Seismic Driven Reservoir Characterization and Production Management

The hybrid waveform classification introduces several additional features that aid the quality control and overcome limitations of conventional seismic attributes and unsupervised classification. These include the tools for interactive computation of waveform similarity map using semblance, and for interactive addition and subtraction of classes in the waveform map. The ability to add a user-defined class or subtract an existing class combines supervised classification with unsupervised approach to produce a hybrid classification map, greatly extend the usefulness of the tool. The key feature is to ensure that classes of prime interest, such as waveforms at well locations, are included in the analysis.

The primary aim of this study is Amplitude Versus Angle (AVA) modelling in derisking hydrocarbon prospects. AVA effects on angle gathers provide basic information on the lithology and porefill contents of the rocks under investigation. To perform the AVA modelling, a series of forward models in association with rock physics-modelled fluid-substituted logs have been developed and associated seismic responses for different pore fluids and rock types studied. The results reveal that synthetic seismic responses together with the AVA analysis show changes for different lithologies. AVA attributes analysis show trends in generated synthetic seismic responses for different fluid-substituted and porosity logs.

The integration of seismic data attributes and passive seismic (micro tremor) can provide a more convincing result about the existence of a reservoir. Micro seismic technology can be used to predict the presence of Hydrocarbon directly from the surface. So, it can be minimized the risk of drilling failure. More than that, this technology was very environmental friendly because it doesn’t need dynamite to get a data or digging a deep hole that can make a damage or social problem. Micro tremor can be used as reference to do or propose further drilling point, and would also have to be compared with the seismic data and wells in the surrounding area.

4D seismic feasibility studies prior to an eventual time-lapse seismic survey are paramount to frame the value proposition based on the geology, reservoir characteristics, production scenarios, type of acquisition, and seismic repeat interval. These considerations will have a profound impact on cost and production. From reports in the literature, it appears that actual time-lapse seismic signals rarely match their predictions. An explanation can be found in a biased rock sampling for core analysis, but also in incorrect modeling methods with the traditional feasibility studies being too simplistic and with a need to introduce additional complexity. For the latter, a remedy is found herein through coupled 3D reservoir geomechanical models and the emerging use of 3D anisotropic full-waveform modeling to generate an accurate time-lapse seismic response in the prestack domain for analysis with appropriate levels of noise added. This field-scale modeling has to take into account the overburden, sideburden, and underburden, since production-related effects radiate out from the reservoir and can have a profound impact on drilling. The value of a detailed 4D feasibility study then further informs the wider framework of closed-loop seismic reservoir monitoring where the modeled data can also be used in parallel with the actual data processing to help with model reconciliation, catering for a dynamic integrated Earth model.

Overlapping compaction trends for sand and shale lithologies combined with complex overburden geology has led to poor seismic imaging capabilities at the Mesozoic level in the Vulcan Sub-Basin, Western Timor Sea, Australia. To achieve a significant improvement in seismic data quality the broadband Sandalford 3D seismic dataset was acquired in 2012. A rock physics study conducted in 2013, using key well data from the area, showed that sand and shale trends were both partially overlapping for both AI and Vp/Vs. However, it was concluded that the two lithologies could be separated through a multi-stack inversion followed by a Bayesian lithology classification. Following the seismic processing and initial seismic interpretation the broadband seismic dataset was subject to a deterministic multi-stack inversion in 2013. This case study demonstrates how new technologies and detailed quantitative interpretation studies can significantly improve the seismic data quality and the definition of the reservoir facies distribution in an area with a history of particular seismic imaging difficulties.

This paper describe about the characterization of Low Resistivity Low Contrast (LRLC) reservoirs. Using P-Impedance (Ip), S- Impedance (Is) and Poisson Impedance (Ip-C*Is), discrimination of reservoir facies was effectively done from background non reservoir facies. Poisson's ratio and density are important in reservoir facies delineation. When the hydrocarbon bearing LRLC zones impedance contrast is very small from background trend, Poisson Impedance (PI) curve is correlated with the Gamma Ray (GR) for different values of c and it is possible to determine the maximum correlation coefficient in each case. The c value corresponding to the Maximum correlation coefficient for GR is used to compute another attribute that would emphasize lithology called lithology Impedance (LI). We can characterize these LRLC zone by cross-plotting Lithology Impedance with GR log.

Angsi field, located 170 km offshore the east coast of Peninsular Malaysia, consists of more than 20 reservoirs from I to K units in a water depth of 69 m. The major oil bearing reservoir I35, containing more than 60% of the STOOIP of the field, has been divided into three subunits I35U, I35M and I35L deposited in a depositional environment in a range from fluvial point bars to coastal plain. 3D and 4D seismic has been used to constrain a new 3D geocellular model to identify un-drained areas and support an EOR project. In this paper, the methodology followed to capture the seismic information into the static model is presented showing the impact of the approach used. Additionally, comparing this new model with the two previous existing static models, it is highlighted the implication of not using or capturing the seismic information while building the model.

The presence of karst in carbonates, as in Central Luconia Province, affects directly their volumetric, flow properties and drilling. Using broadband emphasized pre-processing and advanced multiple elimination methods, the reprocessing of the 3D seismic in the present study area has resulted in significant improvement in image quality, allowing the interpretation of 32 new faults. The interpretation of these faults has considerably improved reservoir characterization, namely: 1) a new volume based modelling technique allowed the integration of all these faults in the static model, preserving the structural complexity, and 2) a karst conduit modelling was performed along these faults, using object-distance simulation method, and integrated in the static model. This seismic-driven reservoir characterization could well be a key success factor for future field development and well target optimization.

A multidisciplinary reservoir study of an offshore field in ultra-deep water environment demonstrates how effectively seismic can be integrated within a reservoir modelling workflow to reveal the hidden petrophysical properties. This workflow consists of well data modelling, statistical rock physics modelling, probabilistic petrophysical seismic inversion and geological modelling, and allows to manage reservoir uncertainty in a scenario of few wells. A log-facies model based on petro-elastic characterization is used as common input to seismic inversion and geostatistical simulation processes. A rock physics model, which is later embedded in the seismic inversion, links petrophysics and impedances in a statistical way on a log-facies basis. The conditional probability of petrophysics given the seismic is obtained by means of seismic inversion and rock physics model, which finally provides, together with cluster analysis, the conditional probability of the facies given the seismic. All the uncertainties are handled and propagated to the 3D geological model. Seismic facies probabilities are directly integrated as a prior trend into geostatistical simulations of reservoir properties in order to recover, from the well data, the high frequency part of signal, which is not carried from seismic.