Biot's equations-based reservoir parameter inversion using deep neural networks

Reservoir parameter inversion from seismic data is an important issue in rock physics. The traditional optimisation-based inversion method requires high computational expense, and the process exhibits subjectivity due to the nonuniqueness of generated solutions. This study proposes a deep neural network (DNN)-based approach as a new means to analyse the sensitivity of seismic attributes to basic rock-physics parameters and then realise fast parameter inversion. First, synthetic data of inputs (reservoir properties) and outputs (seismic attributes) are generated using Biot's equations. Then, a forward DNN model is trained to carry out a sensitivity analysis. One can in turn investigate the influence of each rock-physics parameter on the seismic attributes calculated by Biot's equations, and the method can also be used to estimate and evaluate the accuracy of parameter inversion. Finally, DNNs are applied to parameter inversion. Different scenarios are designed to study the inversion accuracy of porosity, bulk and shear moduli of a rock matrix considering that the input quantities are different. It is found that the inversion of porosity is relatively easy and accurate, while more information is needed to make the inversion more accurate for bulk and shear moduli. From the presented results, the new approach makes it possible to realise accurate and pointwise inverse modelling with high efficiency for actual data interpretation and analysis.

Application of Micro-resistivity Scanning Imaging Logging Data in Reservoir Thickness Division

Reservoir parameter interpretation is one of the main contents of reservoir description, which affects the whole process of oilfield development. According to the characteristics of micro-resistivity scanning imaging logging, which can directly reflect the changes of lithology and physical properties of reservoirs, this paper compares the thickness and interbed division of reservoirs with conventional logging data, this paper finds out the shortcomings of the conventional logging data in the interpretation of thickness and the division of interlayers, and combines the core analysis data to examine the differences in the correlation on the coring wells, and obtains good results, it has laid the foundation for the establishment of new interpretation procedure.

Spectral-ratio tomography for seismic attenuation estimation

When measuring surface seismic data, an accurate attenuation estimation method is necessary to compensate for the energy loss and phase distortion of seismic waves, and is also beneficial for further quantitative amplitude analyses and reservoir parameter predictions. For conventional Q-estimation methods (such as the log spectral-ratio (LSR) method and attenuated traveltime tomography), accuracy may be affected by the differences between the overburden ray paths of two selected reflections (we call it the overburden effect). In this study, we design a more accurate Q-tomography method to estimate Q-values (both in the overburden and target layer simultaneously) without overburden assumptions. We address the overburden effect by using an inversion method, which allows us to separate attenuation effects from the overburden through the traveltime differences in the tomography grid cells. We test the method on synthetic data and prove its feasibility and effectiveness by applying it to field data.

EFFECT OF CO2 REACH ENVIRONMENT ON THE PETROPHYSICAL PROPERTIES OF ARTIFICIALLY CONSOLIDATED CORE SAMPLES

In order to develop, maintain and deplete reservoirs economically around the globe, various measurements are needed with a high demand on natural core samples. The next stage in the life of every reservoir is a secondary or tertiary method to enhance productivity. However, to tailor the available methods and technologies to the reservoir, several screening processes, feasibility studies and pilot experiments are needed. As an aid to these, like a sensitivity analysis, continuous measurements are set up to study fluid flow, chemical reactions, additional recovery and much more, but for all of these, core samples are needed. The lack and high value of natural core samples yield that the demand cannot be satisfied from this source alone. The aim of the study was to create an artificially consolidated stone core sample, a model material, which can be suitable for being the subject of these experiments, with additional benefits in mass production and reservoir parameter-based quality control. In this article the authors wish to present partial results of a big study, this time with comparing the porosity, permeability, connate water and capillary pressure parameters of the core samples used with different after-cure techniques. The process of compaction was the same, but the overburden pressures and the effect of CO2 rich curing were examined. For this, part of the samples was prone to high CO2 environment for different timespans during the after treatment of the samples. The petrophysical parameters were then measured on all of the groups, including a control group and the CO2 affected cores. The focus was on porosity, permeability, connate water saturation/wettability and capillary pressure measurements and the common features and differences in the yielded pore space’s structure are summarized in this article.

Regionalization of Reservoir regulation parameters using physiographic and climatological predictors

<p>Present regional and global scale hydrology has to account for man-made reservoirs that impart significant regulation signature into the downstream streamflow regime. Optimization of domains with large number of reservoirs would incur multitude of reservoir regulation parameters. Such parameter-set-per-reservoir approach not only results in excessive computational costs but also, by principle, lacks effective constraining of the parameter space. We propose an approach to derive single set of parameters for all the reservoirs and lakes in the modelling domain. The hypothesis is that reservoir regulation parameters can be regionalized using physiography and climatology at lakes and their catchments.<br><br>To test this hypothesis, we setup a modeling domain for the São Francisco basin of Northeast Brazil in the mesoscale hydrological model (mHM, www.ufz.de/mhm). The domain consists of climatology ranging from tropical (As) to semi-arid (BSh) and reservoirs with catchment area varying from less than 500 km<sup>2</sup> to greater than 500,000 km<sup>2</sup>. We carried out correlation analysis between selected physiographical and climatological predictors and the reservoir parameters of the multiscale lake module, mLM, of the mHM model (https://presentations.copernicus.org/EGU2020/EGU2020-6047_presentation.pdf). For an instance, the reservoir rule curves in mLM are estimated based on inflow and position of water level. The predictors here are inflow and water level which are normalized using catchment area and the shape of the reservoir, respectively. Similarly, the timing and shape parameters of rule curves were plotted against the climatological characteristics of the upstream catchment. The preliminary results reveal significant trends between the mLM parameters and the normalized predictors. These mathematical relationships, better known as transfer functions, can now be used to generate a single global reservoir parameter set.</p><p>The demonstrated hypothesis helps to optimize regulated hydrology using a single parameter set, irrespective of size, location and inherent climatology of reservoirs involved. This is inline with the pre-existing paradigm of multiscale parameter regionalization (MPR) of mHM. The findings contribute to the contemporary effort of hydrological modeling society towards improved global scale hydrological modeling.</p>

Predicting the gas resource potential in reservoir C-sand interval of Lower Goru Formation, Middle Indus Basin, Pakistan

The integrated study of seismic attributes and inversion analysis can provide a better understanding for predicting the hydrocarbon-bearing zones even in extreme heterogeneous reservoirs. This study aims to delineate and characterize the gas saturated zone within the reservoir (Cretaceous C-sand) interval of Sawan gas field, Middle Indus Basin, Pakistan. The hydrocarbon bearing zone is well identified through the seismic attribute analysis along a sand channel. The sparse-spike inversion analysis has efficiently captured the variations in reservoir parameter (P-impedance) for gas prospect. Inversion results indicated that the relatively lower P-impedance values are encountered along the predicted sand channel. To further characterize the reservoir, geostatistical techniques comprising multiattribute regression and probabilistic neural network (PNN) analysis are applied to predict the effective porosity of reservoir. Comparatively, the PNN analysis predicted the targeted property more efficiently and applied its estimations on entire seismic volume. Furthermore, the geostatistical estimations of PNN analysis significantly predicted the gas-bearing zones and confirmed the sand channel as a major contributor of gas accumulation in the area. These estimates are in appropriate agreement with each other, and the workflow adopted here can be applied to various South Asian regions and in other parts of the world for improved characterization of gas reservoirs.