Algebraic formation damage models. Empirical model becomes important when there is a loss in detailed information about process of damage creation. In this paper, an empirical model was built to describe growth and extent of the damage in Zubair formation caused by the operation of unfiltered produced water re-injection (UPWRI) in North-Rumaila oilfield under matrix condition by using nonlinear regression tools. From growth and extent of the damage, we conclude that reinjected unfiltered produce water under matrix pressure led to large damage in Main Pay formation near and round wellbore injector because of increased volume of produced water injection. In matrix injection, the quality of the injected water must meet stringent requirements. The damage is start from maximum value (75%) at near wellbore and gradually decreases away from it. The damage zone expands symmetrical around axial wellbore injector formed circular dish that increase-by-increase flow rate and duration injection.

This paper includes a finite element analysis study to explore the stress, strain and of a 16-inch trunk line in South Rumaila oil field, Iraq. CAESAR II software was utilized to simulate the expansion loop 3D model. The loop was prepared in the middle of a distance of 300 000 mm. Data such as pipe and oil specifications were used as input parameters for the software. Results revealed that the allowable stress values for both loops in sustainable and expansion load cases were equal to 57 and 12.7% respectively, which is an appropriate indicator to utilize this elevated loop as a replacement for the normal one. The displacement in the first half of the elevated expansion loop was higher by 6% compared to the displacement of the normal expansion loop due to the generated torsion. Also, the force components FX (force in the flow direction) and FZ (force vertical to the flow direction in the horizontal plane) on the supports were not recorded at the normal expansion loop, whereas these values were then increased up to 157 and 240% respectively at the middle of the loops’ legs. These values decreased by 46 and 85% respectively, at the center nodes of the elevated expansion loop. Results revealed that the elevated expansion loop can absorb stress, strain, and displacement values effectively and this can reduce the required area for constructing the pipelines.

Changes in temperature, pressure, and/or oil composition resulting from mixing with other crude oils or gas injection often affect the solubility of asphaltenes in crude oils. This might lead to the precipitation and deposition of asphaltene, permeability reduction, the obstructing of wells and other surface infrastructure, and eventually a reduction or stoppage of production, which has a considerable economic impact. Therefore, it is essential for both upstream and downstream processing to be able to understand and anticipate asphaltene phase behaviour in order to implement the correct preventative and remedial solutions. To forecast and simulate the precipitation of asphaltene, one of two theories is used: the solubility theory or the colloidal theory. In this study, the former one was applied by using cubic-plus-association equation of state (CPA EOS) to predict the asphaltene phase envelope and determine the precipitation zones for different concentrations of asphaltene of an Iraqi live oil using Multiflash software and Python depending on real field data. The results showed that the zone of precipitation becomes smaller with increasing asphaltene concentration, at which the largest area was at the lowest concentration of 0.04 (as a weight ratio of asphaltene/oil), and then it decreased little by little until it reached the smallest area at the bubble point pressure curve (at the highest concentration of 0.32). This confirms the effect and force of the large asphaltene precipitation in light oils, i.e., the low concentration of asphaltene. Also, the highest concentration of asphaltene precipitation occurs at the bubble pressure point.

For water-wet porous media, the literature revealed a poor displacement efficiency for CO₂ relative to N₂. The overall average residual water saturation displaced by CO₂ is ~50.0% compared to ~20.0% displaced by N₂. Furthermore, based on the “Land” trapping model (Land, 1968), the non-wetting phase trapped during a subsequent imbibition displacement would also be reduced due to the low end-point saturation of the non-wetting phase achieved during the drainage flood. In this study, we hypothesize that for a drainage flood with a very low viscosity ratio (μ_displacing/μ_displaced <<1) and low flow rate (c_a < 10^–6) (i.e., conditions that are typical for the displacement of water by CO₂, N₂ in a strongly water-wet porous media) the end-point residual water saturation is predominantly controlled by the interfacial tension of the fluid-gas system. To test our hypothesis, we have performed six pore-scale displacement experiments on a micromodel using both CO₂ and N₂ to understand the influence of different fingering mechanisms (i.e., capillary vs. viscous) on flooding performance. It is observed that, for capillary-dominated floods, IFT values control the displacement efficiency. Therefore, we could conclude for capillary experiments that the poor displacement of the non-wetting (i.e., CO₂) is due to a snap-off model which is closely related to the IFT value.

The unconfined compressive strength (UCS) is a crucial factor of rock strength parameters for estimating the in situ stresses of the rock, designing the most effective fracture design, predicting the best mud weight, and mitigating drilling issues. UCS is commonly determined by subjecting rock samples to uniaxial or triaxial strains until they fail. Laboratory tests provide a direct and more precise estimation of UCS. However, it is unable to generate a continuous profile along the well (i.e., limited to specific depth intervals) due to the presence of specimens, expense, and time consumption. Consequently, other approaches were devised to overcome the gaps in the UCS prediction by utilizing wire-line log data. Several empirical correlations for predicting UCS are derived from well-log data, particularly the porosity, density, and sonic logs. In this paper, the previous correlations for predicting the UCS of carbonate formation have been evaluated using measured data of UCS. The results show that the compressional wave velocity (VP) is the best well log parameter for estimating carbonate formation’s unconfined compressive strength, and Yasar and Erdogan correlation best predicts the UCS that fit the measured data for carbonate formations. Thus, Yasar and Erdogan correlation has been chosen to estimate a continuous profile of UCS across the entire depth of carbonate formation in the Rumaila oil field.