• Tidak ada hasil yang ditemukan

Adapun saran yang dapat diberikan pada penelitian ini adalah sebagai berikut:

1. Data sumur log yang lebih lengkap dan sudah dianalisis atau dilakukan pre-process dengan baik. Data log wajib memiliki data log Vs/DTSM dalam data sumur log tersebut.

2. Seismic gathers yang terkait dengan saran no.1 sehingga proses pre-stack inversion akan lebih ideal dalam melakukan perbandingan.

3. Penelitian selanjutnya diharapkan dapat menciptakan model arsitektur Neural Network yang lebih optimal

Universitas Pertamina - 48 lebih detail sehingga hasil prediksi lebih optimal.

DAFTAR PUSTAKA

Akhundi, H., Ghafoori, M., & Lashkaripour, G. (2014). Prediction of Shear Wave

Velocity Using Artificial Neural Network Technique, Multiple Regression and

Petrophysical Data: A Case Study in Asmari Reservoir (SW Iran). Open

Journal of Geology, 04(07), 303–313. https://doi.org/10.4236/ojg.2014.47023

Batzle, M. L., Kan, T. K., & Castagna, J. P. (2016). Rock Physics - The Link

Between Rock Properties and AVO Response. SEG.

Blevin, J. E., Struckmeyer, H. I. M., Cathro, D. L., Totterdell, J. M., Boreham, G.

J., Romine, K. K., Loutit, T. S., & Sayers, J. (1998). Tectonostratigraphic

Framework and Petroleum Systems of the Browse Basin, North West Shelf.

The Sedimentary Basins of Western Australia.

Bre, F., Gimenez, J. M., & Fachinotti, V. D. (2018). Prediction of wind pressure

coefficients on building surfaces using artificial neural networks. Energy &

Buildings, 158, 1429–1441. https://doi.org/10.1016/j.enbuild.2017.11.045

Castagna, J. P., Batzle, M. L., & Eastwood, R. L. (1984). Relationships between

compressional and shear-wave velocities in clastic silicate rocks. 1984 SEG

Annual Meeting, SEG 1984, 50(4), 582–584.

https://doi.org/10.1190/1.1894108

Crain, E. R. (1971). Prediction of Well Log Interpretation Parameters. 3rd

Formation Evaluation Symposium, CWLS.

https://www.spec2000.net/06-velocity.htm

Fehler, M., & Keliher, P. J. (2011). 2. Model Development. In SEAM Phase 1:

Challenges of Subsalt Imaging in Tertiary Basins, with Emphasis on

Deepwater Gulf of Mexico (pp. 15–50). Society of Exploration Geophysicists.

https://doi.org/10.1190/1.9781560802945.ch2

Freedman, D. A. (2009). Statistical models: Theory and practice. In Statistical

Models: Theory and Practice. Cambridge University Press.

https://doi.org/10.1017/CBO9780511815867

Gardner, G. H. F., Gardner, L. W., & Gregory, A. R. (1974). FORMATION

VELOCITY AND DENSITY—THE DIAGNOSTIC BASICS FOR

STRATIGRAPHIC TRAPS. GEOPHYSICS, 39(6), 770–780.

https://doi.org/10.1190/1.1440465

Goodway, B., Chen, T., & Downton, J. (1997). Improved AVO fluid detection and

lithology discrimination using Lamé petrophysical parameters;

“λρ”,“μρ”,&“λ/μ fluid stack”, from P and S inversions. SEG Technical

Program Expanded Abstracts 1997, 183–186.

Greenberg, M. L., & Castagna, J. P. (1992). Shear‐Wave Velocity Estimation in

Porous Rocks: Theoretical Formulation, Preliminary Verification and

Applications. Geophysical Prospecting, 40(2), 195–209.

https://doi.org/10.1111/j.1365-2478.1992.tb00371.x

Hadi, F. A., & Nygaard, R. (2018). Shear wave prediction in carbonate reservoirs:

Can artificial neural network outperform regression analysis? 52nd U.S. Rock

Mechanics/Geomechanics Symposium, September 2018.

Mason, H., Upton, G., & Cook, I. (2000). Understanding Statistics. The

Mathematical Gazette, 84(499), 157. https://doi.org/10.2307/3621533

OM, V., AE, U., & AC, V. (2017). Analysis of Hydrostatic Pressure Zones in Fabi

Field, Onshore Niger Delta, Nigeria. Journal of Geology & Geophysics,

06(01). https://doi.org/10.4172/2381-8719.1000275

Russell, B. H. (1988). Introduction to Seismic Inversion Methods. In Introduction

to Seismic Inversion Methods. https://doi.org/10.1190/1.9781560802303

Singh, S., & Kanli, A. I. (2016). Estimating shear wave velocities in oil fields: a

neural network approach. Geosciences Journal, 20(2), 221–228.

https://doi.org/10.1007/s12303-015-0036-z

Veeken, P. C. H., & Silva, M. Da. (2004). Technart1June04 Veeken.pdf. 22(June),

47–70.

Whitcombe, D. N., Connolly, P. A., Reagan, R. L., & Redshaw, T. C. (2000).

Extended Elastic Impedance for fluid and lithology prediction. 2000 SEG

Annual Meeting, 67(1), 63–67.

Widiaputra. (2016). Artificial Neural Network. Dosen Perbanas.

https://dosen.perbanas.id/artificial-neural-network/

Wyllie, M. R. J., Gregory, A. R., & Gardner, L. W. (1956). ELASTIC WAVE

VELOCITIES IN HETEROGENEOUS AND POROUS MEDIA.

GEOPHYSICS, 21(1), 41–70. https://doi.org/10.1190/1.1438217

Zhang, Y., Zhong, H. R., Wu, Z. Y., Zhou, H., & Ma, Q. Y. (2020). Improvement

of petrophysical workflow for shear wave velocity prediction based on

machine learning methods for complex carbonate reservoirs. Journal of

Petroleum Science and Engineering, 192(April), 107234.

FORMULIR BIMBINGAN TUGAS AKHIR

FAKULTAS TEKNOLOGI EKSPLORASI & PRODUKSI

PROGRAM STUDI TEKNIK GEOFISIKA

Nama Mahasiswa : Loris Alif Syahputra NIM : 101116100

Nama Pembimbing : Waskito Pranowo, M.T. NIP : 116030

No. 1 Hari/Tanggal: Jum’at/20-03-2020

Hal yang menjadi perhatian:

1. Membuat Reflektifitas EEI

2. Tahapan-tahapan melakukan Inversi

Paraf Pembimbing:

No. 2 Hari/Tanggal: Jum’at/27-03-2020

Hal yang menjadi perhatian:

1. Menambahkan parameter filter tambahan dalam prediksi

2. Persiapan untuk seminar kemajuan TA

Paraf Pembimbing:

FORMULIR BIMBINGAN TUGAS AKHIR

FAKULTAS TEKNOLOGI EKSPLORASI & PRODUKSI

PROGRAM STUDI TEKNIK GEOFISIKA

Nama Mahasiswa : Loris Alif Syahputra NIM : 101116100

Nama Pembimbing : Waskito Pranowo, M.T. NIP : 116030

No. 1 Hari/Tanggal: Jum’at/10-04-2020

Hal yang menjadi perhatian:

1. Menyelesaikan tahapan inversi 2. Meningkatkan akurasi hasil prediksi

Paraf Pembimbing:

No. 2 Hari/Tanggal: Jum’at/17-04-2020

Hal yang menjadi perhatian:

1. Membandingkan hasil inversi dari data sumur dengan data prediksi

2. Konsultasi isi laporan TA

Paraf Pembimbing:

FORMULIR BIMBINGAN TUGAS AKHIR

FAKULTAS TEKNOLOGI EKSPLORASI & PRODUKSI

PROGRAM STUDI TEKNIK GEOFISIKA

Nama Mahasiswa : Loris Alif Syahputra NIM : 101116100

Nama Pembimbing : Waskito Pranowo, M.T. NIP : 116030

No. 1 Hari/Tanggal: Kamis/16-01-2020

Hal yang menjadi perhatian:

1.

Mengamati urutan parameter input untuk model ANN

2.

Melihat parameter log dengan kontribusi terbesar dalam menurunkan nilai RMSE

Paraf Pembimbing:

No. 2 Hari/Tanggal: Kamis/23-01-2020

Hal yang menjadi perhatian:

1. Menentukan parameter-parameter log dengan nilai korelasi terhadap parameter

target (Vs)

2. Menghitung nilai relative error log input dengan parameter target (Vs)

Paraf Pembimbing:

FORMULIR BIMBINGAN TUGAS AKHIR

FAKULTAS TEKNOLOGI EKSPLORASI & PRODUKSI

PROGRAM STUDI TEKNIK GEOFISIKA

Nama Mahasiswa : Loris Alif Syahputra NIM : 101116100

Nama Pembimbing : Waskito Pranowo, M.T. NIP : 116030

No. 1 Hari/Tanggal: Kamis/30-01-2020

Hal yang menjadi perhatian:

1. Melihat RMSE pada data well train, perbaiki parameter atau arsitektur hingga stabil 2. Mencoba algoritma yang lebih ringan

3. Lakukan regresi linear antara parameter log dengan output

4. Membuat model awal menggunakan metode machine learning yang lebih sederhana / cepat sebagai parameter awal untuk prediksi Vs menggunakan ANN

Paraf Pembimbing:

No. 2 Hari/Tanggal: Kamis/23-01-2020

Hal yang menjadi perhatian:

1. Tambah jumlah data well log 1 + well log 2 untuk prediksi well log 3 & well 3 +

well 2 untuk prediksi well 1

2. Perbaiki arsitektur ANN

LAMPIRAN KODE SKRIP

#Library

from keras.models import Sequential from keras.layers import Dense import matplotlib.pyplot as plt import pandas as pd

from sklearn.preprocessing import StandardScaler from sklearn.preprocessing import MinMaxScaler import tensorflow as tf

from keras.regularizers import l1_l2 import numpy as np

#Pre-process the datasets ##WELL TRAIN DATASET

dataset = pd.read_excel('Poseidon11.xlsx')

dataset = dataset.dropna(how='any') #For rows that have any nan value on it

# IQR Rule

Q4 = dataset.quantile(0.25) Q6 = dataset.quantile(0.75) IQRS = Q6 - Q4

dataset = dataset[~((dataset < (Q4 - 1.5 * IQRS)) |(dataset > (Q6 + 1.5 * IQRS))).any(axis=1)]

x_train = dataset.iloc[:, [1,3,4,5,6]].values y_train = dataset.iloc[:,7].values

##WELL TEST DATASET

datasets = pd.read_excel('Poseidon21.xlsx')

datasets = datasets.dropna(how='any') #For rows that have any nan value on it

#IQR Rule

Q1 = datasets.quantile(0.25) Q3 = datasets.quantile(0.75) IQR = Q3 - Q1

datasets = datasets[~((datasets < (Q1 - 1.5 * IQR)) |(datasets > (Q3 + 1.5 * IQR))).any(axis=1)]

x_test = datasets.iloc[:, [1,3,4,5,6]].values y_test = datasets.iloc[:,7].values

#Normalizing the datasets

nor = MinMaxScaler()

x_train = nor.fit_transform(x_train) x_test = nor.transform(x_test)

# Inserting Vs Pred to DF using MLR

from sklearn.linear_model import LinearRegression regressor = LinearRegression()

regressor.fit(x_train,y_train)

y_pred1 = regressor.predict(x_train) y_pred2 = regressor.predict(x_test)

datasets.insert(0, "Vs Pred", y_pred2)

# Using ANN

x_train1 = dataset.iloc[:, [0,2,4,5,6,7]].values x_test2 = datasets.iloc[:, [0,2,4,5,6,7]].values

#Normalizing the datasets

nor = StandardScaler()

x_train1 = nor.fit_transform(x_train1) x_test2 = nor.transform(x_test2)

# #Creating ANN Architecture

lrelu = lambda x: tf.keras.activations.relu(x, alpha=0.1) model = Sequential () #intializer

model.add(Dense(12, input_dim=6, kernel_initializer='glorot_uniform',bias_initializer = 'glorot_uniform', activation= lrelu, kernel_regularizer=l1_l2(l1=0.01, l2=0.01), bias_regularizer=l1_l2(l1=0.01, l2=0.01))) #input layer

model.add(Dense(6, kernel_initializer='glorot_uniform',bias_initializer = 'glorot_uniform', activation =lrelu, kernel_regularizer=l1_l2(l1=0.01, l2=0.01), bias_regularizer=l1_l2(l1=0.01, l2=0.01))) #hidden layer

model.add(Dense(3, kernel_initializer='glorot_uniform',bias_initializer = 'glorot_uniform', activation =lrelu, kernel_regularizer=l1_l2(l1=0.01, l2=0.01), bias_regularizer=l1_l2(l1=0.01, l2=0.01))) #hidden layer

model.add(Dense(1, kernel_initializer='glorot_uniform',bias_initializer = 'glorot_uniform', activation = 'linear', kernel_regularizer=l1_l2(l1=0.01, l2=0.01), bias_regularizer=l1_l2(l1=0.01, l2=0.01))) #output layer

model.compile(loss='mean_squared_error', optimizer='adam', metrics=['mean_squared_error']) #Compiling the ANN

# Training ANN Model

model.fit(x_train1,y_train, batch_size=10, epochs =500, verbose = 1, validation_split = 0.2)

Dokumen terkait