Ilsik Jang

A new automated scheme of quadrilateral mesh generation for randomly distributed line constraints

This paper presents a new automated method of quadrilateral meshes with random line constraints which have not been fully considered in previous models. The authors developed a new looping scheme and a direct quadrilateral forming algorithm based on advanced front techniques. This generator overcomes the limitations of previous studies such as line constraint, unmeshed hole and mesh refinement. A qualitative test reveals that our algorithm is reliable and suitable at the field needed for very accurate results. The developed direct method to handle line-typed features automatically makes the multiple discretizations without any user interaction and modification.

Development of Pareto-based evolutionary model integrated with dynamic goal programming and successive linear objective reduction

This paper presents a new Pareto-based evolutionary model incorporated with preference-ordering and objective-dimension reduction to improve the multi-directional searches for multi-objective problems.It induces a convergence toward the Pareto-optimal front by adjusting aspiration levels allocated to objectives and by excluding redundant objectives during optimization.Its usefulness was validated for multi-objective test problems comparing to conventional single- and multi-objective optimization models. This study investigates the coupling effects of objective-reduction and preference-ordering schemes on the search efficiency in the evolutionary process of multi-objective optimization. The difficulty in solving a many-objective problem increases with the number of conflicting objectives. Degenerated objective space can enhance the multi-directional search toward the multi-dimensional Pareto-optimal front by eliminating redundant objectives, but it is difficult to capture the true Pareto-relation among objectives in the non-optimal solution domain. Successive linear objective-reduction for the dimensionality-reduction and dynamic goal programming for preference-ordering are developed individually and combined with a multi-objective genetic algorithm in order to reflect the aspiration levels for the essential objectives adaptively during optimization. The performance of the proposed framework is demonstrated in redundant and non-redundant benchmark test problems. The preference-ordering approach induces the non-dominated solutions near the front despite enduring a small loss in diversity of the solutions. The induced solutions facilitate a degeneration of the Pareto-optimal front using successive linear objective-reduction, which updates the set of essential objectives by excluding non-conflicting objectives from the set of total objectives based on a principal component analysis. Salient issues related to real-world problems are discussed based on the results of an oil-field application.

Optimal Well Placement Based on Artificial Neural Network Incorporating the Productivity Potential

Abstract This article presents an efficient approach to determine the optimal drilling location for maximizing the cumulative production without the need for a reservoir simulation, of which scheme is based on artificial neural network incorporating the productivity potential. A reservoir simulator can provide an accurate result, but is sometimes inefficient due to the enormous computing requirements. The typical artificial neural network scheme used in multiwell placement shows lower predictability as the size of the input data increases. This work introduces the productivity potential that merges various reservoir properties, such as the permeability, porosity, and saturation, and integrates it into an artificial neural network. The cumulative production is compared with the result of the reservoir simulator to determine the accuracy of the developed method. The efficiency of the conventional artificial neural network is improved by the proposed model, as well by using the productivity potential instead of a lot of separate inputs. The predictability is verified by determining the drilling location in the same way as that of the reservoir simulator in the case of a single infill well. The stability is confirmed by its ability to produce a reliable result even as the number of input data increases.

Development of a robust multi-objective history matching for reliable well-based production forecasts

This article presents a dynamic reservoir characterization using a new multi-objective optimization algorithm to quantify the reservoir uncertainties in history matching. The proposed method formulated Pareto-optimality with preference-ordering to derive multiple trade-off history-matched reservoir models for probabilistic production estimation. The integration of linear programming with multi-objective genetic algorithm enhances the efficiency of a multi-directional search by prioritizing the reservoir models that satisfy the aspiration levels on the discrepancy between the observed and the calculated production data. The preference levels are automatically adjusted in correspondence to the quality of the reservoir models for facilitating the model update process during optimization. An oil-field application result indicates the method outperforms the conventional multi-objective optimization method in terms of the relative average error for the production data despite a small loss of diversity-preservation among the reservoir models.

Numerical Analysis of Diffusion in Discrete Fracture Networks with Fractal Geometry by Using Pressure Transient Data

This article presents numerical analyses on diffusion trend for the fractured media with fractal properties by pressure transient data. The authors develop a discrete fracture model with random fracture pattern and a power-law length distribution, which has not been fully considered in the previous fractal modeling. From the model presented, the fractal dimension and conductivity index is determined, and the pressure behavior is analyzed for the range of length exponent (a) of power-law from 1 and 3. Various fractal geometries are examined depending on the morphology of fractures. A single fractal characteristic is observed in the media dominated by long (a < 2) or small fractures (a > 3), and two distinct fractal dimensions in pertinently mixed system of long and small fractures (2 < a < 3). Also, pressure transient behaviors in the mixed system show their unique characteristics, corresponding to each fractal dimension respectively. It is shown that diffusion effect is relatively small in the early time influenced by a few long fractures, and becomes larger in the late time by long and small fractures.

Multi-Objective History Matching Allowing for Scale-Difference and the Interwell Complication

This study presents a new multi-objective history matching model to predict the individual well performance. Typical single-objective history matching, reducing the linearly averaged form of different-scaled objectives, has not covered the individual well performance properly. Previous multi-objective history matching, which could demonstrate the individual performance, shows the poor applicability as the number of objective function increases. This work aims to develop the accurate and diversity-preserved methodology to accomplish the global optimization. The scheme consists of dynamic goal programming and successive linear objective reduction incorporated with non-dominated sorting genetic algorithm-II. Dynamic goal programming grants priorities to solutions satisfying the expectation values for the objective functions with goal adjustment. SLOR removes redundant objective functions at the fitness evaluation in genetic algorithm. For the case study of waterflood history matching, the model is less sensitive to the form of objective functions and gridblock size. This study proves that reflecting relativity of different performances is able to improve prediction ability of the conventional single- and multi-objective approaches. The model provides a reliable range of uncertainty from diversity-preserved concept. The developed multi-objective optimization algorithm can easily apply to solve the convergence problem and the unrealistic estimation caused by scale-difference and the complication among multi-objective functions.

Inverse Fracture Model Integrating Fracture Statistics and Well-testing Data

Abstract Heterogeneity and poor connectivity of fractures make it difficult to characterize fracture networks and predict flow behavior on them. Previous studies have introduced inversion models integrating the observed pressure data to describe flow patterns. However, they could not consider the statistical properties of fractures because the models are based on regular lattice or continuum approaches. A new inverse fracture model, which simultaneously integrates fracture characteristics, fluid flow, and solute transport data, is proposed. Discretization for the fracture-occurrence points makes it possible to incorporate fracture properties in the inversion. Fluid flow is implemented by the cubic law, and a semi-analytical method is used to include the solute transport data due to its efficient performance. The model can interpret the characteristics of the geometry and conductivity between wells within fractured reservoirs, since the inverse fracture network not only has the same fracture characteristics as observed data, but also reproduces the fluid flow and solute transport data. It is demonstrated that the fracture network that is developed makes responses for additional flow and transport predicted within a reasonable error range.

Hydraulic-Unit-Based Fuzzy Model to Predict Permeability from Well Logs and Core Data of a Multi-Layer Sandstone Reservoir in Ulleung Basin, South Korea:

This paper presents a method to predict permeability of an offset well using well logs and core data from a multi-layer sandstone reservoir with various depositional environments. Many studies have been conducted to predict permeability but limited to wells drilled in a formation with a single depositional environment. This study implements fuzzy models to predict the FZI (flow zone indicator) and permeability, and applied HU (hydraulic unit) grouping to classify the rock properties of various depositional environments in terms of the correlation between permeability and porosity. Then, an individual fuzzy model is developed for each HU group. A field application confirms that the method can be applied to permeability prediction using well data from various depositional environments. A comparative study shows that HU grouping plays a key role in grouping cores of similar flow characteristics in the case of core data from multiple layers with various depositional environments.

Reservoir Heterogeneity Affecting Steam Communication between Multiple Well-Pairs for Steam Assisted Gravity Drainage

This paper investigates the effects of reservoir heterogeneity on steam communication and well-interference between multiple steam chambers during SAGD (steam assisted gravity drainage) process. The conventional steam stimulation for developing oil sands uses well pad system composed of several well pairs, so that the heat interference occurs when the steam chambers merge with each other. The numerical simulations using multiple SAGD well pairs were conducted in a heterogeneous formation and compared with those of a homogeneous case in terms of the production performances. Reservoir heterogeneity could make uneven steam communication, unequal chamber growth and leave unrecovered area, thereby giving the negative effect on oil production. Both production profiles of water and oil, and cSOR (cumulative steam to oil ratio) of individual SAGD pair would be good indicators to diagnose steam movement between chambers. The results discussed the negative effect of reservoir heterogeneity making unequal growth of vapor chamber and energy inefficiency due to horizontal movement of injected steam, and the necessity of identifying the indicators to characterize the chamber growth.

Well-placement optimisation using sequential artificial neural networks

In this study, a new algorithm is proposed by employing artificial neural networks in a sequential manner, termed the sequential artificial neural network, to obtain a global solution for optimizing the drilling location of oil or gas reservoirs. The developed sequential artificial neural network is used to successively narrow the search space to efficiently obtain the global solution. When training each artificial neural network, pre-defined amount of data within the new search space are added to the training dataset to improve the estimation performance. When the size of the search space meets a stopping criterion, reservoir simulations are performed for data in the search space, and a global solution is determined among the simulation results. The proposed method was applied to optimise a horizontal well placement in a coalbed methane reservoir. The results show a superior performance in optimisation while significantly reducing the number of simulations compared to the particle-swarm optimisation algorithm.