Catálogo de publicaciones - revistas

<jats:p> The sand architecture interpretation and modeling of the different orders of sedimentary bodies are of great significance for the efficient development of the unconventional reservoir. However, sand architecture interpretation and modeling using a small amount of observation information are extraordinarily difficult. We took the X36 gas field in Xihu Sag as an example to study the distribution and superimposition characteristics of sandbodies under the influence of tides. We proposed a set of architecture characterization and modeling methods suitable for the condition of a few wells. First, we extracted a variety of seismic attributes from the original seismic data, and we fit the correlations between seismic attributes and sandstone thickness interpreted by logging. Then, we classified and integrated the seismic attributes with a high correlation to calculate the fusion seismic attributes. The fusion seismic attribute has a higher correlation with sandstone thickness. The boundary of the sandstone in the fusion seismic attribute is clearer than a single attribute. We established the sand body distribution model of the study area with the constraints of the fusion seismic attribute and the distance attribute volume. The results indicated that the sand body distribution in the model was more consistent with the sand body development model under the influence of the tide. The results can guide architecture characterization and remaining oil potential tapping of the oil and gas fields offshore. </jats:p>

<jats:p> Carbonate fault-controlled karst reservoirs are unique because they typically are deeply buried and have substantial heterogeneity throughout compared with sandstone or other carbonate reservoirs. Creating a characterization method is critical for the efficient development of this kind of reservoir. The Shunbei 5 fault zone is an example of a carbonate fault-controlled karst reservoir located in the Tarim Basin. We summarized a 3D architectural model of fault-controlled karst reservoir based on the outcrop, drilling, logging, and seismic data. We dissected the architectural characteristics and established a comprehensive technology series to characterize fault-controlled karst reservoirs. The results indicate that the strike-slip fault and its broken strata are the geologic basis for karst development. We divided the architectural elements of the fault-controlled karst reservoirs into five categories: fault core damaged zone, fracture-damaged zone, dissolution pore zone, large cavern zone, and cavern filling zone. The fault core damaged zones form along the main slip surface of the strike-slip faults, and its interior is composed of a fault core and damaged zone. We proposed U-net deep machine learning network based on the seismic data to predict the 3D distribution of the fault core damaged zone. The fracture-damaged zone is the weakest part of karstification with poor reservoir quality. The fracture-damaged zone’s 3D distribution was effectively characterized using the improved seismic data texture attribute. The dissolution pore zone is mainly distributed outward the fracture-damaged zone, with a medium karstification degree. The optimized seismic data energy envelope attribute can effectively characterize the 3D aspect of this region, with the large cavern zone representing the position with the most substantial degree of karstification. The reservoir quality depends on the fillings inside the cavern, and we predicted the large cavern zones based on the energy enhanced residual impedance properties. </jats:p>

<jats:p> Numerous studies have demonstrated the capability of supervised deep learning techniques for predicting geologic features of interest from seismic sections, including features that are difficult to identify using traditional interpretation methods. However, the successful application of these techniques in practice has been limited by the difficulty of obtaining a large training data set where the seismic data and corresponding ground truth labels are well-defined. Manually creating large amounts of labels requires a heavy workload, and the uncertainty of the interpretation and labeling process decreases the model’s ability for making accurate predictions. Using the chalk-flint sequence scenario onshore Denmark as an example, we have developed a novel workflow for predicting subresolution thin layers from seismic sections. It entails generating large quantities of synthetic training data with high-quality labels using stochastic geologic modeling, training a convolutional neural network based on the synthetic data set, and applying it to real seismic data. This is, to our knowledge, the first example of using deep learning to predict subresolution thin layers from seismic data based on geostatistically generated training images. It is shown that a neural network trained on synthetic data can predict a realistic number of subresolution flint layers from the real seismic data that have been collected from the Stevns region in Denmark, which has value for the understanding of the overall geologic characteristics of succession and engineering applications such as construction site evaluation. </jats:p>

<jats:p> Identification and prediction of high-quality source rocks is the key to obtaining new resources in the exploration area of Cenozoic basins in offshore China. We investigate the seismic response and area of hydrocarbon source rocks based on seismic data, well curves, lithologic interpretation, and geochemical analysis. The target is the source rock development zone of the W Formation in a survey of the South China Sea. The results show that the seismic response of thick layer source rocks differ from surrounding rocks in the seismic profile (strong reflections with opposite polarity at the top and bottom and messy or chaotic reflections inside). Seismic reflections of interlayer source rocks have the characteristics of low frequency and continuous strong amplitude. The dominant frequency and maximum amplitude decrease as the number of mudstone layers increases. Through seismic petrophysical analysis, we have obtained three sensitive parameters of source rock in this survey: clay content, P-wave impedance, and elastic impedance. We use different classification methods to realize the classification and prediction of hydrocarbon source rocks, among which the Kernel Fisher Discriminant Analysis (KFDA) method is the best. The prediction results are consistent with the geological background, geochemical information, and well curves. </jats:p>

<jats:p> Under-compaction and hydrocarbon generation are the main factors affecting pore pressure. The current seismic pore pressure prediction method is to obtain the overpressure trend by estimating the normal compaction trend (NCT) to predict the physical parameters during normal compaction and comparing the measured parameters. However, selecting a single parameter to indicate overpressure may cause insufficient consideration of factors such as hydrocarbon generation. Since hydrocarbon generation requires specific temperature and other conditions, we roughly divide the pore pressure into two parts: under-compaction in the early stage and hydrocarbon generation after reaching the hydrocarbon generation threshold. We propose a petrophysical model for estimating the normal compaction trend before hydrocarbon generation, modify the bulk modulus of the model, and use the bulk modulus method to calculate the pressure generated by under-compaction; the pressure is added to obtain the final pore pressure. In the shale gas work area in the Sichuan Basin, the prediction results are more in line with the actual situation, and the petrophysical analysis shows that the ratio of free hydrocarbon content and kerogen to water is the influencing factor indicating pore pressure. The practicality of the pore pressure prediction formula considering hydrocarbon generation in oil and gas sweet spots is illustrated through an example in the research area. </jats:p>

<jats:p> Variations in the migration velocity model directly affect the position of the imaged reflectors in the subsurface, leading to structural imaging uncertainties. These uncertainties are not explicitly addressed when trying to deterministically build an adequate velocity model. This paper presents a new stochastic geology-controlled velocity modeling method handling the possible presence of a salt weld. This permits to generate a large set of geological scenarios and associated velocity models. Each model is used to remigrate the seismic data. Then, a statistical analysis of the resulting seismic images is performed to quantify the local variability of the seismic responses. The approach is applied to the imaging of salt diapirs, in an iterative scheme (migrate, pick and update). The results show that, similarly to stacking common mid-point gathers, the statistical analysis preferentially preserves recurrent features from an image to another. In particular, this analysis permits to distinguish between connected and detached diapirs without prior knowledge about their connectivity, highlighting the potential of the method to resolve important aspects about basin and reservoir architecture. More generally, it provides quantitative information on the parts of the seismic image most sensitive to migration velocity variations, which opens interesting perspective to quantitative interpretation uncertainty assessment. Finally, the presented application also suggests that it is possible to significantly improve the quality of the generated seismic images by sampling many possible geological scenarios. </jats:p>

<jats:p> The complex of depositional, burial, and diagenetic histories of the Late Cretaceous Nezzazat Group sandstones in Northeastern Africa present the main challenges with regard to reservoir quality. The quality of commercial reservoirs is maintained despite deep burial and the associated high temperature and pressure. The study presents optimum integration of different dataset to address the reservoir quality and reservoir performance controllers. The dataset includes measured porosity and permeability, petrographic point counting data, grain size analysis, X-ray diffraction data, scanning electron microscopy and compaction porosity loss by. The depositional controls on the reservoir quality are the facies, where the higher quality found in the channel and the upper shoreface settings. The coarse-grained sandstone associated with better reservoir quality. The large intergranular porosity is the main porosity control to the fluid to flow. The massive and laminated sandstones are the best quality facies. The labile grains (feldspars and mica) control the permeability distribution. While the secondary diagenetic controllers are the carbonate cementation that inhibited the effects of compaction. The siderite cementation has resulted in a micropore dominated and highly tortuous pore system. Total porosity has largely been preserved in the siderite-cemented sample but virtually eliminated in the dolomite cemented. Low volume of illite associated with better reservoir quality. While the better reservoir quality associated with abundant quartz cementation that protected the primary porosity from compaction. Compaction act as a significant porosity loss factor during diagenesis. Authigenic kaolinite does not significantly affect the reservoir quality. The reservoir sensitivity to formation damage come from the potential for fines (kaolinite, illitic clays, siderite and pyrite) migration within the pore system that are readily to mobilize by fluid flow. </jats:p>

<jats:p> The Archie model is the foundation for calculating oil saturation, but limitations exist when the model is used to calculate oil saturation in water-flooded layer. In the process of water injection, the dynamic change in oil saturation will be caused by the different degrees of water flooding and the properties of the injected water. Under the dynamic condition of water flooding, the Archie model is not suitable for calculating the oil saturation of water flooded layers. By combining dynamic and static methods, a "double ratio" model of the same sedimentary facies layer in the later development stage was established: R<jats:sub>t</jats:sub> =  R<jats:sub>0</jats:sub>− R<jats:sub>0</jats:sub>· f( F<jats:sub>w</jats:sub>)=R<jats:sub>0</jats:sub>[1− f( F<jats:sub>w</jats:sub>)]. Based on the parameters of rock resistivity and formation water resistivity, an improved Archie model for calculating oil saturation in water flooded layers of the same sedimentary facies was established. The interpretation of the actual data of the Zhenwu Oilfield in Jiangsu, China shows that the average relative error between the calculation result and the core analysis result is 5.46%. The calculation result is reasonable, which offers a scientific basis for predicting the remaining oil distributions. The computational results have been validated by real datasets. This improved mode can provide experience-based guidance for the calculation of the remaining oil saturation of the water flooded layer in the same sedimentary interpretation layer. </jats:p>

<jats:p> Borehole image logs greatly facilitate detailed characterization of rock formations, especially for the highly heterogeneous and anisotropic carbonate rocks. However, interpreting image logs requires massive time and workforce and lacks consistency and repeatability because it relies heavily on a human interpreter's expertise, experience, and alertness. Thus, we propose to train an end-to-end deep neural network (DNN) for instant and consistent facies classification of carbonate rocks from acoustic image logs and gamma ray logs. The DNN is modified from the well-known U-Net for image segmentation. The training data are composed of two datasets: (1) manually labeled field data measured by different imaging tools from the geologically complex Brazilian pre-salt region and (2) noise-free synthetic data. Some short sections of the field data that are challenging for manual labeling due to entangled features and noises or low resolution are left unlabeled for a blind test after training. All labeled data are divided into a training set, a validation set, and a test set to avoid over-fitting. We demonstrate that the trained DNN achieves 77% classification accuracy for the test set and provides reasonable predictions for the challenging unlabeled sets. It is a great achievement given the complexity and variability of the field data. Compared with manual classification, our DNN provides more consistent and higher-resolution predictions in a highly efficient manner and thus dramatically contributes to automatic image log interpretation. </jats:p>

<jats:p> </jats:p>