- Email: [email protected]
Have you ever heard the sound of well logs or reservoir data?
Accepted Manuscript Have you ever heard the sound of well logs or reservoir data? Ali Kadkhodaie, Reza Rezaee PII:
S0920-4105(17)30514-4
DOI:
10.1016/j.petrol.2017.06.014
Reference:
PETROL 4026
To appear in:
Journal of Petroleum Science and Engineering
Received Date: 5 April 2017 Revised Date:
30 May 2017
Accepted Date: 5 June 2017
Please cite this article as: Kadkhodaie, A., Rezaee, R., Have you ever heard the sound of well logs or reservoir data?, Journal of Petroleum Science and Engineering (2017), doi: 10.1016/ j.petrol.2017.06.014. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Have you ever heard the sound of well logs or reservoir data? Ali Kadkhodaie1,2* and Reza Rezaee1 1. Department of Petroleum Engineering, Curtin University, Perth, Western Australia
RI PT
2. Earth Science Department, Faculty of Natural Science, University of Tabriz, Iran
Abstract
SC
The current study presents an effective approach to convert well logs to music and listen to their sound. Well log data play an important role in different stages of oil and gas field’s exploration and
M AN U
development. Several rock properties such as porosity, lithology, fluids saturation, fluid contacts and pay zones can be obtained through interpretation of well log data. Boring tabulated data or a mass of curves can be converted into joyous and pleasant sounds. For this purpose, four case studies were run to show how borehole quality logs, petrophysical evaluation results, capillary pressure data (pore size distribution) and drilling data can be converted to musical notes. The proposed approach can
TE D
help in quality control or interpretation of well logs or any other reservoir data. The interpreter just needs to wear wireless headphones and listen to the music generated from reservoir data. Service companies may consider the musical interpretation of well logs as an alternative and immediate way
EP
for quality control of logging procedures at well bore site. Meanwhile, visually disabled petroleum engineers may use the aural interpretation of subsurface data. A siren can be sounded when lost
AC C
circulation occurs or a kick is detected to warn well site crew about the drilling risks. The result of musical transformed well logs can be stored in MP3 files for future applications. Keywords: Well logs, sonification, audification, music, quality control, petrophysical interpretation
1. Introduction *
Corresponding author: Tel/Fax: +618 9266 9366 Email addresses: [email protected]; [email protected] (Ali Kadkhodaie), [email protected] (Reza Rezaee)
1
ACCEPTED MANUSCRIPT Subsurface data plays an important role is rock properties estimation and reservoir characterization. They provide very useful information for uncored intervals and making correlations in inter-well spaces. Reservoir rock properties are estimated from interpretation well log data (e.g. Rezaee et al., 2007 & 2008; Ostadhassan et al., 2012 & 2015; Sfidari et al., 2012 & 2014; Zamiran et al., 2015; Kadkhodaie et al., 2016 & 2017 ). Reservoir static models incorporate different rock properties in the
RI PT
form of well logs and propagate them through a simulation grid. Petrophysicists often deal with a mass of curves and boring tabulated well-logging data for each drilled well in a hydrocarbon field. Accordingly, finding alternative ways for quality control and interpretation of petrophysical data
SC
would be worth. Recently, sonification or audification of numerical data has received a considerable attention in different disciplines of science and engineering. Such approaches can help geologists and
M AN U
petrophysicists make aural interpretation of well logs. There are different approaches and methods for data audification or sonification. The first work published on the geoscience’s data sonification is “Listening to the Earth sing” by Hayward (1994). He proposed the idea of audio seismograms and made audio facilities accessible to geophysicists for sonification of seismic attributes. One of the main advantages of Hayward’s approach was the quick-look analysis of audio seismic data besides
TE D
their visual display. Quinn and Meeker (2001) introduced the climate symphony and other audifications such as the sound of seismic data and solar winds. Ekdale and Tripp (2005) introduced sonification of paleontological data and discussed how musical interpretations can help geologists to
EP
interpret subsurface data. Stewart and Brough (2006) studied the musical interpretation of well logs.
AC C
They discussed how audible expression of well-logging data can aid in the immediate identification of lithology changes. Quintero (2013) worked on the turning numerical values of wireline logs into different musical notes or sounds as an alternative way for wireline logs interpretation or making blind scientists aware of the downhole rock and fluids properties. Recently, McGee and Rogers (2016) worked on sonification of seismic data through making a variety of seismically based timbres. They stated that seismic data are the ultimate source of earth’s sound generation because they are in the form of physical waves. They can generate different songs directly using time-stretching and pitch-shifting without mapping to already available sounds. The current study presents an approach for mapping subsurface data to a set of predefined musical notes 2
ACCEPTED MANUSCRIPT by using the Matlab software (MATLAB 2016). The outcome music is helpful in quality control and aural interpretation of well logs. What makes the current study different from the previous ones is that in addition to well logging data, it focuses on the sonification of rock property logs, drilling sensor data and reservoir data such as capillary pressure curves. The musification codes generated in this study are freely available to future researchers working on exploration, description and
RI PT
development of oil and gas fields.
2. Methodology
SC
Audification is a technique for representation of a series of data as a sound. The input of Audification can be a time series or a set of data versus time or depth. Sonification or Audification
M AN U
can help researchers in the different discipline of science to create different songs and music pieces from a sequence of data (e.g., Dombois et al.,2001; Helmen and Ritter, 2004; Olivan et al., 2004; and Pauletto et al., 2005). A music piece or composition comprises three main component including melody, harmony and rhythm. When notes of different pitches are played, the outcome is called a
TE D
melody. Sometimes, melodies can be singable, however only the sequence of pitches does not create a melody. It is often possible to sing melodies that are distinguishable. By definition “harmony is usually referred to as simultaneously occurring frequencies, pitches (tones, notes), or chords (Malm,
EP
1996)”. Harmony is often defined as the vertical dimension of music, as opposed to the melodic line, or the "horizontal" dimension (Jamini, 2005). Each note has a certain duration and the relationship
AC C
between the durations of notes is referred to as rhythm. Seven basic musical notes are “Do, Re, Mi, Fa, Sol, La, Si”. In English and German-speaking countries, they are represented by 'C, D, E, F, G, A, B' letters (Fig. 1). Each basic note has a different pitch for which frequency increases from Do to Si syllable. The eighth note will have the same name as the first one and so on. The span between the first and eighths note is called one octave. Going up one octave in pitch will double the frequency. The combination of three or more musical notes creates a Chord. Accordingly, different songs and melodies can be created through a combination of the aforementioned musical notes. For example, the music sheet of the Old French Song (Tchaikovsky, 2005) is shown in Fig. 2. Another example 3
ACCEPTED MANUSCRIPT illustrating how the bell ring music sheet was mapped to gamma ray log signatures is shown in Fig. 3. The main purpose of the current research is to create a music piece through mapping the basic musical notes to the numerical value of well logs or reservoir data which is interpretable by experts in the petroleum industry. For this purpose, four case studies were run to create songs from borehole
RI PT
quality logs (gamma ray and caliper), petrophysical evaluation results (porosity and water saturation), capillary pressure data (pore size distribution) and drilling data (rate of penetration). The data chosen for this study come from one of the wells drilled in the Crismon Lake field which is
SC
located in the Fitzory Trough, Canning Basin, Western Australia. The Laurel Formation with a complex clastic-carbonate lithological sequence is the main studied interval in Crimson Lake field.
M AN U
The main lithological composition consists of sandstone with interbedded siltstone and claystone and thin beds of carbonate rocks. The sandstone is very 1ight grey to white, 1ight to medium brown and occasiona1 1ight green grey, dominantly with clear, transparent quartz, trace colored quartz and lithic fragments. Accessory grains include trace lithic fragments, glauconite and carbonaceous flecks
TE D
with trace to conman pyrite as loose aggregates and as cement.
3. Results and discussion
Following function of Matlab program (MATLAB 2016) was used to convert a matrix of well logs
EP
and other reservoir data to sound:
(Eq. 1)
AC C
sound(α*data, β*Fs)
where data is the matrix of a predefined song data such as Christmas bell ring, guitar, trumpet, piano or any other sound. It can be an n by 1 column vector for mono or single channel audio files or an n by 2 matrix data in the case of stereo playback. The multiplier α in front of song data sets the master volume, while the multiplier β controls the frequency or pitch of the sound. The frequency parameter, Fs, is the sample rate of the audio file. It depends on the hardware restrictions and can be a value in the range of [1000 Hz, 384000 Hz] for Matlab software (MATLAB 2016). Generally, it can be said that a high pitch sound corresponds to a high-frequency sound wave and a low pitch sound corresponds to a low-frequency sound wave. If the frequency of a sound is set 4
ACCEPTED MANUSCRIPT to Fs = 2*Fs1 the frequency will be cut in half and the pitch will go down as high as one octave. Similarly, setting Fs=0.5*Fs1 will double the frequency and the pitch will go up as high as one octave. Well log data were introduced in the form of α or β matrices to control the master volume and
RI PT
pitch of the chosen song. Depth was considered as the rhythm of the song and playing different pitches created well log melodies. The following examples show different song composed based on different datasets of a drilled well.
Borehole quality logs song: caliper logs in conjunction with gamma ray log are used to detect an
SC
enlarged region of a borehole and bad hole intervals. Washouts and cavities develop due to several reasons such as swelling or weakening of clays as they come in contact with mud filtrate,
M AN U
unconsolidated formations and rock mechanical failure caused by in-situ stresses. The first example of this study focuses on the simultaneous sonification of gamma ray and caliper logs for a test well. The pitch of the drum sound with a sampling frequency of 11025 Hz was mapped to the caliper log. For the case of gamma ray log, the ring sound with a frequency of 44100 Hz was chosen and its pitch
TE D
was mapped to the fluctuation of GR values. As is seen in Fig. 4(a), high gamma-ray intervals can be recognized through changes in the tonality and pitch of the bell ring and piano song. As with the gamma ray log, washout intervals can be identified through listening to the caliper log song (Fig. 4b).
EP
Interpreted well logs song: porosity, shale volume, water saturation and mineral fractions are the
AC C
main outputs calculated from the interpretation of well logs. The second case focuses on simultaneous sonification of porosity and water saturation logs as an example from interpreted petrophysical data (Fig. 5). Porosity log was mapped to the Santoor ding with the base frequency of 44100 Hz. Similarly, water saturation log was correlated to drum sound with the base frequency of 11025 HZ. The codes were modified so that if water saturation is above 50%, then the drum beats will be played. Listening to the joint composition of the porosity and water saturation songs will tell us information regarding good reservoir quality zones for which PHIE is high and Sw falls below 50%. Even, the changes in lithology, rock texture, pore system or other rock and fluid properties can be tracked by listening to the musical interpretation of well logs. Computer Processed Interpretation 5
ACCEPTED MANUSCRIPT (CPI) plot of the studied interval is shown in Fig. 6. The plots shown on left tracks are gamma ray (GR), neutron porosity (NPHI), sonic (DT), density (RHOB) and photo-electic (PEF) logs along with resistivity curves (MSFL, LLS and LLD) of the main reservoir interval (Laurel formation). The petrophysical interpretation of the reservoir rocks including lithology, water saturation and porosity are shown on the right tacks of the plot as well.
RI PT
Capillary pressure curves song: Capillary pressure curves are used for calibration of log-derived water saturation above the transition zone, determination of height above free water level and pore size distribution (PSD) analysis. The third example deals with sonification of pore size distribution
SC
curve determined from capillary pressure data. For this purpose, PSD data were mapped to the Santoor dings and the resulting song will tell us about the pore size, storage capacity and actually
M AN U
reservoir quality of the studied sample. The base frequency of the composed Santoor ding is 44100 Hz. Graphical representation of the pore size distribution data mapped to the Santoor song is shown in Fig. 7.
Drilling data song: there are different types of measurement while drilling data (MWD) used to derive useful information from subsurface rocks and mechanical strength of the formations being
TE D
drilled. Drill sensor data such as rate of penetration (ROP), bit pressure, rotation pressure, head speed, pull down rate and pressure recorded from drill rigs can be sonified by using the approach presented in this study. As an example, the rate of penetration of drill bit for the studied well was
EP
mapped to Christmas bell ring. The base frequency of the composed Christmas bell ring is 44100 Hz.
AC C
As is seen in Fig. 8. the pitch of the composed song changes when the penetration rate increases or decreases and somehow it is indicative of the rock type or changes in their strength. Similarly, a siren can be sounded when lost circulation occurs or a kick is detected to warn well site engineers about the risk of blowout or loss of well. MP3 and Video files of the sonified data discussed above can be found in the following permanent link: https://goo.gl/XfWnki The same procedure explained for the shown examples can actually be applied to any reservoir data to make new songs. Even a 3D grid of reservoir static and dynamic properties can be sonified and transformed into the musical models of rock properties. They can sing pleased songs when mapped 6
ACCEPTED MANUSCRIPT and adjusted to the joyous melodies. Matlab codes (MATLAB 2016) used for the sonification of the case study data are shown in Appendix 1.
4. Conclusions The results of this study showed that sonification of well logs can be considered as an alternative
RI PT
approach in rock properties analysis. Two cases of well logs quality control and interpretation were tested and satisfactory results were achieved. Fluctuations of well logs responses were mapped to the musical notes and their properties such as pitch, volume and tonality. The created music and songs
SC
not only provide joyous and pleasant moments but also make it possible to have a quick interpretation of well log signatures. Through a change in tonality or pitch of the created music, the
M AN U
interpreter can track changes in rock properties or quality of well logs. The proposed approach can be applied to any other reservoir information such as drilling data, production data, flow rate and history match curves. The output musical data can be stored in audio files with popular formats such as MP3 or WAV. It is expected that in near future, clients ask their service companies or contractors to provide an audio file of the well logs in addition to their hard copy or digital files. Matlab codes
References
EP
TE D
for sonification of well logs, reservoir and drilling data are listed in Appendix 1.
AC C
Dombois, F., 2001.Using audification in planetary seismology. Proceedings of the 7th International Conference on Auditory Display (ICAD2001): 227–230 Ekdale, A., and Tripp, A., 2005. Paleontological sonification: Letting music bring fossils to your ears. Journal of Geoscience Education, 53, 271-280. Hayward, C., 1994. Listening to the earth sing. in Auditory Display: Sonification, Audification, and Auditory Interfaces. Addison-Wesley, 1994. Hermann, T., Ritter, H., 2004. Sound and meaning in auditory data display" (PDF), Proceedings of the IEEE, IEEE, 92 (4), 730–741. doi:10.1109/jproc.2004.825904 7
ACCEPTED MANUSCRIPT Jamini, D., 2005. Harmony and Composition: Basics to Intermediate, p.147. ISBN 1-4120-3333-0. Kadkhodaie, A., Rezaee, R., 2016. A new correlation for water saturation calculation in gas shale reservoirs based on compensation of kerogen-clay conductivity. Journal of Petroleum Science and Engineering. Vol. 146, 932-939
RI PT
Kadkhodaie, A., Rezaee, R. 2017. Estimation of vitrinite reflectance from well log data. Journal of Petroleum Science and Engineering 148, 94-102.
Malm, W.P.,1996. Music Cultures of the Pacific, the Near East, and Asia, p.15. ISBN 0-13-182387-
SC
6. Third edition.
MATLAB, 2016 version. The MathWorks, Inc., Natick, Massachusetts, United States.
M AN U
McGee, R., Rogers, D., 2016. Musification of seismic data. The 22nd International Conference on Auditory Display (ICAD–2016). Canberra, Australia.
Olivan, J., Kemp, B., Roessen, M. 2004. Easy listening to sleep recordings: tools and examples. Sleep Medicine, 5 (6): 601–603, doi:10.1016/j.sleep.2004.07.010
TE D
Ostadhassan, M., Zeng, Z., Zamiran, S., 2012. Geomechanical modeling of an anisotropic formationBakken case study. 46th US Rock Mechanics / Geomechanics Symposium 2012, At Chicago, Illinois, USA.
EP
Ostadhassan,M., Zamiran, S., Jabbari, H., 2015. Stability Analysis of Multilateral High Density Pad
AC C
Wells in the Three Forks Formation. SPE Western Regional Meeting, At Garden Grove, California, USA.
Pauletto, S., Hunt, A., 2005. A comparison of audio & visual analysis of complex time-series data sets. Proceedings of the 11th International Conference on Auditory Display (ICAD2005): 175–181 Quinn, M., and Meeker, L., 2001. Research set to music: The climate symphony and other sonifications of Ice Core, Rada, DNA, Seismic and Solar wind data: Proceedings of the 2001 International Conference on Auditory Display, Espoo, Finland. 56-61.
8
ACCEPTED MANUSCRIPT Quintero, G., 2013. Sonification of oil and gas wireline well logs. International Conference on Auditory Display. Pp. 301-306, 6-10 July, Poland. Rezaee, M.R, Kadkhodaie, A., Barabadi, A., 2007. Prediction of shear wave velocity from petrophysical data using intelligent systems, a sandstone reservoir of Carnarvon Basin. Journal of
RI PT
Petroleum Science and Engineering 55, 201-212. Rezaee, M.R., Kadkhodaie, A., 2008. Intelligent approach for the synthesis of petrophysical logs. Journal of Geophysics and Engineering 5, 12-26.
SC
Sfidari, E., Kadkhodaie, A, Najjari, S., 2012. Comparison of intelligent and statistical clustering approaches to predicting total organic carbon using intelligent systems. Journal of Petroleum Science
M AN U
and Engineering 86-87, 190-205
Sfidari, E., Kadkhodaie, A., Rahimpour-Bonab, H., Soltani, B., 2014. A hybrid approach for lithofacies characterization in the framework of sequence stratigraphy: a case study from the South Pars gas field, the Persian Gulf basin. Journal of Petroleum Science and Engineering 121, 87-102.
TE D
Stewart, R.R., Brough, S., 2006. Log jammin’: transforming well logs to music. CREWES Research Report Number: Vol 18. SEG Technical Program Expanded Abstracts 25(1), 1-18.
EP
Tchaikovsky, P.I., 2005. Old French Song music sheet. [email protected] Zamiran, S., Osouli, A., Ostadhassan, M., 2014. Geomechanical modeling of inclined wellbore in
AC C
anisotropic shale layers of bakken formation. 48th US Rock Mechanics / Geomechanics Symposium, At Minneapolis, Minnesota, USA
9
ACCEPTED MANUSCRIPT
RI PT
Fig. 1. Seven basic music syllables: “Do Re Me Fa Sol La Si”. Each basic note has a different pitch
AC C
EP
TE D
M AN U
SC
for which frequency increases from Do to Si syllable
10
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 2. Music sheet of the Old French Song (Tchaikovsky, 2005).
11
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3. Bell ring music sheet for piano (Copyright 2015 Red Balloon Technology Ltd.) mapped to
AC C
EP
TE D
gamma ray log.
12
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 4. High gamma ray intervals can be recognized through changes in the tonality and pitch of the bell ring in piano song (a). Washout intervals can be identified through mapping drum sound to
AC C
EP
TE D
borehole quality logs such as Caliper log.
13
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig 5. Porosity and water saturation mapped to Santoor and Drum song, respectively. Listening to the composed song helps in identification of high porosity zones. If water saturation is over 50%, then drum beats will be played.
14
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6. Computer Processed Interpretation (CPI) plot of the studied interval. The main lithology composition is shale and sandstone with some contributions of carbonate rocks in the lower interval.
15
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 7. Graphical representation of pore size distribution data mapped to Santoor song
16
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
Fig. 8. Graphical representation ROP (Rate of penetration) sonified based on Christmas bell ring song. The composed song will aid in understanding rock types and strength of the rocks being drilled.
17
ACCEPTED MANUSCRIPT
M AN U
% Sonify GR and Caliper logs for i=1:size(CALI,1)
SC
% By Ali Kadkhodaie March 2017 %CASE 1:sonification of GR-CALI logs % Load Christmass bell ring data matrix Load Ring.mat % Load Christmass Drum data matrix Load Drum.mat %Plot input data subplot(2,1,1) plot(depth,CALI) ylabel('CALI'); subplot(2,1,2) plot(depth,GR) ylabel('GR'); xlabel('Depth');
RI PT
Appendix 1: Matlab codes for sonification of well logs, reservoir and drilling data
sound(Ring(10000:20000,:),.05*GR(i)*fs_Ring) if CALI(i)-8.5>9.5 sound(2*Drum(70000:72000,:),fs_Drum*0.5) else
TE D
end pause (10000/fs_Ring)
EP
hold on subplot(2,1,1) hold on plot(depth(i),CALI(i),'ro') hold on subplot(2,1,2) hold on plot(depth(i),GR(i),'ro')
AC C
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 2:sonification of Phie-Sw logs % Load Santoor ding data matrix load Santoor.mat %Plot input data subplot(2,1,1) plot(depth,phie,'r') ylabel('PHIE'); subplot(2,1,2) plot(depth,Sw) ylabel('Sw (v/v)'); xlabel('Depth (m)'); % Sonify porosity and water saturation logs for i=1:size(phie,1) sound(Ding*phie(i),.2*fs_Ding) 18
ACCEPTED MANUSCRIPT if Sw(i)>0.5 sound(Drum(70000:72000,:),fs_Drum*0.5) else end pause (10000/fs_Ding)
RI PT
hold on subplot(2,1,1) hold on plot(depth(i),phie(i),'ro') hold on subplot(2,1,2) hold on plot(depth(i),Sw(i),'go')
M AN U
SC
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 3:sonification of PSD data % Load Santoor ding data matrix Load Ring.mat %Plot input data plot(Ri_SampleNo,Dist) ylabel('Dist (Cm2)'); xlabel('Pore radius (micron)');
% Sonify Pore Size Distribution data for i=1:size (Dist,1)
TE D
sound(2*Ding,((Dist(i)-min(Dist))/(max(Dist)-min(Dist))+1)*fs_Ding) pause (30000/fs_Ding) hold on
plot(Ri_SampleNo(i),Dist(i),'ro')
AC C
EP
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 4:sonification of Drilling data % By Ali Kadkhodaie March 2017 % Load Christmass bell ring data matrix Load Ring.mat plot(depth,ROP) ylabel('ROP (ft/hrs)'); xlabel('Depth'); % Sonify ROP data for i=1:size(ROP,1) pause (1) sound(ROP(i)*Ding,0.6/(0.1*ROP(i))*fs_Ding) hold on plot(depth(i),ROP(i),'ro') 19
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
end
20
ACCEPTED MANUSCRIPT Conversion of well logs to musical notes
-
Boring tabulated log data converted into joyous and pleasant music
-
Aural quality control and interpretation of well logs
-
Sonification of capillary pressure and drilling data
-
Ability to store the musically converted well logs into MP3 files
-
Visually disabled reservoir engineers can continue their work
-
Freely available Matlab codes for future researches
AC C
EP
TE D
M AN U
SC
RI PT
-
S0920-4105(17)30514-4
DOI:
10.1016/j.petrol.2017.06.014
Reference:
PETROL 4026
To appear in:
Journal of Petroleum Science and Engineering
Received Date: 5 April 2017 Revised Date:
30 May 2017
Accepted Date: 5 June 2017
Please cite this article as: Kadkhodaie, A., Rezaee, R., Have you ever heard the sound of well logs or reservoir data?, Journal of Petroleum Science and Engineering (2017), doi: 10.1016/ j.petrol.2017.06.014. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Have you ever heard the sound of well logs or reservoir data? Ali Kadkhodaie1,2* and Reza Rezaee1 1. Department of Petroleum Engineering, Curtin University, Perth, Western Australia
RI PT
2. Earth Science Department, Faculty of Natural Science, University of Tabriz, Iran
Abstract
SC
The current study presents an effective approach to convert well logs to music and listen to their sound. Well log data play an important role in different stages of oil and gas field’s exploration and
M AN U
development. Several rock properties such as porosity, lithology, fluids saturation, fluid contacts and pay zones can be obtained through interpretation of well log data. Boring tabulated data or a mass of curves can be converted into joyous and pleasant sounds. For this purpose, four case studies were run to show how borehole quality logs, petrophysical evaluation results, capillary pressure data (pore size distribution) and drilling data can be converted to musical notes. The proposed approach can
TE D
help in quality control or interpretation of well logs or any other reservoir data. The interpreter just needs to wear wireless headphones and listen to the music generated from reservoir data. Service companies may consider the musical interpretation of well logs as an alternative and immediate way
EP
for quality control of logging procedures at well bore site. Meanwhile, visually disabled petroleum engineers may use the aural interpretation of subsurface data. A siren can be sounded when lost
AC C
circulation occurs or a kick is detected to warn well site crew about the drilling risks. The result of musical transformed well logs can be stored in MP3 files for future applications. Keywords: Well logs, sonification, audification, music, quality control, petrophysical interpretation
1. Introduction *
Corresponding author: Tel/Fax: +618 9266 9366 Email addresses: [email protected]; [email protected] (Ali Kadkhodaie), [email protected] (Reza Rezaee)
1
ACCEPTED MANUSCRIPT Subsurface data plays an important role is rock properties estimation and reservoir characterization. They provide very useful information for uncored intervals and making correlations in inter-well spaces. Reservoir rock properties are estimated from interpretation well log data (e.g. Rezaee et al., 2007 & 2008; Ostadhassan et al., 2012 & 2015; Sfidari et al., 2012 & 2014; Zamiran et al., 2015; Kadkhodaie et al., 2016 & 2017 ). Reservoir static models incorporate different rock properties in the
RI PT
form of well logs and propagate them through a simulation grid. Petrophysicists often deal with a mass of curves and boring tabulated well-logging data for each drilled well in a hydrocarbon field. Accordingly, finding alternative ways for quality control and interpretation of petrophysical data
SC
would be worth. Recently, sonification or audification of numerical data has received a considerable attention in different disciplines of science and engineering. Such approaches can help geologists and
M AN U
petrophysicists make aural interpretation of well logs. There are different approaches and methods for data audification or sonification. The first work published on the geoscience’s data sonification is “Listening to the Earth sing” by Hayward (1994). He proposed the idea of audio seismograms and made audio facilities accessible to geophysicists for sonification of seismic attributes. One of the main advantages of Hayward’s approach was the quick-look analysis of audio seismic data besides
TE D
their visual display. Quinn and Meeker (2001) introduced the climate symphony and other audifications such as the sound of seismic data and solar winds. Ekdale and Tripp (2005) introduced sonification of paleontological data and discussed how musical interpretations can help geologists to
EP
interpret subsurface data. Stewart and Brough (2006) studied the musical interpretation of well logs.
AC C
They discussed how audible expression of well-logging data can aid in the immediate identification of lithology changes. Quintero (2013) worked on the turning numerical values of wireline logs into different musical notes or sounds as an alternative way for wireline logs interpretation or making blind scientists aware of the downhole rock and fluids properties. Recently, McGee and Rogers (2016) worked on sonification of seismic data through making a variety of seismically based timbres. They stated that seismic data are the ultimate source of earth’s sound generation because they are in the form of physical waves. They can generate different songs directly using time-stretching and pitch-shifting without mapping to already available sounds. The current study presents an approach for mapping subsurface data to a set of predefined musical notes 2
ACCEPTED MANUSCRIPT by using the Matlab software (MATLAB 2016). The outcome music is helpful in quality control and aural interpretation of well logs. What makes the current study different from the previous ones is that in addition to well logging data, it focuses on the sonification of rock property logs, drilling sensor data and reservoir data such as capillary pressure curves. The musification codes generated in this study are freely available to future researchers working on exploration, description and
RI PT
development of oil and gas fields.
2. Methodology
SC
Audification is a technique for representation of a series of data as a sound. The input of Audification can be a time series or a set of data versus time or depth. Sonification or Audification
M AN U
can help researchers in the different discipline of science to create different songs and music pieces from a sequence of data (e.g., Dombois et al.,2001; Helmen and Ritter, 2004; Olivan et al., 2004; and Pauletto et al., 2005). A music piece or composition comprises three main component including melody, harmony and rhythm. When notes of different pitches are played, the outcome is called a
TE D
melody. Sometimes, melodies can be singable, however only the sequence of pitches does not create a melody. It is often possible to sing melodies that are distinguishable. By definition “harmony is usually referred to as simultaneously occurring frequencies, pitches (tones, notes), or chords (Malm,
EP
1996)”. Harmony is often defined as the vertical dimension of music, as opposed to the melodic line, or the "horizontal" dimension (Jamini, 2005). Each note has a certain duration and the relationship
AC C
between the durations of notes is referred to as rhythm. Seven basic musical notes are “Do, Re, Mi, Fa, Sol, La, Si”. In English and German-speaking countries, they are represented by 'C, D, E, F, G, A, B' letters (Fig. 1). Each basic note has a different pitch for which frequency increases from Do to Si syllable. The eighth note will have the same name as the first one and so on. The span between the first and eighths note is called one octave. Going up one octave in pitch will double the frequency. The combination of three or more musical notes creates a Chord. Accordingly, different songs and melodies can be created through a combination of the aforementioned musical notes. For example, the music sheet of the Old French Song (Tchaikovsky, 2005) is shown in Fig. 2. Another example 3
ACCEPTED MANUSCRIPT illustrating how the bell ring music sheet was mapped to gamma ray log signatures is shown in Fig. 3. The main purpose of the current research is to create a music piece through mapping the basic musical notes to the numerical value of well logs or reservoir data which is interpretable by experts in the petroleum industry. For this purpose, four case studies were run to create songs from borehole
RI PT
quality logs (gamma ray and caliper), petrophysical evaluation results (porosity and water saturation), capillary pressure data (pore size distribution) and drilling data (rate of penetration). The data chosen for this study come from one of the wells drilled in the Crismon Lake field which is
SC
located in the Fitzory Trough, Canning Basin, Western Australia. The Laurel Formation with a complex clastic-carbonate lithological sequence is the main studied interval in Crimson Lake field.
M AN U
The main lithological composition consists of sandstone with interbedded siltstone and claystone and thin beds of carbonate rocks. The sandstone is very 1ight grey to white, 1ight to medium brown and occasiona1 1ight green grey, dominantly with clear, transparent quartz, trace colored quartz and lithic fragments. Accessory grains include trace lithic fragments, glauconite and carbonaceous flecks
TE D
with trace to conman pyrite as loose aggregates and as cement.
3. Results and discussion
Following function of Matlab program (MATLAB 2016) was used to convert a matrix of well logs
EP
and other reservoir data to sound:
(Eq. 1)
AC C
sound(α*data, β*Fs)
where data is the matrix of a predefined song data such as Christmas bell ring, guitar, trumpet, piano or any other sound. It can be an n by 1 column vector for mono or single channel audio files or an n by 2 matrix data in the case of stereo playback. The multiplier α in front of song data sets the master volume, while the multiplier β controls the frequency or pitch of the sound. The frequency parameter, Fs, is the sample rate of the audio file. It depends on the hardware restrictions and can be a value in the range of [1000 Hz, 384000 Hz] for Matlab software (MATLAB 2016). Generally, it can be said that a high pitch sound corresponds to a high-frequency sound wave and a low pitch sound corresponds to a low-frequency sound wave. If the frequency of a sound is set 4
ACCEPTED MANUSCRIPT to Fs = 2*Fs1 the frequency will be cut in half and the pitch will go down as high as one octave. Similarly, setting Fs=0.5*Fs1 will double the frequency and the pitch will go up as high as one octave. Well log data were introduced in the form of α or β matrices to control the master volume and
RI PT
pitch of the chosen song. Depth was considered as the rhythm of the song and playing different pitches created well log melodies. The following examples show different song composed based on different datasets of a drilled well.
Borehole quality logs song: caliper logs in conjunction with gamma ray log are used to detect an
SC
enlarged region of a borehole and bad hole intervals. Washouts and cavities develop due to several reasons such as swelling or weakening of clays as they come in contact with mud filtrate,
M AN U
unconsolidated formations and rock mechanical failure caused by in-situ stresses. The first example of this study focuses on the simultaneous sonification of gamma ray and caliper logs for a test well. The pitch of the drum sound with a sampling frequency of 11025 Hz was mapped to the caliper log. For the case of gamma ray log, the ring sound with a frequency of 44100 Hz was chosen and its pitch
TE D
was mapped to the fluctuation of GR values. As is seen in Fig. 4(a), high gamma-ray intervals can be recognized through changes in the tonality and pitch of the bell ring and piano song. As with the gamma ray log, washout intervals can be identified through listening to the caliper log song (Fig. 4b).
EP
Interpreted well logs song: porosity, shale volume, water saturation and mineral fractions are the
AC C
main outputs calculated from the interpretation of well logs. The second case focuses on simultaneous sonification of porosity and water saturation logs as an example from interpreted petrophysical data (Fig. 5). Porosity log was mapped to the Santoor ding with the base frequency of 44100 Hz. Similarly, water saturation log was correlated to drum sound with the base frequency of 11025 HZ. The codes were modified so that if water saturation is above 50%, then the drum beats will be played. Listening to the joint composition of the porosity and water saturation songs will tell us information regarding good reservoir quality zones for which PHIE is high and Sw falls below 50%. Even, the changes in lithology, rock texture, pore system or other rock and fluid properties can be tracked by listening to the musical interpretation of well logs. Computer Processed Interpretation 5
ACCEPTED MANUSCRIPT (CPI) plot of the studied interval is shown in Fig. 6. The plots shown on left tracks are gamma ray (GR), neutron porosity (NPHI), sonic (DT), density (RHOB) and photo-electic (PEF) logs along with resistivity curves (MSFL, LLS and LLD) of the main reservoir interval (Laurel formation). The petrophysical interpretation of the reservoir rocks including lithology, water saturation and porosity are shown on the right tacks of the plot as well.
RI PT
Capillary pressure curves song: Capillary pressure curves are used for calibration of log-derived water saturation above the transition zone, determination of height above free water level and pore size distribution (PSD) analysis. The third example deals with sonification of pore size distribution
SC
curve determined from capillary pressure data. For this purpose, PSD data were mapped to the Santoor dings and the resulting song will tell us about the pore size, storage capacity and actually
M AN U
reservoir quality of the studied sample. The base frequency of the composed Santoor ding is 44100 Hz. Graphical representation of the pore size distribution data mapped to the Santoor song is shown in Fig. 7.
Drilling data song: there are different types of measurement while drilling data (MWD) used to derive useful information from subsurface rocks and mechanical strength of the formations being
TE D
drilled. Drill sensor data such as rate of penetration (ROP), bit pressure, rotation pressure, head speed, pull down rate and pressure recorded from drill rigs can be sonified by using the approach presented in this study. As an example, the rate of penetration of drill bit for the studied well was
EP
mapped to Christmas bell ring. The base frequency of the composed Christmas bell ring is 44100 Hz.
AC C
As is seen in Fig. 8. the pitch of the composed song changes when the penetration rate increases or decreases and somehow it is indicative of the rock type or changes in their strength. Similarly, a siren can be sounded when lost circulation occurs or a kick is detected to warn well site engineers about the risk of blowout or loss of well. MP3 and Video files of the sonified data discussed above can be found in the following permanent link: https://goo.gl/XfWnki The same procedure explained for the shown examples can actually be applied to any reservoir data to make new songs. Even a 3D grid of reservoir static and dynamic properties can be sonified and transformed into the musical models of rock properties. They can sing pleased songs when mapped 6
ACCEPTED MANUSCRIPT and adjusted to the joyous melodies. Matlab codes (MATLAB 2016) used for the sonification of the case study data are shown in Appendix 1.
4. Conclusions The results of this study showed that sonification of well logs can be considered as an alternative
RI PT
approach in rock properties analysis. Two cases of well logs quality control and interpretation were tested and satisfactory results were achieved. Fluctuations of well logs responses were mapped to the musical notes and their properties such as pitch, volume and tonality. The created music and songs
SC
not only provide joyous and pleasant moments but also make it possible to have a quick interpretation of well log signatures. Through a change in tonality or pitch of the created music, the
M AN U
interpreter can track changes in rock properties or quality of well logs. The proposed approach can be applied to any other reservoir information such as drilling data, production data, flow rate and history match curves. The output musical data can be stored in audio files with popular formats such as MP3 or WAV. It is expected that in near future, clients ask their service companies or contractors to provide an audio file of the well logs in addition to their hard copy or digital files. Matlab codes
References
EP
TE D
for sonification of well logs, reservoir and drilling data are listed in Appendix 1.
AC C
Dombois, F., 2001.Using audification in planetary seismology. Proceedings of the 7th International Conference on Auditory Display (ICAD2001): 227–230 Ekdale, A., and Tripp, A., 2005. Paleontological sonification: Letting music bring fossils to your ears. Journal of Geoscience Education, 53, 271-280. Hayward, C., 1994. Listening to the earth sing. in Auditory Display: Sonification, Audification, and Auditory Interfaces. Addison-Wesley, 1994. Hermann, T., Ritter, H., 2004. Sound and meaning in auditory data display" (PDF), Proceedings of the IEEE, IEEE, 92 (4), 730–741. doi:10.1109/jproc.2004.825904 7
ACCEPTED MANUSCRIPT Jamini, D., 2005. Harmony and Composition: Basics to Intermediate, p.147. ISBN 1-4120-3333-0. Kadkhodaie, A., Rezaee, R., 2016. A new correlation for water saturation calculation in gas shale reservoirs based on compensation of kerogen-clay conductivity. Journal of Petroleum Science and Engineering. Vol. 146, 932-939
RI PT
Kadkhodaie, A., Rezaee, R. 2017. Estimation of vitrinite reflectance from well log data. Journal of Petroleum Science and Engineering 148, 94-102.
Malm, W.P.,1996. Music Cultures of the Pacific, the Near East, and Asia, p.15. ISBN 0-13-182387-
SC
6. Third edition.
MATLAB, 2016 version. The MathWorks, Inc., Natick, Massachusetts, United States.
M AN U
McGee, R., Rogers, D., 2016. Musification of seismic data. The 22nd International Conference on Auditory Display (ICAD–2016). Canberra, Australia.
Olivan, J., Kemp, B., Roessen, M. 2004. Easy listening to sleep recordings: tools and examples. Sleep Medicine, 5 (6): 601–603, doi:10.1016/j.sleep.2004.07.010
TE D
Ostadhassan, M., Zeng, Z., Zamiran, S., 2012. Geomechanical modeling of an anisotropic formationBakken case study. 46th US Rock Mechanics / Geomechanics Symposium 2012, At Chicago, Illinois, USA.
EP
Ostadhassan,M., Zamiran, S., Jabbari, H., 2015. Stability Analysis of Multilateral High Density Pad
AC C
Wells in the Three Forks Formation. SPE Western Regional Meeting, At Garden Grove, California, USA.
Pauletto, S., Hunt, A., 2005. A comparison of audio & visual analysis of complex time-series data sets. Proceedings of the 11th International Conference on Auditory Display (ICAD2005): 175–181 Quinn, M., and Meeker, L., 2001. Research set to music: The climate symphony and other sonifications of Ice Core, Rada, DNA, Seismic and Solar wind data: Proceedings of the 2001 International Conference on Auditory Display, Espoo, Finland. 56-61.
8
ACCEPTED MANUSCRIPT Quintero, G., 2013. Sonification of oil and gas wireline well logs. International Conference on Auditory Display. Pp. 301-306, 6-10 July, Poland. Rezaee, M.R, Kadkhodaie, A., Barabadi, A., 2007. Prediction of shear wave velocity from petrophysical data using intelligent systems, a sandstone reservoir of Carnarvon Basin. Journal of
RI PT
Petroleum Science and Engineering 55, 201-212. Rezaee, M.R., Kadkhodaie, A., 2008. Intelligent approach for the synthesis of petrophysical logs. Journal of Geophysics and Engineering 5, 12-26.
SC
Sfidari, E., Kadkhodaie, A, Najjari, S., 2012. Comparison of intelligent and statistical clustering approaches to predicting total organic carbon using intelligent systems. Journal of Petroleum Science
M AN U
and Engineering 86-87, 190-205
Sfidari, E., Kadkhodaie, A., Rahimpour-Bonab, H., Soltani, B., 2014. A hybrid approach for lithofacies characterization in the framework of sequence stratigraphy: a case study from the South Pars gas field, the Persian Gulf basin. Journal of Petroleum Science and Engineering 121, 87-102.
TE D
Stewart, R.R., Brough, S., 2006. Log jammin’: transforming well logs to music. CREWES Research Report Number: Vol 18. SEG Technical Program Expanded Abstracts 25(1), 1-18.
EP
Tchaikovsky, P.I., 2005. Old French Song music sheet. [email protected] Zamiran, S., Osouli, A., Ostadhassan, M., 2014. Geomechanical modeling of inclined wellbore in
AC C
anisotropic shale layers of bakken formation. 48th US Rock Mechanics / Geomechanics Symposium, At Minneapolis, Minnesota, USA
9
ACCEPTED MANUSCRIPT
RI PT
Fig. 1. Seven basic music syllables: “Do Re Me Fa Sol La Si”. Each basic note has a different pitch
AC C
EP
TE D
M AN U
SC
for which frequency increases from Do to Si syllable
10
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig. 2. Music sheet of the Old French Song (Tchaikovsky, 2005).
11
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3. Bell ring music sheet for piano (Copyright 2015 Red Balloon Technology Ltd.) mapped to
AC C
EP
TE D
gamma ray log.
12
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 4. High gamma ray intervals can be recognized through changes in the tonality and pitch of the bell ring in piano song (a). Washout intervals can be identified through mapping drum sound to
AC C
EP
TE D
borehole quality logs such as Caliper log.
13
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
Fig 5. Porosity and water saturation mapped to Santoor and Drum song, respectively. Listening to the composed song helps in identification of high porosity zones. If water saturation is over 50%, then drum beats will be played.
14
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6. Computer Processed Interpretation (CPI) plot of the studied interval. The main lithology composition is shale and sandstone with some contributions of carbonate rocks in the lower interval.
15
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 7. Graphical representation of pore size distribution data mapped to Santoor song
16
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
Fig. 8. Graphical representation ROP (Rate of penetration) sonified based on Christmas bell ring song. The composed song will aid in understanding rock types and strength of the rocks being drilled.
17
ACCEPTED MANUSCRIPT
M AN U
% Sonify GR and Caliper logs for i=1:size(CALI,1)
SC
% By Ali Kadkhodaie March 2017 %CASE 1:sonification of GR-CALI logs % Load Christmass bell ring data matrix Load Ring.mat % Load Christmass Drum data matrix Load Drum.mat %Plot input data subplot(2,1,1) plot(depth,CALI) ylabel('CALI'); subplot(2,1,2) plot(depth,GR) ylabel('GR'); xlabel('Depth');
RI PT
Appendix 1: Matlab codes for sonification of well logs, reservoir and drilling data
sound(Ring(10000:20000,:),.05*GR(i)*fs_Ring) if CALI(i)-8.5>9.5 sound(2*Drum(70000:72000,:),fs_Drum*0.5) else
TE D
end pause (10000/fs_Ring)
EP
hold on subplot(2,1,1) hold on plot(depth(i),CALI(i),'ro') hold on subplot(2,1,2) hold on plot(depth(i),GR(i),'ro')
AC C
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 2:sonification of Phie-Sw logs % Load Santoor ding data matrix load Santoor.mat %Plot input data subplot(2,1,1) plot(depth,phie,'r') ylabel('PHIE'); subplot(2,1,2) plot(depth,Sw) ylabel('Sw (v/v)'); xlabel('Depth (m)'); % Sonify porosity and water saturation logs for i=1:size(phie,1) sound(Ding*phie(i),.2*fs_Ding) 18
ACCEPTED MANUSCRIPT if Sw(i)>0.5 sound(Drum(70000:72000,:),fs_Drum*0.5) else end pause (10000/fs_Ding)
RI PT
hold on subplot(2,1,1) hold on plot(depth(i),phie(i),'ro') hold on subplot(2,1,2) hold on plot(depth(i),Sw(i),'go')
M AN U
SC
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 3:sonification of PSD data % Load Santoor ding data matrix Load Ring.mat %Plot input data plot(Ri_SampleNo,Dist) ylabel('Dist (Cm2)'); xlabel('Pore radius (micron)');
% Sonify Pore Size Distribution data for i=1:size (Dist,1)
TE D
sound(2*Ding,((Dist(i)-min(Dist))/(max(Dist)-min(Dist))+1)*fs_Ding) pause (30000/fs_Ding) hold on
plot(Ri_SampleNo(i),Dist(i),'ro')
AC C
EP
end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %CASE 4:sonification of Drilling data % By Ali Kadkhodaie March 2017 % Load Christmass bell ring data matrix Load Ring.mat plot(depth,ROP) ylabel('ROP (ft/hrs)'); xlabel('Depth'); % Sonify ROP data for i=1:size(ROP,1) pause (1) sound(ROP(i)*Ding,0.6/(0.1*ROP(i))*fs_Ding) hold on plot(depth(i),ROP(i),'ro') 19
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
end
20
ACCEPTED MANUSCRIPT Conversion of well logs to musical notes
-
Boring tabulated log data converted into joyous and pleasant music
-
Aural quality control and interpretation of well logs
-
Sonification of capillary pressure and drilling data
-
Ability to store the musically converted well logs into MP3 files
-
Visually disabled reservoir engineers can continue their work
-
Freely available Matlab codes for future researches
AC C
EP
TE D
M AN U
SC
RI PT
-