Model-Based Inversion in North Sea f3-Block Dutch Sector

2015

MODEL-BASED INVERSION IN NORTH SEA F3-BLOCK DUTCH SECTOR By

Hassan Hussien Hassan Khalid El-Maghawry Wafaa Mohamed Hassan B.Sc. Students 2015

To

Geophysics Department Faculty of Science Ain Shams University

Supervised by

Dr. Azza Mahmoud Abd El-Latif El-Rawy Lecturer of Geophysics Geophysics Department- Faculty of Science- Ain Shams University Cairo - 2015

Acknowledgement

We would like to express our gratitude and many thanks to our supervisor

Dr.Azza

El-Rawy   for

her

help,

guidance,

patient,

understanding, continuous support and efforts with us to complete this work. We are also thankful to Mr.Ahmad Hosny from Rashid Petroleum Company for his help and advices. Finally, we would like to appreciate all department staff members they always had time for question and discussion.

I|Page

ABSTRACT

The location of project is in North Sea in Dutch sector where a 3D seismic acquisition. As determination of lithology and fluid content distribution is a desirable objective for reservoir characterization and subsequent reservoir management, we adopt AVO inversion and poststack impedance inversion methods to achieve this goal. Due to the access to available inversion methods, it is advisable on finding out the best method. To this end, we present a comparative performance of the available model based seismic acoustic impedance inversion methods. In this study, we start with the traditional model based seismic acoustic impedance inversion and compare its result with equivalent results from elastic impedance, the separate or independent impedance inversion carried out as a post stack process and simultaneous inversion.

II | P a g e

Acknowledgement  ........................................................................................................................................ I ABSTRACT  ................................................................................................................................................. II Contents  ....................................................................................................................................................... III

List of Figures  .............................................................................................................................................. V 1.

2.

3.

INTRODUCTION  ............................................................................................................................... 1 1.1.

Topographic location  .................................................................................................................. 1

1.2.

Objective of inversion  ................................................................................................................. 2

1.3.

Available data  .............................................................................................................................. 2

1.4.

Methodology  ................................................................................................................................ 5

GENERAL GEOLOGICAL SETTINGS  ......................................................................................... 6 2.1.

Introduction  ................................................................................................................................. 6

2.2.

Stratigraphy  ................................................................................................................................. 7

2.2.1.

Paleozoic (about 590 - 250 million years ago)  ...................................................................... 7

2.2.2.

Mesozoic  ............................................................................................................................... 8

2.2.3.

Cenozoic   ............................................................................................................................. 11

2.3.

Petroleum system  ...................................................................................................................... 11

2.4.

Structural setting  ...................................................................................................................... 15

2.5.

Tectonic history  ......................................................................................................................... 20

THEORITICAL BACKGROUND OF SEISMIC AMPLITUDE INVERSION ........................ 22 3.1.

3.1.1.

Seismic velocity inversion   .................................................................................................. 22

3.1.2.

Seismic Amplitude Inversion  .............................................................................................. 23

3.2.

4.

Introduction  ............................................................................................................................... 22

Post-stack Inversion Methods  .................................................................................................. 33

3.2.1.

Colored inversion  ................................................................................................................ 33

3.2.2.

Sparse spike inversion ......................................................................................................... 34

3.2.3.

Model-based inversion  ........................................................................................................ 36

3.2.4.

Stochastic inversion   ............................................................................................................ 40

SEISMIC INVERSION USING HAMPSON-RUSSELL SOFTWARE.................................. ................. ..................... .... 42 4.1.

Introduction  ............................................................................................................................... 42

4.2.

Available Data  ........................................................................................................................... 43

III | P a g e

4.3.

Inversion Concepts  .................................................................................................................... 44

4.4.

Model-Based Inversion Workflow on F3-Block Seismic Cube ............................................. 47

4.4.1.

Data Importing  .................................................................................................................... 47

4.4.2.

Identifying Target Reservoir in F3-Block Cube  ................................................................. 52

4.4.3.

Depth-Time Conversion and Check Shot Correction  ......................................................... 57

4.4.4.

Wavelet Extraction and Synesthetic Trace   ......................................................................... 59

4.4.5.

Initial Model Building ......................................................................................................... 62

4.4.6.

Quality Control (QC)  .......................................................................................................... 64

4.4.7.

Running Inversion  ............................................................................................................... 67

4.4.8.

Results  ................................................................................................................................. 68

Summary  .................................................................................................................................................... 71 References  .................................................................................................................................................. 73

IV | P a g e

List of Figures Figure 1.1: Location map of Dutch Sector, showing F3 Block noted above (from Remmelts, 1995) -------- 3 Figure 2.2: stratigraphic column (modified from Glennie, 1997a) --------------------------------------------------10 Figure 3.2: Geologic Provinces Total Petroleum System (403601). ------------------------------------------------- 13 Figure 4.2: Petroleum System Glennie (1998)----------------------------------------------------------------------------14 Figure 5.2: structure features of North Sea (Ziegler, 1990) -----------------------------------------------------------18 Figure 6.2: Geological cross section ziegler (1978) ---------------------------------------------------------------------- 19 Figure 7.3: velocity inversion R. Kamei, R.G. Pratt and T. Tsuji-------------------------------------------------------22 Figure 8.3: Modeling and Inversion Rasmussen KB Brunn A and Pedersen JM (2004) -------------------------26 Figure 9.3: Porosity vs Impedance, Salter R Et al (2005) --------------------------------------------------------------- 27 Figure 10.3: Porosity Prediction, Salter R etal (2005) ------------------------------------------------------------------ 28 Figure 11.3: Seismic Inversion ------------------------------------------------------------------------------------------------ 30 Figure 12.3: Synthetic trace and wavelet extraction Buiting,J . J. M. & Bacon,M . (1999) -------------------- 32 Figure 13.3: Colored inversion vs Seismic data (Lancaster and Whitcombe 2000) ----------------------------- 34 Figure 14.3: Sparse spike inversion (modified after Ronghe and Surarat 2002). -------------------------------- 36 Figure 15.3: Model based inversion (Goffe et al.1994, Duboz et al. 1998)---------------------------------------- 38 Figure 16.3: Correlation between check shots and impedance log veeken 2006 ------------------------------- 38 Figure 17.3: Seismic vs Impedance, Veeken 2006 ----------------------------------------------------------------------- 39 Figure 18.3: Stochastic inversion (modified after TorresVerdin et al. 1999). ------------------------------------- 41 Figure 19.4: combination between the geology and seismic data in inversion ---------------------------------- 44 Figure 20.4: Seismic Inversion Components ------------------------------------------------------------------------------ 45 Figure 21.4: Step 1 -------------------------------------------------------------------------------------------------------------- 46 Figure 22.4: Step 2 -------------------------------------------------------------------------------------------------------------- 46 Figure 23.4: Comparison between Input Seismic and Model Based Inversion ----------------------------------- 47 Figure 25.4: Opened Database List ------------------------------------------------------------------------------------------ 47 Figure 24.4: HR Opened Database List ------------------------------------------------------------------------------------- 47 Figure 26.4: Hampson-Russell Interface ----------------------------------------------------------------------------------- 48 Figure 27.4: Well Explorer Window ----------------------------------------------------------------------------------------- 48 Figure 28.4: Importing Logs ---------------------------------------------------------------------------------------------------48 Figure 29.4: Identifying unknown logs ------------------------------------------------------------------------------------- 49 Figure 30.4: STRATA Project Selection Mode----------------------------------------------------------------------------- 49 Figure 31.4: Importing Data---------------------------------------------------------------------------------------------------50 Figure 32.4: SEG-Y Seismic File Open --------------------------------------------------------------------------------------- 51 Figure 33.4: Geometry Grid Page -------------------------------------------------------------------------------------------- 51 Figure 34.4: Well to Seismic Map Menu ----------------------------------------------------------------------------------- 52 Figure 35.4: GR-Porosity Cross plot ----------------------------------------------------------------------------------------- 53 Figure 36.4: Well F020-1 Cross Section ------------------------------------------------------------------------------------ 53

V|Page

Figure 37.4: GR-Impedance Cross Plot ------------------------------------------------------------------------------------- 54 Figure 38.4: F02-2 Cross section for GR-Impedance cross plot ------------------------------------------------------ 55 Figure 39.4: FS8 the target reservoir --------------------------------------------------------------------------------------- 55 Figure 40.4: Picking Horizon FS8 --------------------------------------------------------------------------------------------- 56 Figure 41.4: Actual Picks of FS8 ---------------------------------------------------------------------------------------------- 56 Figure 42.4: Difference between Time logs and Depth logs ---------------------------------------------------------- 57 Figure 43.4: Depth time table ------------------------------------------------------------------------------------------------ 58 Figure 44.4: Check Shot Correction ----------------------------------------------------------------------------------------- 59 Figure 45.4: Wavelet with different phases (0, 90,180 and 270) ---------------------------------------------------- 60 Figure 46.4: Wavelet 270 correlation in well F02-1 --------------------------------------------------------------------- 61 Figure 47.4: Wavelet zero correlation in well F03-4 -------------------------------------------------------------------- 61 Figure 48.4: Initial Model Building ------------------------------------------------------------------------------------------ 62 Figure 49.4 ------------------------------------------------------------------------------------------------------------------------ 62 Figure 50.4 ------------------------------------------------------------------------------------------------------------------------ 62 Figure 51.4 ------------------------------------------------------------------------------------------------------------------------ 63 Figure 52.4: Initial Model for F3-Block Seismic Cube ------------------------------------------------------------------- 63 Figure 53.4: Inversion Analysis for Post-stack Data --------------------------------------------------------------------- 64 Figure 54.4: Number of Iteration --------------------------------------------------------------------------------------------65 Figure 55.4: Error Plot ---------------------------------------------------------------------------------------------------------- 66 Figure 56.4: Inversion Analysis Cross plot --------------------------------------------------------------------------------- 66 Figure 57.4: Inverted Inline ---------------------------------------------------------------------------------------------------67 Figure 58.4: Inverted Xline ---------------------------------------------------------------------------------------------------- 67 Figure 59.4: Inverted Arbitrary Line----------------------------------------------------------------------------------------- 68 Figure 60.4: Reflectivity (Right) Impedance (Left) Slice-750 (upper) Slice-780 (lower) ------------------------69 Figure 61.4: Reflectivity (Right) Impedance (Left) Slice-950 (upper) Slice-1600 (Lower) ---------------------- 70

VI | P a g e

CHAPTER 1: 1. INTRODUCTION 1.1.

Topographic location

The North Sea is a marginal, shallow sea on the European continental shelf. It is more than 970 kilometers from north to south and 580 kilometers from east to west, with an area of around 750,000 square kilometers. The North Sea RAC area is larger, because it includes the Skagerrak and Kattegat which connect the North Sea  proper to the Baltic. The North Sea is bordered by England, Scotland, Norway, Denmark, Germany, the Netherlands, Belgium and France. In the southwest, beyond the Straits of Dover, the North Sea becomes the English Channel which connects to the Atlantic Ocean. The North Sea is a fairly shallow coastal sea and depths in the southern basin do not exceed 50m. The northern areas are deeper but are still generally less than 200m except in the Norwegian Trough, in the north -east, which is the only region of very deep water. The Dutch part of the North Sea (NCP) occupies an area of 57,000 km2 which is larger than the mainland of The Netherlands. In addition to traditional activities, like fishery and shipping, since the 1960s there have been new activities such as oil and gas exploration and exploitation, and extraction of sand from the seabed. New  plans for installation of wind turbines and coastal extensions are in progress. In addition the area forms part of a valuable marine ecosystem. Knowledge of the morphology of the seafloor and the sediment properties is important, both to understand the ecosystem as well as for developing infrastructural activities.

1|Page

1.2.

Objective of inversion

The principle objective of seismic inversion is to transform seismic reflection data into a quantitative rock property, descriptive of the reservoir. In its most simple form, acoustic impedance logs are computed at each CMP... In other words, if we had drill and logged wells at the CMP’s, what would the impedance logs have looked like? Compared to working with seismic amplitudes, inversion results show higher resolution and support more accurate interpretations. This in turn facilitates better estimations of reservoir properties such as porosity and net pay. An additional  benefit is that interpretation efficiency is greatly improved, more than offsetting the time spent in the inversion process. It will also be demonstrated below that inversions make possible the formal estimation of uncertainty and risk 1.3.

Available data

The project is a prospect in North Sea specified in Dutch sector place offshore  Netherlands F3 Block (Figure 1.1) with coordinates N 54°52′0.86″ / E 4°48′47.07 this project was held in 1987 and that by making 3d seismic survey and drilling of 4 wells in different area to cover our cube which has 24*16 km square our target is at late Jurassic early cretaceous for very rich gas prospect. The available data is: Seismic Data: •

A 3D seismic cube that consists of 740 in lines and 949 xlines and it is post

stack cube

2|Page

Figure 1.1: Location map of Dutch Sector, showing F3 Blo ck noted above (from Remmelts, 1995)

3|Page

Horizons •

Demo0 (FS4)

•

Demo1 (MFS4)

•

Demo2 (FS6)

•

Demo3 (Top_Foresets)

•

Demo4 (Truncation)

•

Demo5 (FS7)

•

Demo6 (FS8)

•

Demo7 (Shallow)

Wells and available logs •

F02 – 1 (Caliper, Density, GR, P-Wave, Porosity)

•

F03 – 2 (Density, GR, P-Wave, Porosity)

•

F03 – 4 (Density, GR, P-Wave, Porosity)

•

F06 – 1 (Density, GR, P-Wave, Porosity)

4|Page

1.4.

Methodology

The 3D seismic cube is inputted in the Hampson-Russell software then we start to input the wells in the area and recognize the logs in each wells then start to input the tops and the check shots after that we start to input the seismic data as cube and the horizons then start to extract wavelet to get synthetic seismogram and correlate it with seismic after that the selected wavelet is used to make model based inversion to our cube to get inverted one. Benefits of seismic inversion The benefits of seismic inversion include •

Compensation for and reduction of the effects of wavelet tuning

•

Presentation of output as geologic layers rather than reflection edges

•

Merging of known low frequency geologic and geophysical information with

seismic data •

Modelling and inclusion of layer stratigraphy

•

Inclusion of geophysical constraints from known information or analogues.

•

Calibration to well log data

•

Improved interpretability of seismic horizons

•

Increased bandwidth in the inversion output

•

Attenuation of random noise

5|Page

CHAPTER 2: 2. GENERAL GEOLOGICAL SETTINGS 2.1.

Introduction

The project is a prospect in North Sea specified in Dutch sector place offshore  Netherlands with coordinates N 54°52′0.86″ / E 4°48′47.07 this project was held in 1987 and that by making 3d seismic survey and drilling of 4 wells in different area to cover our cube which has 24*16 km square. Our target is at late Jurassic early cretaceous for very rich gas prospect. To be reliable, Earth models used for mineral exploration should be consistent with all available geologic and geophysical information. Due to data uncertainty and other aspects inherent to the underdetermined geophysical inverse problem, there are an infinite number of models that can fit the geophysical data to the desired degree (i.e. the problem is no unique). Additional information is essential to obtain a unique and useful solution. Incorporating prior geologic knowledge, and combining several complimentary types of geophysical data collected over th e same Earth region, can reduce ambiguity and enhance inversion results, leading to more reliable Earth models. There are two areas of research that are important to help achieve the goal of more reliable Earth models: development of geophysical inversion methods that 1) Increase the kinds of geologic information that can be incorporated 2) Can combine several complimentary types of geophysical data collected over the same Earth region

6|Page

2.2.

Stratigraphy

2.2.1. Paleozoic (about 590 - 250 million years ago) The configuration of Lower Paleozoic crystalline and metamorphic basement rocks that underlie the North Sea sedimentary basins was assembled during the Caledonian Orogeny (about 420 - 390 million years ago) to form the Caledonian  basement. This was achieved through the closure of the Iapetus Ocean and the Tornquist Sea, at the Iapetus Suture and ‘Trans-European Fault Zone’, respectively (e.g. Andrews et al. 1990; Johnson et al. 1993; Gatliff et al. 1994; Glennie and Underhill 1998). Many of the major faults within the Caledonian basement formed lines of weakness that experienced significant reactivation during subsequent phases of earth movements. During the Devonian (about 410 -360 million years ago), there was widespread red-bed molasse and lacustrine sedimentation as the newly-formed Caledonian mountain ranges were eroded. Mid-Devonian (about 375 million years ago) marine limestones in the south of the Central North Sea were probably formed during an early rift phase. This was a precursor to the main phases of Permo-Triassic (about 290 - 210 million years ago) and Late Jurassic rifting (about 160 - 140 million years ago) and associated strike-slip movements. During the early Carboniferous (about 360 - 325 million years ago), fluviodeltaic and shallow-marine sediments and local volcanic accumulated in parts of the Central North Sea at times of regional crustal extension, though the Northern  North Sea area was mainly source of clastic sediments. As in England, these Carboniferous rocks were gently folded, faulted, uplifted and eroded du ring the Late Carboniferous Variscan Orogeny approximately 300-290 million years ago.

7|Page

During the Late Permian (about 270 - 250 million years ago) redbeds and local volcanic (Rotliegend Group) accumulated within the widespread Northern P ermian Basin. Following marine transgression, cyclical evaporitic successions (Zechstein Group) were deposited and locally reach over 1000 m in thickness. The evaporites have been deformed by halokinesis intermittently since mid-Triassic times (about 230 million years ago), leading to the widespread growth of salt pillows and salt diapirs, especially in the Central North Sea 2.2.2. Mesozoic For the ages of the geological systems and events in this section the reader is referred to (Figure 2.1) In the Triassic there was a return to arid, continental climate conditions and both sandstone- and mudstone-dominated redbed successions were laid down. During the Early Jurassic there was a spread of marine deposits over much of the North Sea during a phase of thermal subsidence following PermoTriassic rifting. During the Middle Jurassic, regressive, paralic sediments accumulated when a major subaerial thermal dome formed within the Central North Sea probably due to the development of a warm, diffuse and transient mantle-plume head (Underhill and Partington 1993). The Late Jurassic was a time of major extensional faulting. The rifting was initially most intense at the extremities of the present graben system and as time elapsed it propagated back towards the center of the dome (Rattey and Hayward 1993; Fraser 1993). The onset of major rifting probably occurring in the middle Oxfordian to early Kimmeridgian (approximately 157 - 155 million years ago) (Underhill 1991; Glennie and Underhill 1998). Seismic data reveal that the Upper Jurassic sedimentary successions commonly thicken dramatically towards syndepositional faults. This pattern of

8|Page

sediment thickness variation is in contrast with that formed during the ‘thermal sag’  phase of basin development (e.g. McKenzie 1978) in Early-mid Jurassic times, when the basin was more ‘saucer -shaped’ and the thickest deposits accumulated at its center. Rift styles vary substantially between the northern and the central North Sea and there were two principal controlling factors. Firstly, differences in the basement composition and tectonic grain between the two regions strongly influenced structural development. In the central North Sea, the rifts are more complex and were segmented along NE ‘Caledonide’ and NW ‘Trans- European Fault Zone’ trends (e.g. Errat et al. 1999; Jones et al. 1999). Secondly, in the northern North Sea, Upper Permian salt is largely absent, and there is no major detachment between basement and cover rocks. In contrast, the Zechstein evaporites in the central North Sea  provide a major detachment level that essentially separates the basement rocks from the cover sequence of rocks or ‘carapace’ (e.g. Hodgson et al. 1992; Smith et al. 1993; Helgeson 1999). This structural contrast is reflected in the smaller size of the oil a nd gas fields discovered within the pre- and syn-rift successions of the central North Sea. Local inversion of central North Sea depocentres during the Early Cretaceous is considered to be a response to strike-slip faulting (Pegrum and Ljones 1984). Transpressional  pulses are believed to have triggered the halokinesis of Zechstein salts within the Central Graben, which exerted an additional control on the patterns of subsidence and sedimentation (Oakman and Partington 1998; Gatliff et al. 1994).

9|Page

Figure 2.2: stratigraphic column (modified from Glennie, 1997a)

10 | P a g e

2.2.3. Cenozoic Thermal subsidence in response to Late Jurassic rifting, dominated much of the Cenozoic, with some relatively minor pulses of earth movements (e.g. Pegrum and Ljones 1984). Regional patterns of sedimentation changed dramatically in early Paleocene times, with the influx into the basinal areas of huge volumes of coarse clastic detritus including debris flows and turbidities. This detritus was shed from the uplands of northern Scotland and the Orkney-Shetland Platform, which were undergoing thermal uplift in response to the development of the Iceland Plume (White 1988; White and Lovell 1997). 2.3.

Petroleum system

Upper Jurassic syn-rift, organic-rich marine mudstones (the Kimmeridge Clay Formation) provide the source material for most of the region’s hydrocarbons (Brooks et al. in press). Cretaceous and Cenozoic post-rift thermal subsidence and  burial has enabled the source rocks to become mature for hydrocarbon generation along the rift axes from Paleocene times onward (Johnson and Fisher 1998). Hydrocarbon migration has been mainly vertical, but with significant lateral migration restricted to the Upper Jurassic and Paleocene successions. Hydrocarbons extraction is from almost every clastic and carbonate sedimentary succession, ranging in age from, and including, Devonian and Eocene strata. Pre-rift producing fields comprise Paleozoic, Triassic to Lower Jurassic and Middle Jurassic categories (Brooks et al. in press). The Middle Jurassic tilted fault-block play is best developed in the East Shetland Basin and is one of the most productive in the North Sea. Syn-rift reservoirs within producing fields comprise Upper Jurassic to Lower Cretaceous sandstones according to Brooks et al. (in press), though many authors prefer to interpret the

11 | P a g e

Lower Cretaceous succession as post-rift deposits formed during a phase of local strike-slip tectonism (e.g. Rattey and Hayward 1993; Oakman and Partington 1998). The producing Upper Jurassic reservoirs include both shallow and deep marine sandstones, though Lower Cretaceous reservoirs were almost exclusively formed within deep marine settings. The syn-rift hydrocarbons producing fields display a wide variety of trapping mechanisms, including tilted fault blocks, domes, and stratigraphic closures. Thick, post-rift Lower Cretaceous mudstones also provide a regional seal for many traps. Post-rift thermal subsidence has continued from Cretaceous times to the present day. Within the UK sector, the Upper Cretaceous Chalk is relatively insignificant as a producing reservoir. Mass-flow sandstone reservoirs of Paleocene age are estimated to contain about 20% of the oil province’s proven hydrocarbon reserves (Pegrum and Spencer 1990). Virtually all of the UK sector Paleocene sand systems become progressively distal to the east or SE. There was an evolution from the emplacement of laterally extensive sheet sands on the basin floor during the early Paleocene, to restriction of sand bodies into narrow, elongate channels intercalated within mud-dominated slope facies during the mid-Eocene. The Central North Sea and includes many important producing hydrocarbon fields. These fields mainly produce from syn-rift Upper Jurassic and/or post-rift Lower Cretaceous or Paleocene reservoir sandstones. Halokinesis has generally not exerted a major influence on the structural development. Many of the Upper Jurassic and some of the Lower Cretaceous hydrocarbons traps are located within either the footwall or hanging wall blocks of major faults (e.g. the Piper and Brae fields). Other, younger Lower Cretaceous hydrocarbons traps include a larger component of stratigraphic trapping (e.g. the Britannia Field).

12 | P a g e

all the Mesozoic traps are deeply buried by the thick Cenozoic successions which occur within or overlie the rift-basins. In contrast, many of the Paleocene traps are  buried at relatively shallow depths. For example, the Balmorals Oilfield lies at about 2150 m depth (Tonkin and Fraser 1991).

Figure 3.2: Geologic Provinces Total Petroleum System (403601).

13 | P a g e

Figure 4.2: Petroleum System Glennie (1998)

14 | P a g e

2.4.

Structural setting

In the northern Dutch offshore Yore dale-type cycles were deposited from the Middle Visean to the Early Namurian in relation to northern source. Continued normal faulting resulted in a gradual basin deepening during the course of the Early Carboniferous. Towards the end of the Visean differential subsidence had resulted in a complex series of deep-water basins characterised by faulted basin-floor topography. This caused local stagnation of deep -water circulation at the transition from the Visean to the Namurian and resulted in the deposition of a thick bituminous shale: the Geverik Member (UK: Bowland Shale). Within the Dutch sector, this organic-rich shale was only encountered in proximity to the Brabant Massif; extrapolating its extent into the main basin seems premature. Correlations of the overlying Namurian deltaic show that the organic-rich Geverik Member just northeast of the London Brabant Massif accumulated in a basin that was 200 meters deeper than a similar location north of the London Brabant Massif where Visean carbonates became exposed. The fact that the Geverik Member overlies a Dinantian carbonate sequence suggests that the depositional location was part of the northern flank of the London-Brabant Massif during the Early Carboniferous, instead of the adjacent graben. This observation challenges the interpretation of Middle Carboniferous organic rich shales as deep-water deposits and sheds doubt on its  presumed basin wide extent. Fault orientation and Late-Carboniferous sediment-distribution patterns on the pre-Permian sub crop map show that Early Carboniferous extension may have dominated the entire eastern Netherlands. Similar patterns are observed in the northern offshore. Namurian - Thrust loading during the Namurian resulted in high subsidence and in the formation of a foreland basin north of the London-Brabant Massif. Correlations of the overlying Namurian deltaic cycles suggest that the

15 | P a g e

faulted topography of the Variscan back-arc basin had been leveled by the Early  Namurian. The thickness of the Namurian succession increases away from the massif but subsequently decreases over the Mid- Netherlands Zandvoort-Krefeld High, a NW-SE aligned structural element of low subsidence until well into the Tertiary. The high is represented on the sub crop map by the major NW-SE lineament separating the Lower Westphalian in the north from the Upper Westphalian in the south. Another deep basin extended northward from the Zandvoort- Krefeld High to the Texel-Ijsselmeer High, a structural element that may  be traced across the North Sea towards the northern basin margin. In the northern Dutch offshore the Namurian is characterised by thick fluvial sandstones. Basal turbidity fans, characteristic of the lowermost Namurian in some of the northern sub basins in the UK were not deposited in the southern part of the Southern North Sea Basin. Westphalia  –   The lithological transition from the Namurian to the Westphalian is placed at the first level of regional occurrence of coal beds. Across the onshore Netherlands this is a characteristic series of three coals, interbedded thin shales and sands, which shows that the entire basin The Mid-Netherlands High has since long been regarded as an area of intraWestphalian uplift. Its southwestern boundary is defined by the Zandvoort-Krefeld High. Uplift of the Mid-Netherlands High is attributed to transpressional movements along its southern boundary faults. Evidence of intra-basinal sediment supply is  presented by the stacked fluvial sandstones. These sands were found to contain reworked coal fragments of which vitrinite-reflectance data show that they were  previously buried at intermediate depths. F3 is a block in the Dutch sector of the North Sea. The block is covered by 3D seismic that was acquired to explore for oil and gas in the Upper-Jurassic –  Lower Cretaceous strata, which are found below the interval selected for this demo set. The

16 | P a g e

upper 1200ms of the demo set consists of reflectors belonging to the Miocene, Pliocene, and Pleistocene. The large-scale sigmoidal bedding is readily apparent, and consists of the deposits of a large fluviodeltaic system that drained large parts of the Baltic Sea region (Sørensenetal, 1997; Overeemetal, 2001). The deltaic package consists of sand and shale, with an overall high porosity (20 – 33%). Some carbonate-cemented streaks are present. A number of interesting features can be observed in this package. The most striking feature is the large -scale sigmoidal bedding, with text-book quality downlap, toplap, onlap, and truncation structures. Bright spots are also clearly visible, and are caused by biogenic gas  pockets. They are not uncommon in this part of the North Sea. Several seismic facies can be distinguished: transparent, chaotic, linear, shingles. Well logs show the transparent facies to consist of a rather uniform lithology, which can be either sand or shale. The chaotic facies likely represents slumped deposits. The shingles at the  base of the clinoforms have been shown to consist of sandy turbidites.

17 | P a g e

Figure 5.2: structure features of North Sea (Ziegler, 1990)

18 | P a g e

Figure 6.2: Geological cross section ziegler (1978)

19 | P a g e

2.5.

Tectonic history

The North Sea area was the site of a triple plate collision zone during the Caledonian orogeny Four major tectonic events influenced the area since the Cambrian: (i) the Caledonian collision during Late Ordovician to Early Silurian, (ii) subsequent rifting and basin formation mainly identified in the Carboniferous to Permian, (iii) Mesozoic rifting and graben formation and (iv) inversion during Late Cretaceous to Early Tertiary (Ziegler, 1990). The Caledonian collision involved two large continents, The tectonic regime changed from general extension and subsidence in The Devonian to a strike-slip regime during the late Carboniferous to early Permian. During the late Variscan cycle northwestern Europe was transected by a system of conjugate shear faults, which in the Danish area resulted in the development of the Tornquist Fan of faults that form links between NNE to SSW trending graben structures (Thybo, 1997). At the late Carboniferous, the Variscan orogeny collapsed and uplift caused truncation of the Devonian – Carboniferous successions. The late Carboniferous and early Permian extensional wrench tectonics caused crustal thinning and subsidence of the Northern and Southern Permian Basins, which are separated by the Mid-North Sea-Ringkøbing Fyn High. The crustal thinning was associated with magmatic activity, evidenced in Rotliegendes volcanic deposits and associated sedimentary rocks. The Rotliegendes or top preZechstein reflector marks a regional unconformity between pre-rift and syn-rift deposits and defines the traditional acoustic basement in most reflection seismic data from the North Sea. From the Triassic to the Jurassic The NNW_SSE striking Central, Viking and Horn Grabens developed with extensive normal faulting in the basin areas. In the graben areas the Jurassic and

20 | P a g e

Triassic sediments attain thicknesses in excess of 4 km (Vejbaek, 1992). During the Late Cretaceous Laramide phase of the Alpine orogeny compressional stresses caused inversion and deformation throughout the Danish and the North Sea area. Since the Early Tertiary the geological evolution has been characterized by regional subsidence in the North Sea area and uplift of the Baltic Shield.

21 | P a g e

CHAPTER 3: 3. THEORITICAL BACKGROUND OF SEISMIC AMPLITUDE INVERSION 3.1.

Introduction

3.1.1. Seismic velocity inversion The first type of inversion, velocity inversion, sometimes known as travel time inversion, is used for depth imaging. Using seismic traces at widely spaced locations, it generates a velocity-depth earth model that fits recorded arrival times of seismic waves. The result is a relatively coarse velocity-depth model extending over several kilometers in depth and perhaps hundreds of kilometers in length and width. This solution is applied in data-processing steps such as migration and stacking, eventually producing the type of seismic image that is familiar to most readers. Seismic interpreters use these images to determine the shape and depth of subsurface reflectors.

Figure 7.3: velocity inversion R. Kamei, R.G. Pratt and T. Tsuji

22 | P a g e

3.1.2. Seismic Amplitude Inversion The second type of inversion, amplitude inversion, is the focus of this article. This approach uses the arrival time and the amplitude of reflected seismic waves at every reflection point to solve for the relative impedances of formations  bounded by the imaged reflectors. This inversion, called seismic inversion for reservoir characterization, reads between the lines, or between reflecting interfaces, to produce detailed models of rock properties. In principle, the first step in model-based seismic inversion — forward modeling —   begins with a model of layers with estimated formation depths, thicknesses, densities and velocities derived from well logs. The simplest model, which involves only compressional (P-wave) velocities invert for P-wave, or acoustic, impedance. The simple model is combined with a seismic pulse to create a modeled seismic trace called a synthetic. Inversion takes an actual seismic trace, removes the seismic pulse, and delivers an earth model for that trace location. To arrive at the  best-fit model, most inversion routines iterate between forward modeling and inversion, seeking to minimize the difference between the synthetic trace and the data. In practice, each of these steps may be quite involved and can depend on the type of seismic data being inverted. For vertical-incidence data, creating the initial model requires bulk density measurements from density logs and compressional velocities from sonic logs, both spanning the interval to be inverted. Unfortunately, the necessary logs often are acquired only in the reservoir. In the absence of sonic logs

23 | P a g e

Borehole seismic surveys — vertical seismic profiles (VSPs) — can provide average velocities across the required interval. If no borehole velocity data governs reflection-driven changes in normal interface through the acoustic impedance contrast; reflectivity is the ratio of the difference in acoustic impedances to their sum.3 the result in depth-based reflectivity model is converted to a time-based model through the velocities combining the time-based model with a seismic pulse creates a synthetic trace. Mathematically, this process is known as convolution. The seismic  pulse, or wavelet, represents the packet of energy arriving from a seismic source. A model wavelet is selected to match the amplitude, phase and frequency characteristics of the processed seismic data. Convolution of the wavelet with the reflectivity model yields a synthetic seismic trace that represents the response of the earth as modeled to the input seismic pulse. Additional steps are needed when Models that include noise, attenuation and multiple reflections are to be included in the modeled trace. The inverse operation starts with an actual seismic trace because the amplitude and the shape of each swing in the seismic trace affect the outcome it is vital that the  processing steps up to this point converse signal phase and amplitude. Different types of inversion start with different types of traces the main distinction is between  performed before stacking and inversion after it. Prestack and post stack. Most seismic surveys provides images using data that have been stacked. Stacking is a signal enhancement technique that averages many seismic traces the traces represent recording from a collection of different sources and receivers offset with common reflecting midpoint each trace assumed to minimal random noise and with signal amplitude equal to the average of the signal in the Stacked traces the resulting stacked trace is taken to be the response of normal- incidence reflection at the common midpoint (CMP).

24 | P a g e

Stacking is a reasonable processing step if certain assumptions hold: the velocity of the medium overlying the reflector may vary only gradually, and the average of the amplitudes in the stacked traces must be equivalent to the amplitude that would be recorded in a normal- incidence trace. In many cases, these assumptions are valid, and inversion may be performed on the stacked data in other words, post stack. In contrast, when amplitude varies strongly with offset, these assumptions do not hold, and inversion is applied to unstacked traces pre stack. Before discussing pre stack situations in detail, we continue with the simpler case of  posts tack inversion. A stacked trace is compared with the synthetic trace computed from the model and wavelet. The differences between the two traces are used to modify the reflectivity model so that the next iteration of the synthetic trace more closely resembles the stacked trace. This process continues, repeating the generation of a synthetic trace, comparison with the stacked trace, and modification of the model until the fit between the synthetic and stacked traces is optimized. There are many ways to construct synthetic traces, and various methods may be used to determine the best fit. A common approach for determining fit is least-squares inversion, which minimizes the sum of the squares of the differences at every time sample. This inversion technique operates on a trace-by-trace basis. In the simplest case, inversion produces a model of relative reflectivity at every time sample, which can be inverted to yield relative acoustic impedance. To obtain formation properties such as velocity and density, a conversion to absolute acoustic impedance is necessary. However, such a conversion requires frequencies down to near 0 Hz, lower than contained in conventional seismic data. An absolute acoustic impedance model can be constructed by combining the relative acoustic impedance

25 | P a g e

model obtained from the seismic frequency range with a low-frequency model derived from borehole data.

Figure 8.3: Modeling and Inversion Rasmussen KB Brunn A and Pedersen JM (2004)

Relating seismically derived acoustic impedance to formation properties makes use of correlations between logging measurements. For example, cross  plotting acoustic impedance and porosity measured in nearby wells establishes a transform that allows seismically measured acoustic impedance to be converted to  porosity values throughout the seismic volume. An example from a carbonate reservoir in Mexico demonstrates the power of this technique (Figure 9.3).

26 | P a g e

Figure 9.3: Porosity vs Impedance, Salter R Et al (2005)

Seismic inversion or stratigraphic deconvolution tries to put a spiked response at geological boundaries (lithology changes) and the main reservoir characteristic interfaces. This is done by the inversion of the seismic cube into an Acoustic Impedance cube (Figure 10.3). The link between the seismic cube and the AI cube is the seismic wavelet. Seismic inversion is a rather confusing expression. Inversion in itself means to undo an operation, but here in fact it is used for the transformation of a seismic amplitude cube into an acoustic (or elastic) impedance cube. One of the  benefits of inversion is that the seismic resolution is increased (e.g. Veeken et al. 2004). Hill (2005) has investigated this phenomenon and found an improvement in thickness prediction that was clearly beyond the seismic tuning thickness. In the stratigraphic inversion scheme a comparison is made between the synthetic trace at the well and the seismic trace. A wavelet is established by applying 27 | P a g e

cross correlation techniques. Or the wavelet is derived from the shaping filter that  permits transformation of the reflection coefficient trace into the seismic trace. This wavelet is then used to perform the seismic inversion, whereby the seismic traces are transformed into blocky AI traces. The spiked response is expressed by the limits of these AI units (Veeken et al. 2002a). These spikes are assumed to correspond  better to meaningful geological boundaries and reservoir characteristic interfaces. If all works well there is relation between acoustic impedance and reservoir characteristics like porosity, permeability, net-to-gross, HC saturation. Even if the well trends are not exactly honored, the filtered well average may follow the general  picture provided by the inversion and that gives the possibility to delineate ‘sweet spots’. A simple relationship often does not exist (Dvorkin and Alkhater 2004), but in individual cases it can be different. For instance in the Kraka field, located offshore Denmark, the Chalk porosity is linear correlated with AI (Klinkby et al. 2005) and this can be used in later prediction.

Figure 10.3: Porosity Prediction, Salter R etal (2005)

28 | P a g e

3.1.2.1. Principles The reflection coefficient series is convolved with the seismic wavelet to give the seismic trace. Inversion aims to start from the seismic trace, remove the effect of the wavelet to get back to the reflection coefficient series, and from them derive the layer impedances. It has to assume that starting seismic data are free from correlated noise (e.g. Multiples).Also, the wavelet presenting the data has to be estimated many inversion methods derive the wavelet from a well tie and have to assume that it does not change laterally away from the well. There is also an amplitude calibration to be taken into account real seismic traces are not directly output as reflection coefficient values, but are scaled to give some convenient but arbitrary RMS average over a trace. At least over a limited time-gated, the Ratio of reflection coefficient of trace amplitude has to be constant if inversion is going to work. Care is needed during seismic processing to avoid steps that might introduce artificial amplitude changes vertically or horizontally. However, locally variable effects in the overburden (e .g. shallow gas) can reduce the Seismic energy penetrating to deeper reflectors and so reduce the reflected signal left to itself, the inversion process would try to interpret this as a decrease in impedance contrast across the deeper interfaces. To remove such artefacts, a long-gate AGC may be applied, which scales amplitudes so as to remove lateral variation when averaged over a TWT interval of 1s, for example. Another issue is that the seismic traces contain data of limited bandwidth; the frequencies Present depend on the rock properties and the seismic acquisition technique, but might be in the range from 5 to 50 Hz. This means that the low frequencies in particular, which are critical for the estimation of absolute impedance values, cannot be obtained from the seismic data, but have to be added from elsewhere. Usually from a model based on well data and geological knowledge.

29 | P a g e

Figure 11.3: Seismic Inversion

3.1.2.2. Extending the bandwidth To get more information into a seismic section than is actually present in the seismic traces, extra data have to be obtained from elsewhere; different algorithms differ in detail, but they all have the following general features (1) Begin with a model that describes the subsurface explicitly or implicitly, this will contain a number of layers of different acoustic impedance. (2) Calculate the seismic response from this model using the wavelet representing the seismic dataset. (3) Compare the calculated seismic response with the real data.

30 | P a g e

(4) Modify the models to reduce the misfit between the calculated and real seismic, perhaps iteratively and perhaps incorporating constraints on the impedance values that may be assigned to particular layers on the complexity of the model and on the variation of layer parameter from one seismic trace to the next along a seismic line. (5) Perhaps add low-frequency information obtained from a model based on geological data. The additional information that has been taken into account is thus: (a) The wavelet removal of its effects is equivalent to a de convolution extending bandwidth at the high-frequency end (b) The model the geological input extends bandwidth at the low-frequency end. To make all this more concrete it is helpful to work through an actual  processing flow. A common approach to building a subsurface model is to split it up into macro layers, probably several hundred so Fms thick, bounded by the main semi-regional seismic markers and consisting of a single broad lithology .This makes it easier to construct a geological model to constrain impedance variation within a macro layer Inside the macro layer the subsurface is represented by means of a series of reflectivity spikes. These spikes when convolved with the wavelet should reproduce the observed seismic Trace and integration of the reflectivity will give the impedance variation within the Macro layer To prevent the software from trying to reproduce all the noise presenting the seismic section it is usual to impose a requirement that the reflectivity spike series is as simple as possible either in the sense of using a small total number of spikes or using spikes of small total absolute amplitude; the algorithm will trade off the misfit between real and calculated seismic section against the complexity of the spike model, us ually under user control of the

31 | P a g e

acceptable degree of misfit. The result is usually called a 'sparse spike' representation of the subsurface. It is obviously critical to the success of this process that we know the wavelet accurately. This is usually obtained from a well-tie study. Well synthetic or VSP information will tell us how zero-phase seismic ought to look across the well; comparison with the real data tells us that wavelet is present shows an example display from such a study. To the left, in red is the candidate wavelet, which in this case is close to being symmetrical zero-phase to; the right is a panel of six (identical)traces showing the result of convolving this wavelet with the well reflectivity Sequence derived from log data .They should be compared with the panel of traces in the middle of the figure, which are the real seismic traces around the well location. Various geologically significant markers are also shown. There is clearly a very good match between the real and synthetic data so far as the principal reflections are concerned, so it is possible to have confidence that the wavelet shown is indeed that presenting the data. The real so some differences in detail which will limit the accuracy of an inversion result. These may arise from imperfections in the seismic processing perhaps the presence of residual multiples or minor imaging  problems With luck, the inversion process will leave some of this low-energy noise out of the inverted image because of the sparseness of the spike reflectivity series.

Figure 12.3: Synthetic trace and wavelet extraction Buiting,J . J. M. & Bacon,M . (1999)

32 | P a g e

3.2.

Post-stack Inversion Methods

The post-stack time-migrated data is input for the inversion algorithm. It is important to have clean seismic data as input and that proper data conditioning is done (Da Silva et al. 2004). This entails CDP Gather cleaning, preserved amplitude  processing, amplitude balancing with sophisticated gain control, multiple suppression and 3D noise attenuation. Comparative analysis of seismic processing (CASP, Ajlani 2003) is a multi-disciplinary approach to secure the QC of the seismic  processing results, whereby quality, turnaround and cost are taken into account. Usually the seismic is processed with a certain target zone in mind. All the  processing parameters have been tuned to this objective. The target in the inversion may be somewhat different. Therefore careful examination of all processing steps is needed before embarking on the exercise to transform amplitude into acoustic impedance. 3.2.1. Colored inversion The colored inversion method is basically a trace integration, achieved by applying a special filtering technique in the frequency domain. The amplitude spectrum of the well log is compared with that of the seismic data and that is where the word ‘colored’ comes from. An inversion operator is designed that brings the seismic amplitudes of the frequencies in correspondence with that seen in the well. This operator is then applied to the whole seismic cube (Lancaster and Whitcombe 2000). A cross plot is made between the amplitude and the logarithm of the frequency to compute the operator. A linear fit is performed to calculate an exponential function and this serves as a shaping filter (cf Walden and Hosken 1985, Velzeboer 1981). This filter transforms the seismic trace into an assumed acoustic impedance

33 | P a g e

equivalent. The assumption is made that the seismic cube is zero phase, which is hardly the case. In individual cases (Giroldi et al. 2005) the results can be spectacular with a flat spot DHI suddenly appearing in the inverted dataset (Figure 13.3). Again this type of Colored inversion for land data from the Chaco Basin in Bolivia. The reflectivity section on the left does not show the same break at the gas – water contact as seen by the relative AI volume. It illustrates the benefit of extracting the relative AI attribute in these Cretaceous reservoirs.

Figure 13.3: Colored inversion vs Seismic data (Lancaster and Wh itcombe 2000)

3.2.2. Sparse spike inversion The seismic trace is simulated by a minimum amount of AI spikes. The spikes are placed in such a way that they explain best the seismic response. Amplitude, time  position and number of the AI spikes is not always realistic, i.e. not conform the geological constraints. To be more specific: if a starting model is not available, the spikes might be placed in unrealistic positions and still the model generates a synthetic that highly resembles the seismic trace. The recursive method uses a feedback mechanism to obtain more satisfactory results. A low frequency AI

34 | P a g e

variation trend can be imported to generate more appropriate results with a better convergence of the found solution. The inversion algorithm was initially working on a trace by trace basis, but now a multi-trace approach is implemented. The inversion solution may vary considerably from trace to trace, making the reliability of the output weaker. The constrained option uses a low frequency model as a guide. The low frequency variation is estimated from blocked well logs and this gives much better results (e.g. Ronghe and Surarat, 2002). QC tests are normally necessary, based on the match between the up scaled well log impedance and the inverted traces, because little well info is required in the construction of the a -priori impedance model (Klefstad et al. 2005). It provides a consistency check of the inversion results. The inversion replaces the seismic trace by a pseudo acoustic impedance trace at each CDP position (Pendrel and Van Riel 2000). The sparse spike hypothesis implies, however, that a thin bed geometry will not always be mimicked in the most optimal way. The zero phase requirement can be circumvented by choosing a compound wavelet for the inversion, thus compensating the non-zero phase characteristics of the input data. A multi-trace approach results in better stability of the generated inversion solution. Sophisticated model-driven sparse spike inversion gives more realistic output. In some cases the interpreter gets away with the approximation, but in the majority of the cases yet a better job is needed.

35 | P a g e

Figure 14.3: Sparse spike inversion (modified after Ronghe and Surarat 2002).

3.2.3. Model-based inversion A simple initial AI model is perturbed, a synthetic computed using the seismic wavelet and the difference with the seismic trace is established (Cook and Sneider 1983, Fabre et al. 1989, Gluck et al. 1997). The AI model with a very small difference is retained as solution (Figure 6.22). A simulated annealing technique using a Monte Carlo procedure is applied (Goffe et al. 1994, Duboz et al. 1998). This technique shows an analogue to the growth of crystals in a cooling volcanic melt (Ma 2003). It starts with a reflectivity model M0 and computes the difference with the seismic input data after convolution with a wavelet. The model is now perturbed and a new model Mn is simulated, where for the same difference is established. The two differences are compared and if the misfit for f(Mn) is smaller than that for M0,

36 | P a g e

than the Mn model is unconditionally accepted. If not, than the Mn model is accepted  but with a probability: P = e(−f(Mn)−f(M0))/T , (6.22) Whereby T is a control parameter (acceptance temperature). This acceptance rule is known as the Metropolis criterion (Metropolis et al. 1953). The process is repeated a large number of times until a very small residual difference is found (threshold value) that is stable. Computation of Cost functions permit to determine a real regional minimum for the difference. The initial AI model is made up of macro-layers defined by the shape of the seismic mapped horizons. Micro-layers are automatically introduced in this macro model. It provides a stratigraphic grid cell volume together with the inline and crossline subdivision for storing constant AI values. The use of micro-layers makes sure that an adequate number of spikes (i.e. vertical change in AI) is utilized in the modelling (Figure 6.23). This model-driven inversion method is much more robust and often a real 3D inversion algorithm is applied to stabilize the solution (Duboz et al. 1998, Coulon. et al. 2000, Veeken et al. 2002b; Figure 6.24). A bulk phase rotation can be applied to zero phase a seismic sub cube. Normally this procedure is valid for a small time window (<1.5 sec TWT), where a stable wavelet is derived (Figure 6.25). The results are evaluated at the well control points in so-called composite well plots (Figure 6.26). The seismic and the AI cubes are compared (Figure 6.27). Layer maps are quite useful to delineate the extent of AI anomalies (Figure 6.28). The method can give satisfactory results even when well control is limited and the seismic quality is rather poor (Veeken et al. 2002a; Figures 6.29 and 6.30). It is also possible to derive a wavelet straight from the seismic dataset. The model-driven inversion does not always honour the well control completely, but a great advantage is that the seismic data is the guide for the inversion. Errors in the well logs do not propagate in the inversion.

37 | P a g e

This is an advantage when the old well database is unreliable. The averaging effect, introduced by the 3D approach, results in small discrepancies at the well locations that are in fact quite acceptable. Also in carbonates good inversion results are possible. The relative velocity and density changes induced by the pore fill are decisive for creating AI anomalies.

Figure 15.3: Model based inversion (Goffe et al.1994, Duboz et al. 1998)

Figure 16.3: Correlation between check shots and impedance log veeken 2006

38 | P a g e

Figure 17.3: Seismic vs Impedance, Veeken 2006

39 | P a g e

3.2.4. Stochastic inversion Geostatistics are used to build a complete subsurface model and constrain the inversion solutions (Dubrule 2003). Simulation is done on local level as well as globally, on the totality of the generated model (Haas and Dubrule 1994, Dubrule 2003). All models honour the well data, otherwise they are rejected. The architecture of reservoirs is classified in various ways (Weber and Van Geuns 1990) and this helps in selecting the simulation approach. Probability density functions (PDF’s) are established for each grid point and these are used to perform a random simulation (Van der Laan and Pendrel 2001). The input for the PDF comes from well logs, spatial properties (variograms) and lithological distributions. The stochastic algorithm calculates for each simulation a synthetic trace, compares it with the real seismic trace and accepts or rejects it. A simulated annealing process is utilized. The number of solutions is reduced in this way and  probability maps are produced to assess the risk. The retained simulations are examined on their variance. If they closely resemble, then the prediction is rather good and the confidence level of the output is increased A probability volume is generated for grid points with porosities above 10%, using the simulation histograms. Subsequently bodies are outlined where the  probability is above 70% for the porosity to be higher than 10%. As more wells are drilled in the same petroleum system, the best matched simulations are retained to further refine the predictions (Sylta and Krokstad 2003). Drawback of the  probabilistic method is that the interpreter has to quantify the uncertainties in a realistic way. This is a tedious and precarious task (cf Klefstad et al. 2005). Areas without proper well control are still difficult to predict and assumptions have to be made. There is a cumulative increase in prediction error as various reservoir  parameters have to be estimated at the same time. 40 | P a g e

Figure 18.3: Stochastic inversion (modified after To rresVerdin et al. 1999).

41 | P a g e

CHAPTER 4: 4. SEISMIC

INVERSION

USING

HAMPSON-RUSSELL

SOFTWARE 4.1.

Introduction

Inversion is the process of extracting, from seismic data, the underlying geology which gave rise to that seismic. Traditionally, inversion has been applied to  post-stack seismic data, with the aim of extracting acoustic impedance volumes (Strata).Recently, inversion has been extended to pre-stack seismic data, with the aim of extracting both acoustic and shear impedance volumes. This allows the calculation of pore fluids (Strata + AVO). Hampson-Russell has been providing innovative geophysical software since 1987. The Hampson-Russell software suite encompasses all aspects of seismic exploration and reservoir characterization, from AVO analysis and inversion to 4D and multicomponent interpretation. There are several types of seismic inversion can be done using STRATA-HR: Post-stack: -Recursive: Traditional band-limited inversion. -Model Based: Iteratively updates a layered initial model. -Sparse Spike: Constrained to produce few events. -Colored: Modern derivative of Recursive Inversion. Pre-stack: -Elastic Impedance: Enhancement for pre-stack data.

42 | P a g e

-Independent Inversion: Enhancement for pre-stack data. -Lambda-mu-rho (LMR): Enhancement for pre-stack data. -Simultaneous Inversion: Enhancement for pre-stack data. 4.2.

Available Data

The project is a prospect in North Sea specified in Dutch sector place offshore  Netherlands F3 Block with coordinates N 54°52′0.86″ / E 4°48′47.07 this project was held in 1987 and that by making 3d seismic survey and drilling of 4 wells in different area to cover our cube which has 24*16 km square our target is at late Jurassic early cretaceous for very rich gas prospect. The available data is: Seismic Data: •

A 3D seismic cube that consists of 740 in lines and 949 xlines and it is

 post stack cube Horizons •

Demo0 (FS4)

•

Demo1 (MFS4)

•

Demo2 (FS6)

•

Demo3 (Top_Foresets)

•

Demo4 (Truncation)

•

Demo5 (FS7)

•

Demo6 (FS8)

43 | P a g e

•

Demo7 (Shallow)

Wells and available logs

4.3.

•

F02 – 1 (Caliper, Density, GR, P-Wave, Porosity)

•

F03 – 2 (Density, GR, P-Wave, Porosity)

•

F03 – 4 (Density, GR, P-Wave, Porosity)

•

F06 – 1 (Density, GR, P-Wave, Porosity)

Inversion Concepts

All inversion algorithms suffer from “non -uniqueness”. There is more than one possible geological model consistent with the seismic data. The only way to decide between the possibilities is to use other information, not present in the seismic data. Figure (4.1) represents the combination between the geology and seismic data in inversion. This other information is usually provided in two ways: •

The initial guess model

•

Constraints on how far the final result may deviate from the initial

guess. The final result always depends on the “other information” as well as the seismic

Figure 19.4: combination between the geology and seismic data in inversion

44 | P a g e

Figure 20.4: Seismic Inversion Components

Model Based Inversion starts with the equation for the convolutional model:

Assume that the seismic trace, S, and the wavelet, W, are known. Assume that the Noise is random and uncorrelated with the signal. Solve for the reflectivity, R, which satisfies this equation. This is actually a non-linear problem, so the solution is done iteratively: Step 1: The initial background model for Model Based Inversion is formed by  blocking an impedance log from a well (Figure 21.4). Step 2: Using the blocked model, and the known wavelet, a synthetic trace is calculated. This is compared with the actual seismic trace. By analyzing the errors or “misfit” between synthetic and real trace, each of the layers is modified in 45 | P a g e

thickness and amplitude to reduce the error. This is repeated through a series of iterations (Figure 22.4).

Figure 22.4: Step 2

Figure 21.4: Step 1

Model Based Inversion produces a broad-band, high frequency result a  potential problem is that the high frequency detail may be coming from the initial guess model, and not from the seismic data. This problem is minimized by using a smooth initial model. Issues in Model Based Inversion: (1)Because the wavelet is known, its effects are removed from the seismic during the calculation. For example, the seismic does not have to be zero-phase, as long as the wavelet has the same phase as the seismic. (2)Errors in the estimated wavelet will cause errors in the inversion result. (3)The effective resolution of the seismic is enhanced. (4)The result can be dependent on the initial guess model. This can be alleviated by filtering the model.

46 | P a g e

(5)There is a non-uniqueness problem, as with all inversion.

Figure 23.4: Comparison between Input Seismic and Model Based Inversion

4.4.

Model-Based Inversion Workflow on F3-Block Seismic Cube

4.4.1. Data Importing Start with creating new well database "Inv.wdb"

Figure 25.4: HR Opened Database List

Figure Database List a g Opened e 47 | P24.4:

HR Software interface appear as shown (Figure 25.4)

Figure 26.4: Hampson-Russell Interface

Well Explorer is used to import well information (Logs,Tops or Geometry)

Figure 27.4: Well Explorer Window

Import LAS files and use default settings till the shown window in (Figure 27.4)

Figure 28.4: Importing Logs

48 | P a g e

Sometimes unknown logs problem happen in this case identifying logs again is required (Figure 28.4)

Figure 29.4: Identifying unknown logs

Check Shots and well tops can be imported with the same steps. STRAT is the responsible part from HR for running inversion. Thus, after creating well database start new project from STRATA as shown (Figure 29.4)

Figure 30.4: STRATA Project Selection Mode

49 | P a g e

Importing F3-Block Seismic Cube (Figure 30.4)

Figure 31.4: Importing Data

Identifying seismic parameters and byte location is a critical step in importing seismic data, In F3-Block the following information was used (Figure 31.4): Inline byte location= 189 Xline byte location= 193 Data Sample Format: IBM  NOTE: In Post-Stack Inversion we can ignore Receiver X & Y Coordinates but we cannot in Pre-stack Inversion.

50 | P a g e

Figure 32.4: SEG-Y Seismic File Open

In Geometry Grid page, we can see number of inline and xlines, no. of traces and other parameters shown in (Figure 32.4)

Figure 33.4: Geometry Grid Page

51 | P a g e

Then each well location is determined in the cube by inserting their xline and inline in Well to Seismic Map Menu where the CDPs are calculated automatically (Figure 33.4)

Figure 34.4: Well to Seismic Map Menu

4.4.2. Identifying Target Reservoir in F3-Block Cube The target reservoir is known in this area but the available geologic information told enough to identify the reservoir. Geology told that to look for deltaic package consists of sand and shale, with an overall high porosity (20 – 33%) for Gas sand. Cross plot tool in HR eLog is used to make a cross plot between GR values and Porosity. (Figure 34.4). The plot shows two zones (Grey and Blue) with low GR values (Sand) and porosity values 20-40 % what matches the reservoir conditions told by Geology. By creating cross-section on log for this zones the possible reservoir zones could be seen. Blue zone at depth range 720-800m and Grey zone at depth range 1050-1100m at well F02 (Figure 35.4).

52 | P a g e

Figure 35.4: GR-Porosity Cross plot

Figure 36.4: Well F020-1 Cross Section

53 | P a g e

To make sure about reservoir another cross plot between GR Values and PImpedance Values (Figure 36.4). found that at the same GR Values of Blue and Grey zones in the GR-Porosity cross plot P-Impedance are mostly high between 42005600 ((m/s)*(g/cc)) which is quite near to Gas Sand impedance values that confirmed by Lithology Prediction using Seismic Inversion Attributes, Dan Hampson, 2010. Using cross-section for well F02 for Blue Zone we could identify reservoir zone clearly at depth 720-800m besides small adjacent zones (Figure 37.4) From well tops data target reservoir is FS8 (Figure 38.4).

Figure 37.4: GR-Impedance Cross Plot

54 | P a g e

Figure 38.4: F02-2 Cross section for GR-Impedance cross plot

Figure 39.4: FS8 the target reservoir

55 | P a g e

Then FS8 Horizon is picked as target reservoir on seismic in addition to FS11 as a good marker in region. Import Horizons file in STRATA as shown in (Figure 39.4)

Figure 40.4: Picking Horizon FS8

Figure 41.4: Actual Picks of FS8

56 | P a g e

4.4.3. Depth-Time Conversion and Check Shot Correction The initial guess model for each trace consists of an impedance log, usually derived  by multiplying a real sonic log by a real density log. The impedance log model must  be measured in 2-way travel time. The original logs are measured in depth. A critical step is depth-to-time conversion. (Figure 41.4)

Figure 42.4: Difference between Time logs and Depth logs

The depth-to-time conversion is made using a depth-time table which maps each depth to the two-way travel time from the datum (surface) to that depth and back (Figure 42.4). The depth-time table is usually calculated from the sonic log velocities using this equation:

57 | P a g e

Figure 43.4: Depth time table

The time to an event depends on all the velocities above that layer, including the first velocity to the surface, V1. That velocity is unknown and is usually approximated  by extrapolating the first measured velocity back to the surface. The depth-time table calculated from the sonic log is rarely sufficient to produce model impedance which ties the seismic data properly because: -The seismic datum and log datum may be different. -The average first layer velocity is not known. -Errors in the sonic log velocities produce cumulative errors in the calculated traveltimes. -The events on the seismic data may be mis-positioned due to migration errors. -The seismic data may be subject to time stretch caused by frequency-dependent absorption and short-period multiples.

58 | P a g e

Check Shot Correction: Check shot table is a series of measurements of actual 2-way time for a set of depths. The depth-time table calculated from the sonic log must be modified to reflect the desired check shot times. (Figure 43.4)

Figure 44.4: Check Shot Correction

4.4.4. Wavelet Extraction and Synesthetic Trace After check shot correction, Correlation between seismic data and synthetic trace is made to extract the appropriate wavelet and get the hi ghest correlation value.

59 | P a g e

We used statistical method to extract wavelet with different phases (0, 90,180 and 270) and compare its correlation percentage to get the best wavelet parameters in each well. (Figure 44.4) In well F02-1 Wavelet-270 gives the best correlation about 30% at time range 5001100 ms and the percentage increases at the reservoir zone 750-900 ms to 75.5 %( Figure 45.4). While in well F03-4 main phases gives low correlation so we used scan tool to get the best phase with highest possible correlation which is 38 deg. with correlation about 45.5%. (Figure 46.4).

Figure 45.4: Wavelet with different phases (0, 90,180 and 270)

60 | P a g e

Figure 46.4: Wavelet 270 correlation in well F02-1

Figure 47.4: Wavelet zero correlation in well F03-4

61 | P a g e

4.4.5. Initial Model Building  Now, we have two synthetic traces in both wells (F02-1 and F03-4) with good correlation with seismic data besides two picked horizons (FS8 and FS11). All of them enables me to start building initial model for inversion process. In STRATA, Choose Model

 Build/Rebuild

a Model as shown (Figure 47.4)

Figure 48.4: Initial Model Building

Insert correlated wells only (F02-1 and F03-4) (Figure 48.4)

Figure 49.4

In each well we used Corrected P-wave log resulted from correlation and computed Impedance resulted from multiplying p-wave log by Density log (Figure 49.4)

Figure 50.4

62 | P a g e

Use default settings for Modeled Trace Filtering Options (Figure 50.4)

Figure 51.4

Resulted Initial Model (Figure 51.4)

Figure 52.4: Initial Model for F3-Block Seismic Cube

63 | P a g e

4.4.6. Quality Control (QC) Before running inversion we carried out some QC analysis which is very important to make sure that inversion results would be effective and the whole work is not in vain. In STRATA, Analysis tool provide good QC functions (Figure 52.4): Red curve represents Inverted result and Blue one represents original log. This window shows us the matching between them with error = 238 which is good error value and the correlation between Synthetic traces (in Red) and Seismic (in Black) with value 0.93 which is good but we can get higher by modifying some parameters such as number of iteration.

Figure 53.4: Inversion Analysis for Post-stack Data

64 | P a g e

In the same window, Invert enables us to run virtual Inversion before real process to get the appropriate parameters and avoid run time consumption. Number Number of Iteration is one of this critical parameters as if it exceeds certain limit would lead to high error. It found that iteration value between 10-20 10 -20 gives best results in most cases we used 20. (Figure 53.4)

Figure 54.4: Number of Iteration

65 | P a g e

Error plot enables us to get the error value at each well (Figure 54.4)

Figure 55.4: Error Plot 

Inversion Analysis Cross plot create plot between computed impedance and original one in log. The heaviest the cloud of dots the better inversion results (Figure 55.4).

Figure 56.4: Inversion Analysis Cross plot 

66 | P a g e

4.4.7. Running Inversion Once running inversion process program calculate P-Impedance at each well and interpolate between them to extend all over the cube.

Figure 57.4: Inverted Inline

Figure 58.4: Inverted Xline

67 | P a g e

Figure 59.4: Inverted Arbitrary Line

4.4.8. Results Model based inversion shows high impedance zones at reservoir zone with values  between 4500-6500 which is the same for for Gas Sand. By taking time slices from inverted model based cube and comparing it with same slices from reflectivity cube (Figure 59.4)(Figure60.4), more features appear showing what may could be channels matching with described geology.

68 | P a g e

Figure 60.4: Reflectivity (Right) Impedance (Left) Slice-750 (upper) Slice-780 (lower)

69 | P a g e

Figure 61.4: Reflectivity (Right) Impedance (Left) Slice-950 (upper) Slice-1600 (Lower)

70 | P a g e

Summary The project is held in The North Sea is a marginal, shallow sea on the European continental shelf. It is more than 970 kilometers from north to south and 580 kilometers from east to west, with an area of around 750,000 square kilometers. The North Sea RAC area is larger, because it includes the Skagerrak and Kattegat which connect the North Sea proper to the Baltic. The North Sea is bordered by England, Scotland, Norway, Denmark, Germany, the Netherlands, Belgium and France. In the southwest, beyond the Straits of Dover, the North Sea becomes the English Channel which connects to the Atlantic Ocean. The North Sea is a fairly shallow coastal sea and depths in the southern basin do not exceed 50m. The northern areas are deeper but are still generally less than 200m except in the Norwegian Trough, in the north-east, which is the only region of very deep water. The Dutch part of the North Sea (NCP) occupies an area of 57,000 km2 which is larger than the mainland of The Netherlands. In addition to traditional activities, like fishery and shipping, since the 1960s there have been new activities such as oil and gas exploration and exploitation, and extraction of sand from the seabed. New  plans for installation of wind turbines and coastal extensions are in progress. In addition the area forms part of a valuable marine ecosystem. Knowledge of the morphology of the seafloor and the sediment properties is important, both to understand the ecosystem as well as for developing infrastructural activities. The available data is a 3D seismic cube in Dutch sector with 24*16 km square consists of 740 in lines and 949 xlines and it is post stack cube there were four drilled wells in the area in the logs that is measured are •

F02 – 1 (Caliper, Density, GR, P-Wave, Porosity)

•

F03 – 2 (Density, GR, P-Wave, Porosity)

71 | P a g e

•

F03 – 4 (Density, GR, P-Wave, Porosity)

•

F06 – 1 (Density, GR, P-Wave, Porosity)

Our target is at late Jurassic to early cretaceous and the trap is shallow full of  biogas and the formation name is fs8 which is gas sand Then use Hampson Russell to convert reflectivity cube to impedance cube so the bright spot appear on the section could be an indication of gas presence Conclusion The reason that we use seismic inversion that it allow to make interpolation  between wells in the area so that the view of impedance extended from only under the wells to the whole cube and in opposite to reflectivity cube the inverted do not deal with layer interfaces but deal with zones of layer give the advantage of interpreting the area and gives push in reservoir characterization as we can use impedance in the prediction of porosity so evaluation of hydrocarbon in reservoir is enhanced by inversion and using this porosity in permeability computation and shale volume and effective porosity

72 | P a g e

References Abreu, V S and Anderson, JB 1998:" Glacial eustasy during the Cenozoic: sequence stratigraphic implications" American Association Petroleum Geologists Bulletin, 82, 1385-1400. Ad Stolk and Cees Laban, 2002:" The Dutch Sector of the North Sea ". Allen D J, Brewerton L J, Coleby L M, Gibbs B R, Lewis M A, MacDonald A M, Wagstaff SJ and Williams A T , 1997:" The physical properties of major aquifers in England and Wales, British Geological Survey Technical" Report WD/97/34, 312pp, Environment Agency R&D Publication 8. Andrews, I.J., Long, D.; Richards, P.C.; Thomson, A.R.; Brown, S; Chesher, J.A. and McCormac, M, 1990:" United Kingdom offshore regional report: the geology of the Moray" HMSO, London. Associated British Ports, 1996:"Southern North Sea sediment transport study. Literature Review and conceptual sediment transport model" ABP Research and Consultancy Ltd. Report No. R546. Balson, P. S., 1992:" The Geology of the Southern North Sea." United Kingdom Offshore Regional Report. London: HMSO for the British Geological Survey. Balson, P. S., 1999:" The Holocene Coastal Evolution of Eastern England: Evidence from the Southern North Sea” Proceedings of Coastal Sediments ’99. Pp1284-1293. Jean Virieux, 2015: " Hierarchical Seismic Imaging: A Multi-scale Approach" CSEG Recorder.

73 | P a g e