جستجوی مقالات مرتبط با کلیدواژه « seismic attributes » در نشریات گروه « زمین شناسی »

Nowadays, the use of artificial intelligence is common to increase the accuracy of the study and, close to reality, is used in the oil industry to increase the accuracy of studying and understanding the relationship between various parameters. The main purpose of this study is to compare the performance of the two methods of Extreme Learning Machine (ELM) and Radial Basis Function (RBF) in porosity estimation, which is static oil modeling. The data from seven wells in the offshore field (Hendijan Oilfield) of the northwestern Persian Gulf were examined. In this regard, post-stack seismic attributes which have a significant relationship with porosity and porosity log for each well were used to compare the performance of the ELM and RBF networks under the same conditions. Eventually, it reveals that ELM is quite sensitive to the data set and needs more data points to prepare a map (quantitatively), but is better than RBF in terms of classification (qualitative). On the other hand, RBF is one of the most powerful algorithms in mapping, especially in low numbers of data points, which can be challenging for others.

One of the basic steps of oil exploration is to define the hydrocarbon zone. Different methods have been used so far for defining such zones. For a specific dataset, finding the most appropriate method leads to more accurate estimates and predictions of analysis besides improving the speed of calculations. Support Vector Machine (SVM), which is one of the methods for analyzing the data, uses kernel functions. It finds a better relationship between data factors and hydrocarbon zone leading to better estimates and classifications.    In this article, hydrocarbon zone detection has been done using seismic and well data.    The purpose of facies analysis is to obtain important petrophysical parameters of the reservoir and to identify heterogeneous boundaries below the ground. The results of the interpretation of petrophysical parameters are the input of the three-dimensional reservoir modeling process and through these parameters, the reservoir parameters are distributed in three-dimensional space. This model is widely used in various sections such as exploration and drilling of new wells, overdraft from a reservoir, determination of suitable areas for overdraft, reduction of drilling risk and risk, determination of reservoir lithology and identification of key well and its extension to other wells in the region. The most important petrophysical parameters are shale volume, porosity, permeability, reservoir fluid saturation and reservoir lithology.    The study of seismic facies has been started since the 90's and due to its importance and application in reservoir description, it has always been considered by many researchers.    To perform the analysis above, first, the hydrocarbon zones were spotted across the Asmari Formation using well logs and well geology reports. Next, the SVM method was used to detect each hydrocarbon zone using well logs. There was an acceptable agreement between the results of SVM method and well geology reports. Second, hydrocarbon zones detection was done using seismic data by SVM. At this stage, seismic attributes were extracted from the seismic trace in the well location. Then, covariance matrix and cross plots of seismic attributes used to identify the most effective attributes to hydrocarbon zones detection. In order to validate the results, the seismic attributes of another trace near the well location were used for hydrocarbon zone detection. SVM results matched hydrocarbon zones with low error.

Introduction:

 In this study, using principal component analysis on seismic attributes, the main component with a high percentage of variance are prepared and presented in the form of appropriate sections. These sections can demonstrate smaller fractures and faults with better resolution. Although these images and sections diagnose faults well, presentation of these images in RGB color enhances the resolution of fractures. 

Methodology and Approaches:

 To raise quality levels of seismic sections, the attributes that have the optimal band combination of using statistical indicators are selected and by applying principal components analysis on these attributes, we obtain images that contain more information. 

Results and Conclusions:

 The resulting images, using the methods proposed in this study, have more details, and are more accurate, than the images obtained from other methods in showing faults and fractures.

Fault identification and investigating their evolution is of special importance in the exploration and development of hydrocarbon resources. Success in exploration and development of hydrocarbon fields, need to recognition of petroleum systems and in this regard one of the most important topics is identifying faults and their extension condition as a main fluid migration path, specially in deeper zones. Faults and fractures have crucial role in making high permeable and porous segments and cut reservoir and cap rock in the fluid migration path. In addition, for maximizing the production of hydrocarbon from reservoirs and also for reducing the risk of drilling, it is necessary to gain information about geometry and nature of faults of reservoirs. In this paper, the purpose is investigating the performance of combination of neural networks and Fault Likelihood auto-tracking algorithm for identification and interpretation of faults in seismic data. At first using the Dip-steering feature of software, the early filter for accurate identification of dip of structures in the data, have been designed and applied. Then with designing and applying the appropriate filters, the seismic data have been improved. After that proper seismic attributes for fault identification have been calculated from seismic data. With picking fault and non-fault points from data, a supervised neural network using the selected attributes was formed and after training the network, the appropriate output achieved. Then the output of neural network has been used as a input for Thinned Fault Likelihood auto-tracking algorithm. The output of this part contains a volume of tracked faults. Finally using sub-tools of TFL and optimal setting of parameters, 3D fault planes has been interpreted and extracted.

The analysis of seismic facies is a technique for mapping geological features and properties using seismic data. To analyze seismic facies, seismic attributes are categorized and classified using machine learning algorithms to identify different seismic facies. Seismic facies analysis due to the nature of seismic data, which always has a degree of uncertainty, can produce different results with even small changes in input parameters of the analyzing method. For this reason, it is necessary to select the different stages of analysis, including the selection of input parameters and algorithm of machine learning, with high accuracy with regard to the objective of the seismic facies analysis. In this study, an interactive method with the supervision of interpreter is proposed for producing seismic facies map, using the optimal selection of the input parameters and the the proper selection of clustering and classification algorithms. In this method, the interpreter in a recursive and rotational process can compare the results of the analysis and generate thr optimal results by changing the input parameters. The method presented in this article is implemented in SeisART software. SeisART has a complete environment for data initialization (importing seismic data and well data). A user-friendly interactive environment allows the user to implement several methods and monitor the corresponding result in 2D and 3D.
SeisART software makes the possibility of the interpreter contribution in the whole stages of seismic facies analysis procedure. The interpreter can select the input attributes and chose the proper methods of pattern recognition to reach the best possible result. In the software, various evaluation utilities have been provided in each stage of seismic facies analysis. These utilities allow the interpreter to monitor the results of each method quantitatively and qualitatively. In the unsupervised system, clustering quality factors are used. The interpreter calculates the validation indices for different methods of clustering and identifies the proper method which has been more successful in discovering the natural grouping of patterns in the data set. Afterward, if there is structural geology information about the horizon of interest, the interpreter can decide on the clustering result with more accuracy. In the supervised system, the most proper method is feasible using minimization of training data and validation data errors. In this case, the interpreter can use geological knowledge and well data information to verify obtained results. In this method, the interpreter can obtain different results by changing the input parameters. Comparing these results, and taking into account the path leading to this result, the interpreter gains more knowledge of existing facies.
This method has been applied to the MSF4 horizons of the 3D seismic data of the North Sea F3 and has been shown which method is more efficient for different purposes

In this paper, by combination of seismic data and well log data, the effective porosity in the space between wells is estimated. One of the important petroleum reservoir features is effective porosity that engineers are always looking to find an appropriate model for distribution of this parameter in the reservoir. The petrophysical properties of petroleum reservoirs are very complex. In the last few decades, the effective porosity estimation procedures have become one of the hot topics in the industry to evaluate these procedures or methods. In the current research, by integration of petrophysical, seismic data and seismic attributes classification using Adaboost algorithm, it is tried to estimate the effective porosity in a twodimensional seismic cross section of the block F3 Dutch sector of the North Sea. In the first step, seismic attributes of two-dimensional seismic section has been extracted. Based on the feature selection methods, six seismic attributes have positive impact on the process of the classification. The six attributes of seismic data and effective porosity have been classified and labeled in the positions of wells. Adaboost algorithm has been used for training, and training has been led to the accuracy of 76.6 percent. The next step is the accuracy of Adaboost algorithm or validation, which has been led to accuracy of 71.7 percent. Adaboost algorithm can also be used for classification of effective porosity in other parts of the reservoir where data are not available. The algorithm outputs have shown a good effective porosity in the lower layers, where there may be an economic horizon for the oil production.

Channels are one of the most important stratigraphic and morphological events. If channels place in a suitable position such as enclosed in impermeable place can make suitable oil and gas reservoir; So identifying channels are crucial. Different tools such as filters, seismic attributes, artificial neural networks, and meta-attributes have played an important role in this regard. In this paper dip-steering cube, dip-steer median filter, dip-steer diffusion filter, and fault enhancement filter, have been used. Then, various seismic attributes such as similarity, texture, spectral decomposition, energy and polar dip have been defined and studied. Therefore, work on F3 real seismic data of Dutch part of the North sea for detecting channels has been started by detecting suitable attributes. For identifying the channel in data, it has been used from compilation and combination of seismic attributes using supervised ANN (multi-layer perceptron), and development of mata-attributes, then recombine meta-attributes created along the channel, and using different interpretation point, for eliminating the impact of facies and lithology changes along the channel. Among the advantages and the reasons for using this kind of neural network (supervised), which increases the effect of the neural network and improves the result, is the ability to train the network by specifying the channel and non-channel points used in this paper. Finally, using the above methods, the identification of the channel examined in the above seismic data has been improved, and the channel has been properly detected and extracted throughout its entire length.

Porosity is one of the most important parameters in evaluation of reserves storage capacity and hydrocarbon production. Commonly, this parameter is calculated via core and well log data. One of the most important limitation of these two data groups, despite their high values are lack of development which is limited to well positions. Estimation of petrophysical parameters in inter- and outer-well spaces is valuable and important. The use of seismic data with a large lateral expansion has an important and critical role in the estimation of petrophysical parameters in the inter-well spaces. In this research, Neutron porosity estimation using seismic attributes, on Asamri Formation in the Chashmeh Khush Field (N. Ahwaz) was done. In the first, by using of Sonic and Density logs, synthetic seismogram were created and then they were correlated with seismic data and average seismic wavelet used in inversion process, were extracted from three well with total correlation 87.47%. Afterward, three dimensional seismic data were converted to the acoustic impedance via inversion process and were used as external attributes to the accompaniment of internal attributes resulted from raw seismic data during the estimation. using stepwise regression, the relationship between porosity and seismic attributes were determined and eight seismic attributes including inversion results, derivative Instantaneous Amplitude, Integrated Absolute Amplitude, Derivative Instantaneous Frequency, Filter 25/30-35/40, Cosine Instantaneous Phase and Raw Seismic were selected as optimized attributes. Then via two Multiattribute linear regression (MLR) and Probabilistic neural network methods (PNN), porosity estimated and specified according to validation result and error, PNN method was proven to be more efficient. Finally, based on the porosity estimated from seismic data, it was cleared that mixed clastic-carbonate intervals in the lower sequence, generally have a better porosity condition than dolomite and limestone in the upper part, this trend is easily recognizable from well log data.

Time representation was the first way to describe a signal, and later on the frequency representation was introduced as another important way to describe a signal for its physical significance. Due to the non-stationary property of seismic data, time-frequency transform has to be used to analyze it. During the last decade, spectral decomposition techniques have proven to be an excellent tool to describe thin beds associated with channel sands, alluvial fans, and the like. However, with the traditional spectral decomposition method based on the short time Fourier Transform, it is difficult to acquire the accurate time-frequency spectrum for non-stationary seismic signals. Recently, the emergence of seismic attribute co-rendering, principal component analysis, cluster analysis, and neural networks has partially solved the problem, but the extraction of spectral attributes from spectral-decomposition tightly linked to the geology has more advantages over other approaches. Popular time–frequency methods have some disadvantages. A good time resolution requires a short window and a good frequency resolution require a narrow-band filter, i.e. a long window, but unfortunately, these two cannot be simultaneously realized. The Wigner-Ville Distribution (WVD) of a signal is the Fourier Transform of the signal’s time-dependent auto-correlation function, a quadratic expression which is bilinear in the signal. As a result, the cross-terms appear in the locations of the resulting time-frequency spectra that either interfere with the interpretation of auto-terms or for which we can provide no physical interpretation. Due to the existence of cross-terms, WVD is not often used. Reduction of the cross-terms is achieved by manipulating the ambiguity function as a mask that reduces the cross-terms while preserving the time and frequency resolution of WVD. The short-time Fourier Transform (STFT) spectrogram, which is the squared modulus of the STFT, is a smoothed version of WVD. An STFT spectrogram is a 2-D convolution of the signal WVD and the utilized window function. In this paper, we introduce a Deconvolutive Short-Time Fourier Transform (DSTFT) spectrogram method, which improves the time-frequency resolution and reduces the cross-terms simultaneously by applying a 2-D deconvolution operation on the STFT spectrogram. Compared to the STFT spectrogram, the spectrogram obtained by this method shows a significant improvement in the time-frequency resolution. In this study, we extract two attributes namely the peak frequency and the peak amplitude, based on the Deconvolutive Short- Time Fourier Transform. The maximum frequency attribute is directly related to the thickness of the thin-bed, like channel, and the maximum amplitude attribute also responds to the thin-bed. We use instantaneous seismic attributes: maximum instantaneous frequencies and their associated amplitudes, as a tool to detect seismic geomorphologic bodies and to identify thin layers. Then we use attributes extracted by Deconvolutive Short Time Fourier Transform to detect the burial channel in both synthetic and real 3D seismic data. Usually, the center of the channel is recognized by the lower maximum frequency and when the thickness of the channel gets thinner away from the center of the channel, the maximum frequency increases correspondingly. Therefore, this attribute could clearly describe the distribution of channel both vertically and horizontally. Results of this study on the synthetic and real seismic data examples illustrate the good performance of the DSTFT spectrogram compared with other traditional time-frequency representations.

Prediction of spatial distribution of porosity in a reservoir is an essential issue forestimating reserves and planning production operations. In most cases, however, lateralvariations of porosity cannot be delineated from measurements made at sparsely located wells. The integration of 3D seismic data with petrophysical measurements cansignificantly improve the spatial description of porosity. In the last two decades, severalmethods have been developed for the estimation of reservoir porosity. A number ofinversion methods are available in the industry to convert seismic amplitude into acousticimpedance. Acoustic impedance is indirectly related to porosity. Alternate integrativeapproaches for estimating porosity include geo-statistical methods, such as kriging andco-kriging using well and seismic data. One of the most accurate methods for estimatingreservoir parameters is the application of seismic attributes by a nonlinear estimator, suchas neural network or neuro-fuzzy model. Both neural networks and neuro-fuzzy modelscan be good estimators but the latter has the benefit of being interpretable. In this study, aneuro-fuzzy model called NEFPROX was used to estimate porosity in a gas reservoirlocated in the Gorgan Basin.NEFPROX is a Mamdani-type neuro-fuzzy model, so it has an advantage of beinginterpretable that makes it distinct from other type of neuro-fuzzy models. The timeconsumingcharacteristic of that method is irrelevant in this case because the prediction ofporosity is an offline prediction problem.The Gorgan Basin is located in the northern part of Iran, southeast of the Caspian Sea.This area consists mainly of three formations:the upper formation, called Clay-SandGroup , belongs to quaternary period. Below Clay-Sand Group  is a tertiary formationcalled Clay-Sand Group. A formation of Brown Beds is also a tertiary formation thatlies below Clay-Sand Group. All of these formations consist mainly of shale and sand.The discovery of gas in the Brown Beds Formation has persuaded explorationists toincrease their activities in the Gorgan Basin. The purpose of this study is to recognizeshale and sand bodies in the Brown Beds formation that consists of alternative sand andshale layers with variable thickness.First, a list of 20 seismic attributes was prepared to extract from raw seismic data inthe location of wells. Stepwise regression was used to select four appropriate attributes.The maximum number of attributes was set at four to avoid the model complexity. Theseattributes, in the order of priority, are instantaneous frequency, amplitude weightedfrequency, apparent polarity, and second derivative instantaneous amplitude. Then thesefour selected attributes were introduced into the neuro-fuzzy model as input to predictporosity as an output of the model.The neuro-fuzzy model was trained with the data of well GO3. Based on hydrocarboncore samples obtained from the Brown Beds formation and the potential of this formationas a probable reservoir, the data corresponding to this formation were selected as atraining data. A model blind test was also conducted with the data of well GO5.Porosity sections generated as the output of the model showed two low porosity sandyand shaly-sand channels in the Brown Beds formation. Lateral variations of thesechannels can clearly be recognized in these sections. The core samples available in wellGO3 (containing hydrocarbon) confirm the existence of the two inferred channels. Thisclear image of channels is simply unidentifiable from raw seismic data. Hence,NEFPROX can be very helpful in supplying valuable information about extent, shape andlithological variation of a reservoir. Finally, compared comparison was made between theperformance of the neuro-fuzzy model and regular neural networks in estimation ofporosity. The comparison indicates that the accuracy of the NEFPROX estimation isequal to that of MLP and is greater than that of RBF.

Reservoir properties, such as porosity and permeability, can be derived at well locations from core samples or well log measurements. Since these properties vary laterally from one well to another, it is normally very difficult to predict reservoir properties away from wells. Seismic data, particularly 3D surveys, contain valuable information about the lateral variation of reservoir properties. When wells fall within the seismic coverage, it is logical to predict reservoir properties between wells by interpreting seismic data and using reservoir properties at well locations as spatial control points. Artificial neural networks (ANN) may be used to aid the estimation of reservoir properties between wells. In this case, a training sample set of input and output data pairs can be collected. The input of neural networks is seismic data relevant to the reservoir at well locations. The expected output from neural networks is reservoir properties at well locations. This paper uses a local linear neuro-fuzzy model to predict reservoir properties from seismic attributes in one of the oil fields in the central region of Iran. The fundamental approach with the locally linear neuro-fuzzy model is dividing the input space into small linear subspaces with fuzzy validity functions. Any linear model produced, along with its validity function, can be described as a fuzzy neuron. Thus, the total model is a neuro-fuzzy network with one hidden layer and a linear neuron in the output layer, which simply calculates the weighted sum of the outputs of locally linear neurons. An incremental tree-based learning algorithm, a locally linear model tree (LOLIMOT), is appropriate for tuning rule premise parameters, i.e. determining the validation hypercube for each locally linear model. In iteration, the worst performing locally linear neuron is determined and then divided. All of the possible divisions in the p dimensional input space are checked, and the best is performed. The splitting ratio is simply adjusted as 1/2, which means that the locally linear neuron is divided into two equal halves. The fuzzy validity functions for the new structure are updated. Their centers are identical with the centers of the new hypercubes and the standard deviations are usually set as 0.3. Just one parameter, the embedding dimension, should be defined before running the algorithm. In this work, the number of attributes is the embedding dimension. To evaluate the performance of a locally linear neuro-fuzzy model in extracting the relationship between seismic attributes and reservoir property, this method was applied in an oil field located in central Iran. This field has two exploration wells. Additionally, a 3D seismic survey, recorded with a one-millisecond sampling interval, covers the area of the reservoir. First, the logs were mapped to the time domain and then blocked it at each one millisecond to resolve the frequency difference of the logs and seismic data. Then, multi-attribute analyses were performed and the best seismic attributes, which had good correlations with porosity, were found to be acoustic impedance and amplitude weighted cosine phase. By applying the locally linear neuro-fuzzy model to train and validate the network, good results were obtained. The correlation coefficient between the modeled and original logs was 80% and the error was 2.6% in validation. Finally, to compare the neuro-fuzzy model with traditional methods, the work was repeated with a probabilistic neural network (PNN) and a multi-layer forward neural network (MLFN). The results obtained by applying the MLFN (correlation 83% and error 4.5) and PNN (correlation 68% and error 5.7) were not better than neuro-fuzzy model.

During the past two decades, marine researches have focused on gas hydrates because of many reasons that the most important of them is future energy resource potential. Gas hydrates are ice-like crystals that form a rigid cage of water molecules and entrap hydrocarbon and non-hydrocarbon gas by hydrogen bonding. There are many methods for prospecting gas hydrates. Seismic methods are the most applicable tools to study them. The presence of the suitable thermodynamic conditions, gas and water in appropriate amount and gas migration pathway from depth to sea floor sediment are necessary for growth and stability of gas hydrates and evaluation of these conditions are necessary for prospecting gas hydrates. The occurrence of gas hydrates and free gas in host sediments changes the elastic properties and makes them detectable with seismic methods. The important seismic indicators of gas hydrates are bottom simulating reflector, flat spot and bright spot. Therefore, the presence of the gas hydrates in host sediments are detectable with seismic indicators and the study of the gas hydrate anomalies are applicable with seismic attributes. In this article, the appropriate conditions for growth and stability of gas hydrates in OmanSea were examined with 2D seismic sections, the occurrence of the gas hydrates in sediments was proven with seismic indicators and the properties of their indicators were studied with reflection strength and apparent polarity attributes. Also the effect of the gas hydrates in attenuation of seismic wave amplitude was studied with spectral decomposition attribute. The results of this study prove the presence of gas hydrates in OmanSea.