Mixed precision parallel preconditioner and linear solver for compositional reservoir simulation

Reservoir simulation role in value creation and strategic management decisions cannot be over emphasized. Simulation of complex challenging reservoirs with millions of grid blocks especially in compositional mode is very time-consuming even with fast modern computers. On the other hand, high price of cluster supercomputers prevents them for being commonly used for fast simulation of such reservoirs. In recent years, the development of many-core processors like cell processors, DSPs, and graphical processing units (GPUs) has provided a very cost-effective hardware platform for fast computational operations. However, programming for such processors is much more difficult than conventional CPUs, and new parallel algorithm design and special parallel implementation methods are needed. Using the computational power of CPUs, GPUs, and/or any other processing unit, Open Computing Language (OpenCL) provides a framework for programming for heterogeneous platforms. In this paper, OpenCL is used to employ the computational power of a GPU to build a preconditioner and solve the linear system arising from compositional formulation of multiphase flow in porous media. The proposed parallel preconditioner is proved to be quite effective, even in heterogeneous porous media. Using data-parallel modules on GPU, the preconditioner/solver runtime reduced at least 1 order of magnitude compared to their serial implementation on CPU.     

Distribution of heavy metals in the clastic fine-grained sediments of Gavkhuni playa lake (Southeast of Isfahan,Iran)

The sedimentary basin of Gavkhuni playa lake includes two sedimentary environments of delta and playa lake. These environments consist of mud, sand and salt flats. There are potentials for concentration of heavy metals in the fine-grained sediments (silt and clay) of the playa due to existence of Pb/Zn ore deposits, industrial and agricultural regions in the water catchment of Zayandehrud River terminating to this area. In order to study the concentration of heavy metals and the controlling factors on their distribution in the fine-grained sediments, 13 samples were taken from the muddy facies and concentration of the heavy metals were determined. The results showed that the heavy metal concentrations range in the sediments (in ppm) are Mn (395.5–1,040), Sr (100.4–725.76), Pb (14.66–91.06), Zn (23.59–80.9), Ni (37–73.66), Cu (13.83–29.83), Co (5.73–13.78), Ag (3.03–4.76) and Cd (2.3–5.5) in their order of abundances. The concentration of Ag is noticeable in the sediments relative to the average concentration of this element in mud sediments. The amounts of Pb and Zn are relatively high in all the samples in comparison with the other elements. The concentration of Ni is relatively high in the oxidized samples. The distribution of Pb is directly related to organic matter content of the sediments. The concentrations of Zn, Sr, Cu, Co and Cd in the samples of the playa are lower than those in the delta. The amount of illite is another factor influencing Zn and Pb concentrations. Sr is more concentrated in the sediments with the high content of calcium carbonate. The distribution pattern of Cu, Co, Pb and Mn resembles to that of the clay content of the sediments. The clay content shows positive correlations with Co, Cu and Mn concentrations and negative correlation with Ag. The Sr and Ag concentrations are positively correlated with the amount of CaCO₃. The amounts of Co, Cu, Ni and Mn show negative correlations with the calcium carbonate content. Pb and Co are noticeably correlated with Mn.     

Optimal water and waste load allocation in reservoir–river systems: a case study

In this paper, a new methodology is developed for optimization of water and waste load allocation in reservoir–river systems considering the existing uncertainties in reservoir inflow, waste loads and water demands. A stochastic dynamic programming (SDP) model is used to optimize reservoir operation considering the inflow uncertainty, and another model called PSO-SA is developed and linked with the SDP model for optimizing water and waste load allocation in downstream river. In the PSO-SA model, a particle swarm optimization technique with a dynamic penalty function for handling the constraints is used to optimize water and waste load allocation policies. Also, a simulated annealing technique is utilized for determining the upper and lower bounds of constraints and objective function considering the existing uncertainties. As the proposed water and waste load allocation model has a considerable run-time, some powerful soft computing techniques, namely, Regression tree Induction (named M5P), fuzzy K-nearest neighbor, Bayesian network, support vector regression and an adaptive neuro-fuzzy inference system, are trained and validated using the results of the proposed methodology to develop real-time water and waste load allocation rules. To examine the efficiency and applicability of the methodology, it is applied to the Dez reservoir–river system in the south-western part of Iran.     

Monitoring above-zone temperature variations associated with CO2 and brine leakage from a storage aquifer

CO₂ injection in saline aquifers induces temperature changes owing to processes such as Joule–Thomson cooling, endothermic water vaporization, exothermic CO₂ dissolution besides the temperature discrepancy between injected and native fluids. CO₂ leaking from the injection zone, in addition to initial temperature contrast due to the geothermal gradient, undergoes similar processes, causing temperature changes in the above zone. Numerical simulation tools were used to evaluate temperature changes associated with CO₂ leakage from the storage aquifer to an above-zone monitoring interval and to assess the monitorability of CO₂ leakage on the basis of temperature data. The impact of both CO₂ and brine leakage on temperature response is considered for three cases (1) a leaky well co-located with the injection well, (2) a leaky well distant from the injector, and (3) a leaky fault. A sensitivity analysis was performed to determine key operational and reservoir parameters that control the temperature signal in the above zone. Throughout the analysis injection-zone parameters remain unchanged. Significant pressure drop upon leakage causes expansion of CO₂ associated with Joule–Thomson cooling. However, brine may begin leaking before CO₂ breakthrough at the leakage pathway, causing heating in the above zone. Thus, unlike the pressure which increases in response to both CO₂ and brine leakage, the temperature signal may differentiate between the leaking fluids. In addition, the strength of the temperature signal correlates with leakage velocity unlike pressure signal whose strength depends on leakage rate. Increasing leakage conduit cross-sectional area increases leakage rate and thus increases pressure change in the above zone. However, it decreases leakage velocity, and therefore, reduces temperature cooling and signal. It is also shown that the leakage-induced temperature change covers a small area around the leakage pathway. Thus, temperature data will be most useful if collected along potential leaky wells and/or wells intersecting potential leaky faults.     

Application of artificial intelligence techniques for predicting the flyrock distance caused by blasting operation

Flyrock arising from blasting operations is one of the crucial and complex problems in mining industry and its prediction plays an important role in the minimization of related hazards. In past years, various empirical methods were developed for the prediction of flyrock distance using statistical analysis techniques, which have very low predictive capacity. Artificial intelligence (AI) techniques are now being used as alternate statistical techniques. In this paper, two predictive models were developed by using AI techniques to predict flyrock distance in Sungun copper mine of Iran. One of the models employed artificial neural network (ANN), and another, fuzzy logic. The results showed that both models were useful and efficient whereas the fuzzy model exhibited high performance than ANN model for predicting flyrock distance. The performance of the models showed that the AI is a good tool for minimizing the uncertainties in the blasting operations.     

Estimation and validation of atmospheric water vapor content using a MODIS NIR band ratio technique based on AIRS water vapor products

Atmospheric water vapor validation needs simultaneous, well-defined, and independent information which are not easily available causing limitations in the development of remote sensing water vapor retrieval algorithms. This study is concerned with the retrieval of total atmospheric water vapor content and its validation. A band ratio method has been used to estimate the water vapor content based on Moderate Resolution Imaging Spectroradiometer (MODIS) Near InfraRed (NIR) data. The method uses MODIS bands 17, 18, and 19 as NIR bands and band 2 to remove the land cover reflectance. Furthermore, the Atmospheric Infrared Sounder (AIRS) has been used for both algorithm development and analysis of the results. The method has been modified to take into account the dry condition of the central parts of Iran. Using some various datasets, the method is implemented and evaluated quantitatively. The validation of the water vapor estimates has been undertaken by an analysis of AIRS data. The validation results shows error as low as 9 % for the estimated water vapor using the MODIS NIR band ratio method.     

Diagenetic model of carbonate rocks of Guri Member of Mishan Formation (Lower to Middle Miocene) SE Zagros basin,Iran

In order to understand the post-depositional history of carbonate rocks of Guri Member (Lower to Middle Miocene), three stratigraphic sections were selected in north Bandar-Abbas in southeast of Iran. Sampling was carried out, analyzed for selective parameters such as oxygen and carbon isotopic compositions (δ¹⁸O and δ¹³C) and interpreted in the present study. We recognized several diagenetic processes including micritization, cementation, neomorphism, compaction, dissolution, silicification, dolomitization, fracturing and vein filling. Some of the diagenetic processes occurred at different conditions, so in order to achieve precise interpretation, samples from different carbonate components such as, micrite, fracture cement, solution pore cement, intergranular cement, and some biotic allochems were analyzed. In this study micrite samples were subdivided into two groups including micro-spary and micrite. They were recognized under Cathodoluminescence microscope. In addition, micrite samples were classified into five groups based on their depositional environments: supratidal, lagoon, coral bar, open sea, and open basin. There were minor changes in stable isotope ratios based on the sedimentary environments, stratigraphy successions, and micro-spary or micrite properties. In this study, similar calcite cements in petrography studies were differentiated by stable isotope data. Those calcite cements have formed in different diagenetic environments such as meteoric and burial cements. Paragenetic sequence of carbonate rocks were interpreted by integration of petrographic and isotopic studies. We have reconstructed diagenetic models of Guri Member into four stages including marine, meteoric, burial, and uplifting.     

3D Elastic Solutions for Laterally Loaded Discs: Generalised Brazilian and Point Load Tests

This paper investigates the application of a double Fourier series technique to the construction of an elastic stress field in a cylindrical bar subject to lateral boundary loads. The lateral loads, including the constant load boundary conditions, are represented by two Fourier series: one on the perimeter of the circular section (r ₀, θ) and the other on the longitudinal curved surface parallel to the bar axis (z). The technique invokes acceptable potential functions of the Papkovich–Neuber displacement field, satisfying the governing partial differential equations, to assign appropriate odd and even trigonometric Fourier terms in cylindrical coordinates (r, θ, z). The generic solution decomposes the problem of interest to a state of stress caused by two independent boundary conditions along the z axis and θ-polar angle, both superimposed on a solution for which these potentials are the product of the trigonometric terms of the independent variables (θ, z). Constants appearing in the resultant second-order partial differential equations are determined from the generally mixed (tractions and/or displacements) boundary conditions. While the solutions are satisfied exactly at the ends of an infinite bar, they are satisfied weakly on average, in the light of Saint Venant’s approximation at the two ends of a finite bar. The application of the proposed analysis is verified against available elastic solutions for axisymmetric and non-axisymmetric engineering problems such as the indirect Brazilian Tensile Strength and Point Load Strength tests.     

Brittleness Effect on Rock Fatigue Damage Evolution

The damage evolution mechanism of rocks is one of the most important aspects in studying of rock fatigue behavior. Fatigue damage evolution of three rock types (onyx marble, sandstone and soft limestone) with different brittleness were considered in the present study. Intensive experimental tests were conducted on the chosen rock samples and acoustic emission (AE) sensors were used in some of them to monitor the fracturing process. Experimental tests indicated that brittleness strongly influences damage evolution of rocks in the course of static and dynamic loading. AE monitoring revealed that micro-crack density induced by the applied loads during different stages of the failure processes increases as rock brittleness increases. Also, results of fatigue tests on the three rock types indicated that the rock with the most induced micro-cracks during loading cycles has the least fatigue life. Furthermore, the condition of failure surfaces of the studied rocks samples, subjected to dynamic and static loading, were evaluated and it was concluded that the roughness of failure surfaces is influenced by loading types and rock brittleness. Dynamic failure surfaces were rougher than static ones and low brittle rock demonstrate a smoother failure surface compared to high brittle rock.     

An alternative method for non-negative estimation of variance components

A typical problem of estimation principles of variance and covariance components is that they do not produce positive variances in general. This caveat is due, in particular, to a variety of reasons: (1) a badly chosen set of initial variance components, namely initial value problem (IVP), (2) low redundancy in functional model, (3) an improper stochastic model, and (4) data’s possibility of containing outliers. Accordingly, a lot of effort has been made in order to design non-negative estimates of variance components. However, the desires on non-negative and unbiased estimation can seldom be met simultaneously. Likewise, in order to search for a practical non-negative estimator, one has to give up the condition on unbiasedness, which implies that the estimator will be biased. On the other hand, unlike the variance components, the covariance components can be negative, so the methods for obtaining non-negative estimates of variance components are not applicable. This study presents an alternative method to non-negative estimation of variance components such that non-negativity of the variance components is automatically supported. The idea is based upon the use of the functions whose range is the set of all positive real numbers, namely positive-valued functions (PVFs), for unknown variance components in stochastic model instead of using variance components themselves. Using the PVF could eliminate the effect of IVP on the estimation process. This concept is reparameterized on the restricted maximum likelihood with no effect on the unbiasedness of the scheme. The numerical results show the successful estimation of non-negativity estimation of variance components (as positive values) as well as covariance components (as negative or positive values).