Pore size classification and prediction based on distribution of reservoir fluid volumes utilizing well logs and deep learning algorithm in a complex lithology

Pore size analysis plays a pivotal role in unraveling reservoir behavior and its intricate relationship with confined fluids.Traditional methods for predicting pore size distribution(PSD),relying on drilling cores or thin sections,face limitations associated with depth specificity.In this study,we introduce an innovative framework that leverages nuclear magnetic resonance(NMR)log data,encompassing clay-bound water(CBW),bound volume irreducible(BVI),and free fluid volume(FFV),to determine three PSDs(micropores,mesopores,and macropores).Moreover,we establish a robust pore size classification(PSC)system utilizing ternary plots,derived from the PSDs.Within the three studied wells,NMR log data is exclusive to one well(well-A),while conventional well logs are accessible for all three wells(well-A,well-B,and well-C).This distinction enables PSD predictions for the remaining two wells(B and C).To prognosticate NMR outputs(CBW,BVI,FFV)for these wells,a two-step deep learning(DL)algorithm is implemented.Initially,three feature selection algorithms(f-classif,f-regression,and mutual-info-regression)identify the conventional well logs most correlated to NMR outputs in well-A.The three feature selection algorithms utilize statistical computations.These algorithms are utilized to systematically identify and optimize pertinent input features,thereby augmenting model interpretability and predictive efficacy within intricate data-driven endeavors.So,all three feature selection algorithms introduced the number of 4 logs as the most optimal number of inputs to the DL algorithm with different combinations of logs for each of the three desired outputs.Subsequently,the CUDA Deep Neural Network Long Short-Term Memory algorithm(CUDNNLSTM),belonging to the category of DL algorithms and harnessing the computational power of GPUs,is employed for the prediction of CBW,BVI,and FFV logs.This prediction leverages the optimal logs identified in the preceding step.Estimation of NMR outputs was done first in well-A(80%of data as training and 20%as testing).The correlation coefficient(CC)between the actual and estimated data for the three outputs CBW,BVI and FFV are 95%,94%,and 97%,respectively,as well as root mean square error(RMSE)was obtained 0.0081,0.098,and 0.0089,respectively.To assess the effectiveness of the proposed algorithm,we compared it with two traditional methods for log estimation:multiple regression and multi-resolution graph-based clustering methods.The results demonstrate the superior accuracy of our algorithm in comparison to these conventional approaches.This DL-driven approach facilitates PSD prediction grounded in fluid saturation for wells B and C.Ternary plots are then employed for PSCs.Seven distinct PSCs within well-A employing actual NMR logs(CBW,BVI,FFV),in conjunction with an equivalent count within wells B and C utilizing three predicted logs,are harmoniously categorized leading to the identification of seven distinct pore size classification facies(PSCF).this research introduces an advanced approach to pore size classification and prediction,fusing NMR logs with deep learning techniques and extending their application to nearby wells without NMR log.The resulting PSCFs offer valuable insights into generating precise and detailed reservoir 3D models.

Cobalt Deposits of China: Classification, Distribution and Major Advances

The important strategic metal cobalt has diverse uses and the majority of world cobalt deposits have been found in China. The deposits can be classified into four types, i.e., magmatic Ni-Cu-Co sulfide deposits, hydrothermal and volcanogenic cobalt polymetallic deposits, strata-bound Cu-Co deposits hosted by sedimentary rocks and lateritic Ni-Co deposits, of which the former two types are the most important. There are six principal metallogenic epochs and seven important metallogenic belts according to their distribution and tectonic position. Although cobalt generally occurs in nickel-copper, copper and iron deposits as an associated metal, great developments in exploration for independent cobalt deposits have happened in China, and, in recent years, many independent deposits with different elementary assemblages and different genetic types have been discovered in the eastern part of the northern margin of the North China platform, the Central Orogenic Belt of China, western Jiangxi and northeastern Hunan. In addition, it is inferred that the Kunlun-Qinling Orogenic Belt has great potential for further exploration of new types of independent cobalt deposits.

Classification and distribution of Comastoma (Gentianaceae) in Helan Mountains in China by floristic, ecological and geographical approaches

Four species of the Comastoma genus （Geianaceae） in Helan Mountains between Inner Mongolia and Ningxia Province in China have been recognized by morphological and geographical taxonomy. These four species are C. falcatum （Turcz.） Toyokuni, C. polycladum （Diels et Gilg） T. N. Ho, C. tenellum （Rottb.） Toyokuni and C. acutum （Michx.） Y. Z. Zhao et X. Zhang. Among them, C. tenellum （Rottb.） is a new recorded species and C. acutum （Michx.） Y. Z. Zhao et X. Zhang is a new combination. The floristic, ecological and geographical distribution of each species was analyzed and then a new key of Comastoma in Helan Mountains and the distribution maps have been generated, which will provide a reference for the revision of this genus and the analysis of the .flora in Helan Mountains. Key words Comastoma, classification, distribution, Helan Mountains

Distribution and Classification of Cobalt-Rich Crust

Based on the on the spot investigation and related data, cobalt rich crust is mainly distributed in the low latitude area near the equator, mostly within 20°S to 20°N, especially 5°-15°(S and N). The analysis of the microtopographic and microphysicognomy features shows that crusts are often present in the complicated topographic regions such as seamount slopes, convex parts of seamounts and joint faults, of which the ideal region is seamount slopes in water depth of 1 500-2 500 m. The authors analyze the relation of the crusts and their bedrock, bedrock type, crust thickness and occurrence, and then attempt to classify the crusts as different types.

Investigation on fracture models and ground pressure distribution of thick hard rock strata including weak interlayer

Dynamic disasters,such as rock burst due to the breaking of large area stiff roof strata,are known to occur in the hard rock strata of coal mines.In this paper,mechanical models of the fracturing processes of thick hard rock strata were established based on the thick plate theory and numerical simulations.The results demonstrated that,based on the fracture characteristics of the thick hard rock strata,four fracture models could be analyzed in detail,and the corresponding theoretical failure criteria were determined in detail.In addition,the influence of weak interlayer position on the fracture models and ground pressure of rock strata is deeply analyzed,and six numerical simulation schemes have been implemented.The results showed that the working face pressure caused by the independent movement of the lower layer is relatively low.The different fracture type of the thick hard rock strata had different demands on the working resistance of the hydraulic powered supports.The working resistance of the hydraulic powered supports required by the stratified movements was lower than that of the non-stratified movements.

Effective distributed convolutional neural network architecture for remote sensing images target classification with a pre-training approach

How to recognize targets with similar appearances from remote sensing images(RSIs) effectively and efficiently has become a big challenge. Recently, convolutional neural network(CNN) is preferred in the target classification due to the powerful feature representation ability and better performance. However,the training and testing of CNN mainly rely on single machine.Single machine has its natural limitation and bottleneck in processing RSIs due to limited hardware resources and huge time consuming. Besides, overfitting is a challenge for the CNN model due to the unbalance between RSIs data and the model structure.When a model is complex or the training data is relatively small,overfitting occurs and leads to a poor predictive performance. To address these problems, a distributed CNN architecture for RSIs target classification is proposed, which dramatically increases the training speed of CNN and system scalability. It improves the storage ability and processing efficiency of RSIs. Furthermore,Bayesian regularization approach is utilized in order to initialize the weights of the CNN extractor, which increases the robustness and flexibility of the CNN model. It helps prevent the overfitting and avoid the local optima caused by limited RSI training images or the inappropriate CNN structure. In addition, considering the efficiency of the Na¨?ve Bayes classifier, a distributed Na¨?ve Bayes classifier is designed to reduce the training cost. Compared with other algorithms, the proposed system and method perform the best and increase the recognition accuracy. The results show that the distributed system framework and the proposed algorithms are suitable for RSIs target classification tasks.

Strata Architectural Variability and Facies Distribution in a Structural Transfer Zone: A Case Study of Fushan Sag, Northern South China Sea

Structural transfer zones in a half-graben rift basin play a significant role in controlling sandy sediments and providing a target for hydrocarbon exploration. Previous studies have classified the transfer zone in lacustrine environments into two different patterns: synthetic approaching transfer zones and synthetic overlapping transfer zones. However, the evolution of the depositional pattern and the controlling factors of the above transfer zones are still unclear. In the Fushan Sag, the northern South China Sea, an overlapping transfer zone developed in the early Eocene Epoch, while a synthetic approaching transfer zone developed in the late Eocene, due to tectonic uplift. This evolutionary process provided an opportunity to study the stacking pattern of strata architectural variability and facies distribution in the structural transfer zone of the Eocene lacustrine basin. In this study, following the indications of the oriented sedimentary structures in core samples and heavy mineral assemblages of 18 wells, the evolution of the paleo-hydrodynamic distribution during the early and late Eocene has been reconstructed. The sequence-stratigraphy was then divided and the sand body parameters calculated, according to the seismic data and well log interpretations. During the early Eocene, the lake level was at a low stand, the faults broken displacement in the East block being over 50 m. The prograding delta and turbidites are oriented perpendicular to the structural transfer zone. According to the quantitative analysis of the flow rate and the depositional parameters, we speculate that gravity transportation of the sediment and the sediment-supply are the dominating factors during this period. Up to the late Eocene, the rising lake level and the decreased fault displacement leads the flow to divert to a NE-direction, resulting in it being parallel to the axis of the transfer zone. Thus, we speculate that the accommodation space is predominant in this period. In comparison with the above two periods, a braided river delta with an isolated sand body and turbidites developing in the deep area is prominent in the overlapping transfer zone, while a meandering river delta is characteristic of the synthetic approaching transfer zone.

Classification with Convolutional Neural Networks in MapReduce

Deep learning (DL) techniques, more specifically Convolutional Neural Networks (CNNs), have become increasingly popular in advancing the field of data science and have had great successes in a wide array of applications including computer vision, speech, natural language processing, etc. However, the training process of CNNs is computationally intensive and has high computational cost, especially when the dataset is huge. To overcome these obstacles, this paper takes advantage of distributed frameworks and cloud computing to develop a parallel CNN algorithm. MapReduce is a scalable and fault-tolerant data processing tool that was developed to provide significant improvements in large-scale data-intensive applications in clusters. A MapReduce-based CNN (MCNN) is developed in this work to tackle the task of image classification. In addition, the proposed MCNN adopted the idea of adding dropout layers in the networks to tackle the overfitting problem. Close examination of the implementation of MCNN as well as how the proposed algorithm accelerates learning are discussed and demonstrated through experiments. Results reveal high classification accuracy and significant improvements in speedup, scaleup and sizeup compared to the standard algorithms.

Analysis of Distributed and Adaptive Genetic Algorithm for Mining Interesting Classification Rules

Distributed genetic algorithm can be combined with the adaptive genetic algorithm for mining the interesting and comprehensible classification rules.The paper gives the method to encode for the rules,the fitness function,the selecting,crossover,mutation and migration operator for the DAGA at the same time are designed.

Types and Spatial Distribution of Geo-tourism Resources in Jilin Province of China

Natural geological conditions and geo-tourism resources in Jilin Province were introduced,and distribution features of the local major tourist resources(vegetation-covered eastern region,grass swamp on western plain) were studied.Jingyu Volcanic Mineral Spring Geo-park,Changbai Mountain Geo-park and Qian'an Mud Forest Geo-park were studied as typical geo-tourism resources,so as to provide basic data for the systematic development and construction of geo-tourism resources in Jilin Province.

Distribution characteristics of historical earthquake classes in Jiangsu Province and South Huanghai Sea region

According to the analysis on the characteristics of historic earthquakes in Jiangsu Province and South Huanghai Sea region, the historical earthquakes in the studied area are divided into two kinds of comparatively safe class and comparatively dangerous class. Then the statistical result of earthquake class, the characteristics of geo-graphical distribution and geological structures are studied. The study shows: a) In Jiangsu Province and South Huanghai Sea region, the majority of historical strong earthquakes belong to comparatively safe class, only 13.8% belong to comparatively dangerous class; b) Most historical earthquakes belong to comparatively safe class in the land area of Jiangsu, eastern sea area of Yangtze River mouth and northern depression of South Huanghai Sea region. However, along the coast of middle Jiangsu Province and in the sea area of South Huanghai Sea, the distribution of historical earthquake classes is complex and the earthquake series of comparatively dan-gerous class and comparatively safe class are equivalent in number; c) In the studied area, the statistical results of historical earthquake classes and the characteristics of spatial distribution accord very well with the real case of present-day earthquake series. It shows that the seismic activity in the region has the characteristic of succession, and the result from this study can be used as a reference for early postseismic judgment in the earthquake emer-gency work in Jiangsu Province.

Statistical Characteristics of Raindrop Size Distribution in the Tibetan Plateau and Southern China

The characteristics of raindrop size distribution （DSD） over the Tibetan Plateau and southern China are studied in this paper, using the DSD data from April to August 2014 collected by HSC-PS32 disdrometers in Nagqu and Yangjiang, com- prising a total of 9430 and 63661-rain raindrop spectra, respectively. The raindrop spectra, characteristics of parameter variations with rainfall rate, and the relationships between reflectivity factor （Z） and rainfall rate （R） are analyzed, as well as their DSD changes with precipitation type and rainfall rate. The results show that the average raindrop spectra appear to be one-peak curves, the number concentration for larger drops increase significantly with rainfall rate, and its value over southern China is much higher, especially in convective rain larger drops, especially for convective rain in southern China. Standardized Gamma distributions better describe DSD for All three Gamma parameters for stratiform precipitation over the Tibetan Plateau are much higher, while its shape parameter （,u） and mass-weighted mean diameter （Dm）, for convective precipitation, are less. In terms of parameter variation with rainfall rate, the normalized intercept parameter （Nw） over the Tibetan Plateau for stratiform rain increases with rainfall rate, which is opposite to the situation in convective rain. The/1 over the Tibetan Plateau for stratiform and convective precipitation types decreases with an increase in rainfall rate, which is opposite to the case for Dm variation. In Z-R relationships, like ＂Z = ARb＇＇, the coefficient A over the Tibetan Plateau is smaller, while its b is higher, when the rain type transfers from stratiform to convective ones. Furthermore, with an increase in rainfall rate, parameters A and b over southern China increase gradually, while A over the Tibetan Plateau decreases sub- stantially, which differs from the findings of previous studies. In terms of geographic location and climate over the Tibetan Plateau and southern China, the precipitation in the pre-flood seasons is dominated by strong convective rain, while weak convective rain occurs frequently in northern Tibet with lower humidity and higher altitude.

Parallel naive Bayes algorithm for large-scale Chinese text classification based on spark

The sharp increase of the amount of Internet Chinese text data has significantly prolonged the processing time of classification on these data.In order to solve this problem,this paper proposes and implements a parallel naive Bayes algorithm(PNBA)for Chinese text classification based on Spark,a parallel memory computing platform for big data.This algorithm has implemented parallel operation throughout the entire training and prediction process of naive Bayes classifier mainly by adopting the programming model of resilient distributed datasets(RDD).For comparison,a PNBA based on Hadoop is also implemented.The test results show that in the same computing environment and for the same text sets,the Spark PNBA is obviously superior to the Hadoop PNBA in terms of key indicators such as speedup ratio and scalability.Therefore,Spark-based parallel algorithms can better meet the requirement of large-scale Chinese text data mining.

A parametric bootstrap approach for one-way classification model with skew-normal random effects

In this paper,several properties of one-way classification model with skew-normal random effects are obtained,such as moment generating function,density function and noncentral skew chi-square distribution,etc.Based on the EM algorithm,we discuss the maximum likelihood(ML)estimation of unknown parameters.For testing problem of fixed effect,a parametric bootstrap(PB)approach is developed.Finally,some simulation results on the Type I error rates and powers of the PB approach are obtained,which show that the PB approach provides satisfactory performances on the Type I error rates and powers,even for small samples.For illustration,our main results are applied to a real data problem.

Review of Text Classification Methods on Deep Learning

Text classification has always been an increasingly crucial topic in natural language processing.Traditional text classification methods based on machine learning have many disadvantages such as dimension explosion,data sparsity,limited generalization ability and so on.Based on deep learning text classification,this paper presents an extensive study on the text classification models including Convolutional Neural Network-Based(CNN-Based),Recurrent Neural Network-Based(RNN-based),Attention Mechanisms-Based and so on.Many studies have proved that text classification methods based on deep learning outperform the traditional methods when processing large-scale and complex datasets.The main reasons are text classification methods based on deep learning can avoid cumbersome feature extraction process and have higher prediction accuracy for a large set of unstructured data.In this paper,we also summarize the shortcomings of traditional text classification methods and introduce the text classification process based on deep learning including text preprocessing,distributed representation of text,text classification model construction based on deep learning and performance evaluation.

Self Organization Map for Clustering and Classification in the Ecology of Agent Organizations

Development of computational agent organizations or “societies” has become the domiant computing paradigm in the arena of Distributed Artificial Intelligence, and many foreseeable future applications need agent organizations, in which diversified agents cooperate in a distributed manner, forming teams. In such scenarios, the agents would need to know each other in order to facilitate the interactions. Moreover, agents in such an environment are not statically defined in advance but they can adaptively enter and leave an organization. This begs the question of how agents locate each other in order to cooperate in achieving organizational goals. Locating agents is a quite challenging task, especially in organizations that involve a large number of agents and where the resource avaiability is intermittent. The authors explore here an approach based on self organization map (SOM) which will serve as a clustering method in the light of the knowledge gathered about various agents. The approach begins by categorizing agents using a selected set of agent properties. These categories are used to derive various ranks and a distance matrix. The SOM algorithm uses this matrix as input to obtain clusters of agents. These clusters reduce the search space, resulting in a relatively short agent search time.

Classification of tight sandstone reservoirs based on NMR logging

The traditional reservoir classification methods based on conventional well logging are inefficient for determining the properties,such as the porosity,shale volume,J function,and flow zone index,of the tight sandstone reservoirs because of their complex pore structure and large heterogeneity.Specifically,the method that is commonly used to characterize the reservoir pore structure is dependent on the nuclear magnetic resonance(NMR)transverse relaxation time(T2)distribution,which is closely related to the pore size distribution.Further,the pore structure parameters(displacement pressure,maximum pore-throat radius,and median pore-throat radius)can be determined and applied to reservoir classification based on the empirical linear or power function obtained from the NMR T2 distributions and the mercury intrusion capillary pressure ourves.However,the effective generalization of these empirical functions is difficult because they differ according to the region and are limited by the representative samples of different regions.A lognormal distribution is commonly used to describe the pore size and particle size distributions of the rock and quantitatively characterize the reservoir pore structure based on the volume,mean radius,and standard deviation of the small and large pores.In this study,we obtain six parameters(the volume,mean radius,and standard deviation of the small and large pores)that represent the characteristics of pore distribution and rock heterogeneity,calculate the total porosity via NMR logging,and classify the reservoirs via cluster analysis by adopting a bimodal lognormal distribution to fit the NMR T2 spectrum.Finally,based on the data obtained from the core tests and the NMR logs,the proposed method,which is readily applicable,can effectively classify the tight sandstone reservoirs.

Dense Pellicular Agarose-Glass Beads for Expanded Bed Application: Flow Hydrodynamics and Solid Phase Classifications

Two dense pellicular agarose-glass matrices of different sizes and densities, i.e., AG-S and AG-L, have been characterized for their bed expansion behavior, flow hydrodynamics and particle classifications in an expanded bed system. A 26 mm ID column with side ports was used for sampling the liquid-solid suspension during expanded bed operations. Measurements of the collected solid phase at different column positions yielded the particle size and density distribution data. It was found that the composite matrices showed particle size as well as density classifications along the column axis, i.e., both the size and density of each matrix decreased with increasing the axial bed height. Their axial classifications were expressed by a correlation related to both the particle size and density as a function of the dimensionless axial bed height. The correlation was found to fairly describe the solid phase classifications in the expanded bed system. Moreover, it can also be applied to other two commercial solid matrices designed for expanded bed applications.

Variability of Raindrop Size Distribution during a Regional Freezing Rain Event in the Jianghan Plain of Central China

The characteristics of the raindrop size distribution(DSD)during regional freezing rain(FR)events that occur throughout the phase change(from liquid to solid)are poorly understood due to limited observations.We investigate the evolution of microphysical parameters and the key formation mechanisms of regional FR using the DSDs from five disdrometer sites in January 2018 in the Jianghan Plain(JHP)of Central China.FR is identified via the size and velocity distribution measured from a disdrometer,the discrete Fréchet distancemethod,surface temperature,human observations,and sounding data.With the persistence of precipitation,the emergence of graupel or snowflakes significantly reduces the proportion of FR.The enhancement of this regional FR event is mainly dominated by the increase in the number concentration of raindrops but weakly affected by the diameters.To improve the accuracy of quantitative precipitation estimation for the FR event,a modified second-degree polynomial relation between the shapeμand slopeΛof gamma DSDs is derived,and a new Z-R(radar reflectivity to rain rate)relationship is developed.The mean values of mass-weighted mean diameters(D_(m))and generalized intercepts(lgN_(w))in FR are close to the stratiform results in the northern region of China.Both the melting of tiny-rimed graupels and large-dry snowflakes are a response to the formation of this regional FR process in the JHP,dominated by the joint influence of the physical mechanism of warm rain,vapor deposition,and aggregation/riming coupled with the effect of weak convective motion in some periods.