improved flow zone definition of the scott field reservoir ... · • fault seal capacity,...

Date goes here

Improved Flow Zone Definition of the Scott Field Reservoir

using a Quantitative Electrofacies ApproachPhilip Highton, Lyudmila Ouagueni (now Tailwind Energy) & Tatiana Christie (now

ConocoPhillips)

Date: 07 May 2019

Scott Field (Block II) Modelling Approach02

Approaches to Facies Definition for Geomodelling03

Facies Definition in the Scott Field04

Summary & Conclusions05

Introduction01

Presentation Structure

Scott and Telford Fields -Description

Block 1

Block

1A

Block 3

Block 4

Marmion

Telford

• Large oilfield discovered in 1983

• Structurally complex field with several fault-

bounded pressure compartments

• Two main reservoir units of Upper Jurassic age:

• Piper Formation

• Scott Member (Sgiath Formation)

• Field property ranges

• N:G ~ 0.85 - 0.90

• Average Ø ~ 10% - 20%

• Permeability ~ up to 8D

BCU

KPT

(top reservoir)

Saltire Shale

(base reservoir)

Hudlestoni MFS

Kim

me

rid

gia

nV

olg

ian

Oxfo

rdia

n

Rosenkrantzi MFS

Serratum MFS

‘Cinctum MFS’

KPT DSB

M.Piper DSB

U.Scott FS

MFS B

MFS C DSB ?

Base KPT=Baylei MFS

=AOI of Block II Geomodel

Block 2A

Block 2

Stratigraphy

Introduction – Orientation


Cla

y c

on

ten

t

Gra

in s

ize

Excellent quality

clean coarse

sands

Poor quality

shaley sands

Large range of permeabilities for a given porosity, depending mainly on clay content, grain size and

quartz cementation

Sedimentary Facies

Diagenesis

Tar MatPresence

Small ScaleFaulting

Tar Mat

In Core

No Tar Mat

In Core

White Light

U.V. Light

White Light

U.V. Light

Introduction – Controls on Reservoir Quality


• Aim of modelling work: to identify areas of unproduced oil poorly accessed by current

well-stock

• Single structural model selected dynamically from a large range of possibilities

• Variable geological inputs include

• Fault seal capacity, Electrofacies vs FZI, Facies variogram ranges, Tar mat

presence/absence and permeability modifiers, Free-water level

• Selection of geological models using quality of match to the dynamic for subsequent

history matching.

DYNAMIC SCREENING OF REALISATIONS ALLOWS US TO TEST THE EFFACY OF FACIES MODELLING APPROACHES

Scott Field (Block II) Modelling Approach

• Facies Analysis

• Sedimentary facies (Lithofacies)

• Subjective characterisation based upon lithology and sedimentary structures

• Core dependent – large amount of error in defining facies in uncored wells

• Facies often described in terms of interpreted depositional processes

• Diagenesis not usually incorporated into classification

• Categories not usually optimised for flow unit definition (although there is usually some relationship)

6

• Approach to Block II

• Trial analysis indicated the logged core lithofacies had a relatively poor match to wireline log values

• Unable to objectively populate facies throughout the geomodel

• High degree of scatter on PoroPerm plots for different lithofacies indicated reduced utility for populating Geomodel facies

DECIDED NOT TO APPLY SEDIMENTARY FACIES TO POPULATE GEOMODEL

Core

Lithofacies

Log

Approach to Facies Definition for Geomodelling

• Facies Analysis

• Flow Zone Indicators (FZI’s)

• A theoretical method to define distinct rock types with similar fluid-flow characteristics

• Should be related to rock texture, grainsize, mineralogy, sedimentary structures and petrophysical properties

• Based upon the relationship between porosity and permeability

• FZI rock types classified using equations:

• RQI=0.0314*SQRT(Permeability/Porosity)

• FZI= RQI/NormPorosity

• Relationships between FZI and wireline log data to predict FZI rock types in uncored wells

7


• Defined different FZI rock types using core analysis data

• Log data showed few reliable predictive relationships with FZI rock types

• Petrophysically derived porosity and permeability values used to classify FZI rock types in uncored wells

APPLIED FZI ROCK TYPING APPROACH TO POPULATE GEOMODEL

Porosity vs

Permeability vs

FZI Rock Types


• Facies Analysis

• Electrofacies

• Facies description based upon measured log parameters

• Objective classification using statistical, machine learning or artificial Intelligence

• Core used to calibrate poroperm relationships

• Can be used for wells without core

• Categories are based upon physical and chemical rock characteristics and a strong link to reservoir properties is common


• Quantitative approach

• Calibrated with data from14 cored wells

• Trial analysis indicated the identified electrofacies could be predicted from wireline log values

• Provided a means to populate facies throughout the geomodel

• Relatively tight scatter of core plug data on PoroPerm plots for different electrofacies indicated good utility for populating Geomodel facies

Cluster Analysis

Discriminant

Function

Analysis

APPLIED ELECTROFACIES FACIES APPROACH TO POPULATE GEOMODEL


• Cluster Analysis

• Used heavily in biology, ecology, palaeontology, geochemistry and quantitative finance

• Objectively groups samples based upon multiple measured parameters

• Discriminant Function Analysis

• Used heavily in biology, ecology, palaeontology, geochemistry and quantitative finance

• Objectively groups samples into pre-defined categories based upon multiple measured parameters

• Two techniques can be used together to objectively identify different rock types using cluster analysis (available core can be used to define poroperm characteristics) then applied to uncored wells using discriminant function analysis

9

Electrofacies Definition in the Scott Field

Dissimilarity #1 #2 #3 #4 #5 #6 #7 #8 #9 #10

#1

#2 0.59#3 0.5 0.46

#4 0.55 0.42 0.48

#5 0.92 0.6 0.87 0.88

#6 0.76 0.5 0.72 0.81 0.8#7 0.93 0.7 0.88 0.89 0.79 0.83

#8 0.92 0.89 0.88 0.91 0.81 0.87 0.75

#9 0.95 0.87 0.86 0.92 0.82 0.88 0.66 0.69

#10 0.95 0.95 0.91 0.97 0.89 0.88 0.8 0.75 0.61

• Cluster Analysis – Used to define Electrofacies in cored wells

10

Cluster analysis is a data exploration (mining)

tool for dividing a multivariate dataset into

“natural” clusters (groups).

Variable 1

Va

ria

ble

2

Define appropriate suite of

logs(GR, Neutron, Density &

Sonic) Calculate sample

dissimilarity

Euclidian Distance = distance in

n dimensions

Calculated dissimilarity matrix Clustering successively linking

more dissimilar samples

Cluster Analysis can classify large numbers of samples with many variables


Sample 1

Sample 2

x

• Discriminant function analysis is a statistical analysis to predict a categorical dependent variable (called a grouping variable) by one or more continuous or binary independent variables (called predictor variables).

11

Variable 1

Resultant duel variable distributions

along a discriminant function

a

b

a b

Va

ria

ble

2Class AClass B

Discriminant Function Analysis can use many variables for classification


• Discriminant function analysis is a statistical analysis to predict a categorical dependent variable (called a grouping variable) by one or more continuous or binary independent variables (called predictor variables).

12

Discriminant Function Analysis can use many variables for classification

Variable 1

Resultant duel variable distributions

along a discriminant function

a

b

a b

Va

ria

ble

2

New unknown sample

Class AClass B

New sample

belongs to Class B


• Workflow Summary

13

Objectively identify

electrofacies from wireline log

data using Cluster Analysis on

cored wells

Ensure identified clusters have

distinct poroperm and flow unit

character

Ensure identified clusters look

geological

Define discriminant functions

using cored well training set

Use discriminant functions to

define facies in uncored wells

from wireline log data

Understand Geological

Significance of Clusters

Model Facies in Geomodel

Truth Test with Quality of Dynamic Matches (Objective Functions)


• Cluster Analysis Results – Electrofacies Groups

Net Electrofacies Class Centroids

1

9

7,5 6 2 3 8 10 4 11

10

1 Predominantly Non-Net Electrofacies

Predominantly Net Electrofacies


CLASS DENSITY GAMMA NEUTRON SONIC

3 2.397 20.334 0.104 72.549

4 2.463 19.651 0.076 66.673

8 2.393 13.873 0.086 71.212

10 2.465 12.917 0.05 64.813

11 2.544 19.266 0.052 60.326

Tightened scatter than lithofacies

Can be identified with 86% success from log data

Can be populated through geomodel

• Net Rock Electrofacies Porosity vs Permeability from Core

Electrofacies 3 Electrofacies 4 Electrofacies 8

Electrofacies 10 Electrofacies 11


Density of sample points

suggest further cluster

subdivision possible but

would need to be

geologically based



• Net Rock Electrofacies Porosity vs Permeability from Core


Lower permeability

grouping is strongly

stratigraphically

constrained: Highlighted

points Lower Scott C

Allows further refinement of

poroperm groups for

modelling

Lower K

population

stratigraphically

restricted

L.Scott C

L.Scott

A/B

U.Scott

U.Scott




• Electrofacies distribution appears geological

• 15 of the 17 stochastic realisations taken to history matching were populated with Electrofacies rather than FZI rock indicating that they provided a better match to the subsurface ‘plumbing’

19

NESW

3475m section length

SW NEA1Z 35 A2 A4 J3Z J42

Section

through a

model

realisation

showing

populated

electrofacies

distribution

Well correlation panel showing electrofacies

Section location map

Electrofacies Definition in the Scott Field – Populated Geomodel

• Block II of the Scott Field has been stochastically modelled using variable Geological inputs

• Sedimentary facies were considered but not used in the Scott Block II model

• Both Flow Zone Indicators (FZI’s) and Electrofacies have been used as alternatives for facies modelling

• Electrofacies have been defined using ‘Cluster Analysis’

• Core analysis data was used to define PoroPerm character

• Discriminant Function Analysis was trained using the Cluster Electrofacies groupings and applied to uncored wells

• An important stratigraphically constrained difference in PoroPerm character has been identified in the Lower Scott C reservoir zone significantly improving permeability prediction in this and the overlying reservoir zones

• 90% of the successfully screened model realisations were constructed using the Electrofacies approach

• High quality of dynamically screened realisations significantly reduced Reservoir Engineer patching

20

Summary and Conclusions

improved flow zone definition of the scott field reservoir ... · • fault seal capacity,...

Documents