الفهرس 
Geologic SettingOnly 14 pages are availabe for public view
 |  | from | 344 |  |  |
| | | from | 344 | | |
Abstract
This study is focused on the Sienna Field. It is a gas field in the southern portion of the West Delta Deep Marine (WDDM) concession, approximately 120 km Northeast of Alexandria, Egypt.
The main purpose of this study is how to discriminate between fluid and lithology by using well logging data with different seismic interpretation methods as AVO modeling, Extended Elastic Impedance and Artificial Neural Network.
This work is represented in seven chapters:
Chapter One discussed the location of the study area, objectives of the present study, available data for the study, methodology and exploration history of Sienna field.
Chapter Two debated the geologic setting of the West Delta Deep Marine (WDDM), including detailed discussion of subsurface stratigraphy and the regional stratigraphic column of (WDDM), structural settings and tectonic framework of (WDDM), where the study area is located, especially Sienna Field. It is based on previously published literatures and the available subsurface geological data.
Chapter Three Figured out the well logging analysis and formation evaluation of the productive zones to determine the petrophysical properties and the reservoirs characteristics for the Sienna channel rocks. The well log data analysis has been carried out using Tech Log® Schlmberger Software.
In this study, the petrophysical characterizations of the subsurface sandstone Plio-Pleistocene reservoir in the offshore Nile Delta Sienna field has been defined based on the well log data available in three wells. The detailed petrophysical analysis revealed the presence of a gas-bearing sandstone interval ranged from 1 to 12 m of net pay zone with good reservoir characteristics in terms of good porosity (26%), low shale volume (16%) and water saturation (66%).
The lithology of the Sienna reservoir rock was studied using NPHI – RHOB and PEFZ-RHOB cross-plots. These plots show that the reservoir rock is mainly composed of sandstone with considerable amount of shale and silt, the effect of the gas shift points to upward while the effect of the shale shift points downwards.
The clay must be identified due to their great effect on the hydrocarbon reservoir evaluation. The Thorium (Th) – Potassium (K), and Photoelectrical effect (PEF) -TH/K cross-plots showing the type of clay minerals in different sand bodies in each of the studied wells. The mineralogy of the Sienna reservoir rock lies between mixed layer clay, mica, illite, montmorillonite and glauconite. The Sienna reservoir was deposited mostly in fluvial to shallow marine environments according to the presence of these clay minerals.
The analysis of pressure data is concerned mainly with locating the different fluid contacts and determining the pressure gradients of the gas-bearing zone. Very close pressure regimes are detected for the investigated gas anomaly throughout the study area. Pressure gradient of gas ranges from 0.078 to 0.084 psi/ft while, the water gradient is 0.46 psi/ft.
Differences in relaxation times of NMR data are used to differentiate between irreducible water, and movable water, gas depending on the T2 cutoff value. NMR-log data also provide indication about porosity according to the amplitude of the decay curve. NMR reveals that the pay zones (Sand-1, Sand-2, and Sand-3) have good petrophysical parameters for production, free fluid with good effective porosity and good permeability.
Chapter Four discussed AVO modeling and inversion which carried out though many steps; first, AVO analysis then the initial low frequency model constructing using well log data and horizons, finally making the simultaneous pre-stack inversion which is produced six volumes of P-velocity (Vp), S-Velocity (Vs), and density (ρ), P-impedance, s-impedance, and Vp/vs ratio. In addition, this chapter displayed AVO attributes combination to produce fluid and lithology stacks.
Chapter Five figured out the Rock physics models to create the bridge between elastic properties (density, VP/VS ratio, bulk modulus, Shear modulus, young modulus impedance and Lame’s parameter) and reservoir properties (porosity, permeability, and saturation) to guide for lithology and fluid discrimination.
Chapter six discussed how can discriminate between fluid and lithology content using Extended Elastic Impedance (EEI) to highlight the difference between reservoir and non-reservoir and identify hydrocarbon zones.
The general workflow of EEI is to define, from well data, chi angles for each petrophysical parameter as Vsh, Sw, ……. and then derive EEI reflectivity volumes for each parameter at its chi angle (by combining (AI) acoustic impedance and Gradient impedance (GI).This process often shows that high correlations to certain petrophysical curves occur at very similar angles, making it difficult to separate out different properties. In addition to this, the well data that are fed into this process can often be biased (e.g., with a majority of wells being hydrocarbon bearing) and can therefore give rise to results that are skewed (particularly important when correlating water saturation curves). Therefore, a second complementary approach, which uses a facies-based scheme is often more instructive, and directly relates log-based facies to seismic properties. The objective now becomes that of defining chi angles at which key facies cross over.
Chapter Seven discussed the Neural Networks; NN were used to predict many petrophysical properties as consisting of the following steps:
1. Three wells that contain gas were used as sample wells in forming the training and testing sets of the neural networks.
2. The training and testing of the neural networks is used different single and multi-attribute.
3. The multi-attribute that gave the best or highest training and testing neural networks performance was used to predict petrophysical properties distribution in Study area.
4. Extraction volume of petrophysical properties for the study area from seismic and inverted volume.
Emerge is a powerful geostatistical module from Hampson Russell of CGG Geo Software. It can predict log property volumes from well logs and attributes from seismic data. Besides applying Emerge to predict a volume of log property, Emerge can increase the resolution of the inversion results. This presentation describes a series of advanced options that can be used in order to improve Emerge training and obtain more accurate and better final predicted results.
By defining the NN layers (input data, output, and Neural network structure) for each predicted property; The main input to the neural networks is well log data for many petrophysical properties as density, porosity, saturation, and elastic properties (K, E, Vp/Vs, possion’s ratio, ………etc.). PNN is used for Prediction and the output is the volume for each petrophysical property.
The low values of training error and validation error make reliable prediction of physical and seismic rock properties. It suggests that reservoir properties can be estimated from seismic data using neural network analysis and thereby helps in better understanding of the lateral variations of reservoir properties away from the wells. Probabilistic neural network (PNN) predicted results show lateral variation of physical and seismic rock properties in reservoir and non-reservoir parts under the study area. High porosity, low Vsh, low density, low water saturation, high Young modulus, high shear, low bulk modulus, high Mu-Rho, low Lambda-Rho, low Vp/Vs ratio, low possion’s ratio, and good permeability are traced corresponds to Sienna channel which indicated to good pay zone with gas content.
Last step of this study is making lithofacies distribution volume by using the volumes of Lambda-Rho and Mu-Rho with the cross-plot facies clusters. The gas-sand, brine sand and shale distribution probability volumes are the final output which give the accurate picture of the distribution of gas inside the channel, which applies to the places of producing wells and predicted a new area in the western part of the study area.