Iranian Journal of Oil and Gas Science and Technology

Abstract Gas lift is a principal technique used for artificial lifting and Enhanced Oil Recovery (EOR) method, facilitates oil flow in mature fields those in the latter half of their productive life by injecting high-pressure gas into the wellbore to reduce column density. However, the thermodynamics of gas injection in wells requiring high differential pressures introduce severe flow assurance challenges. The intense Joule-Thomson cooling effect across injection chokes is the primary driver of gas hydrate formation. In the studied field (Aghajari), the current hardware-based mitigation strategy employs Thermal Chokes, which utilize the enthalpy of live crude oil to heat the injection gas. Despite this, operational evidence indicates that during cold seasons and for wells with high pressure drops, this system proves inefficient, leading to freezing in injection lines and flow interruption. In the absence of inhibitor injection systems, operators are compelled to resort to reactive measures such as flaring injection gas to induce pressure shocks and clear blockages. This vicious cycle not only results in capital loss but also leads to production deferment and excessive workload for human resources. This research aims to propose a proactive process-based solution by synergizing data mining and computational intelligence. Through the analysis of 5,960 operational records from 101 wells (extracted from the WIMS system), an Artificial Neural Network (ANN) model was developed to serve as a virtual sensor, predicting gas thermodynamic behavior and post-choke temperature with 98.5% accuracy. The core novelty of this study lies in the simulation and validation of a dual-stage pressure reduction strategy. Results demonstrate that splitting the pressure drop profile reduces cooling intensity by up to 60%, maintaining the fluid outside the hydrate stability zone throughout the expansion path. This approach enhances safety and production stability while eliminating the need for costly physical interventions.

Abstract The spatial distribution of petrophysical properties within the reservoirs is one of the most important factors in reservoir characterization. Flow units are the continuous body over a specific reservoir volume within which the geological and petrophysical properties are the same. Accordingly, an accurate prediction of flow units is a major task to achieve a reliable petrophysical description of a reservoir. The aim of this paper was core flow unit determination by using a new intelligent method. Flow units were determined and clustered at specific depths of reservoir by using a combination of artificial neural network (ANN) and a metaheuristic optimization algorithm method. At first, artificial neural network (ANN) was used to determine flow units from well log data. Then, imperialist competitive algorithm (ICA) was employed to obtain the optimal contribution of ANN for a better flow unit prediction and clustering. Available routine core and well log data from a well in one of the Iranian oil fields were used for this determination. The data preprocessing was applied for data normalization and data filtering before these approaches. The results showed that imperialist competitive algorithm (ICA), as a useful optimization method for reservoir characterization, had a better performance in flow zone index (FZI) clustering compared with the conventional K-means clustering method. The results also showed that ICA optimized the artificial neural network (ANN) and improved the disadvantages of gradient-based back propagation algorithm for a better flow unit determination.

Abstract In this work, artificial neural network (ANN) has been employed to propose a practical model for predicting the surface tension of multi-component mixtures. In order to develop a reliable model based on the ANN, a comprehensive experimental data set including 15 ternary liquid mixtures at different temperatures was employed. These systems consist of 777 data points generally containing hydrocarbon components. The ANN model has been developed as a function of temperature, critical properties, and acentric factor of the mixture according to conventional corresponding-state models. 80% of the data points were employed for training ANN and the remaining data were utilized for testing the generated model. The average absolute relative deviations (AARD%) of the model for the training set, the testing set, and the total data points were obtained 1.69, 1.86, and 1.72 respectively. Comparing the results with Flory theory, Brok-Bird equation, and group contribution theory has proved the high prediction capability of the attained model.

Abstract Porosity is one of the fundamental petrophysical properties that should be evaluated for hydrocarbon bearing reservoirs. It is a vital factor in precise understanding of reservoir quality in a hydrocarbon field. Log data are exceedingly crucial information in petroleum industries, for many of hydrocarbon parameters are obtained by virtue of petrophysical data. There are three main petrophysical logging tools for the determination of porosity, namely neutron, density, and sonic well logs. Porosity can be determined by the use of each of these tools; however, a precise analysis requires a complete set of these tools. Log sets are commonly either incomplete or unreliable for many reasons (i.e. incomplete logging, measurement errors, and loss of data owing to unsuitable data storage). To overcome this drawback, in this study several intelligent systems such as fuzzy logic (FL), neural network (NN), and support vector machine are used to predict synthesized petrophysical logs including neutron, density, and sonic. To accomplish this, the petrophysical well logs data were collected from a real reservoir in one of Iran southwest oil fields. The corresponding correlation was obtained through the comparison of synthesized log values with real log values. The results showed that all intelligent systems were capable of synthesizing petrophysical well logs, but SVM had better accuracy and could be used as the most reliable method compared to the other techniques.