Tuesday, December 13, 2016

Xplotter - added histograms

The name for Glacier Geosciences came from my wonderful wife.  She was taking my love of outdoors (mountain biking and hiking) and mixing it with business.  The more I thought about it I liked the name.  I, of course, know what glaciers are but I looked up the word.  The definition was something along the lines of "a constant moving force".  I liked that.  It may not always appear that I am doing something but the results are very evident.  

So in this update I'm sharing some eye candy of my first attempt at putting histograms into my cross plot software.  <start shameless brag>  I only worked on it for one two nights.  <end shameless brag>  I'm sure I will change, modify and make it prettier.  There is no great scientific revelations here.  I'm just slowly improving my toolbox.  All of my software has been written in python and uses no extra plugins.  I'm using straight-forward Python 3.5 and Tkinter.  


This first example is just a simple plot of density and gamma ray.  The importance of histrograms are pretty self evident.  You can quickly get the P10, P90 and average from each bi-modal distribution by eye-balling the x-axis for the density curve without having to filter and perform the math.  


The example above is where I have flipped the display axis and made it plot the histrograms on the right and top of the plot window.  The data points are colored by VCLAY increasing from light to dark using my coffee palette.  

Options I've coded in:
  • Display on Top or Bottom, Left or Right
  • Change bin distribution line color and thickness
  • Change color fill for the bins
  • Change number of bins to display



More eye candy with different colors.  



This last example is of three different wells that were filtered to the same interval.  I also pull in tops from my log viewing software automatically.  It is nice to get stats on multiple wells all at once.


It really starts to look good when I add in a chart.  Ooooohhh! Aaaaaahhhh!  On the color filel that time I used Hot Pink (#FF69B4).  I'm actually a fan.  At a company I used to work at they were against "girlie" colors.  I think it was more of a joke than a rule but still funny.  The data points are again colored by VCLAY going from light to dark.  

But in the UMAA / RHOMA plot aboev you can see the need for histograms.  There is a distribution on the x axis that favors values around 10 and 12.  If you were to just look at the data points you would think that 8 had an equal representation.  



I did a quick linear transposition for logarithmic curves.  I will eventually expand it where the bins are within the logarithmic scales.  To complete the Pickett Plot I just need to put the slopes for M and N and have them interactive.  



Last one for eye candy sake and bi-modal distributions are cool.  That color fill is cornflower blue (#6495ED) while the data points are lime green (#32CD32).

Next up:  Getting interactive!  I want to be able to filter to data graphically by clicking around the plot and get reports; depth, well name, etc.  I really want to put in the functionality of the Pickett plot because I am growing tired of drawing the lines manually and putting points into Excel for my M for water saturation.

Thanks for looking!  It has been fun!

Regards,
Jon



Sunday, September 11, 2016

Xplotter - chart updates and some clay typing

This post is a continuation of the other week's mention of Xplotter - Cross Plots and Color Palettes.  While consulting and building a petrophysics business I am making my own tools instead of spending 60k+ on existing solutions.  Adding the ability to load charts onto plots went a lot quicker than I anticipated.  Last time we left off I was simply drawing the charts on the plot using MS Paint.  Now I've added a feature to save chart data in XML and load directly into Xplotter.

Below is an example of the RHOMA vs. UMAA chart.  When you correct density for clay, organics and porosity the values will fall within the ternary diagram so that you can determine the volume of calcite, dolomite and quartz.  In this example below I have just corrected for organics so it is along the way to mineralogy determination and still needs correction for clay and porosity.

Hand Drawn Ternary Diagram - last week


Program Generated Ternary - this week


I am now going through the different chart books and digitizing some of the more popular charts.  I'm going to avoid things like tornado charts and their million lines as much as possible.  I can only torture myself so much.

Below is an example of Gamma Ray - Potassium vs. Photoelectric Factor.  This chart is used for clay typing.  The example is from the Wolfcamp in south Midland Basin.  The expected clay type is illite and very little smectite / montmorillonite.  It is nice when the charts work out; makes my job easier.  All I did outside of Xplotter was add the colorful text boxes.  The colored outline boxes to highlight clay types were all drawn from XML templates.  

Gamma Ray - Potassium vs. Photoelectric Factor


Sure, I may have calibrated the boxes on the Gamma Ray axis, but that is because GR can be a highly qualitative measurement.  

There is another example you can use to do clay typing using spectral gamma ray; potassium vs. thorium.  To be honest, I've not used it very much because I've typically had XRD data to rely on.  But I am definitely going to use it for a basis of comparison for different resource shale plays.  I wonder how it will look when I compare SCOOP Woodford vs. Permian Wolfcamp vs. Ft. Worth Barnett.

Potassium vs. Thorium




XML template snapshot



And no example is complete without cross plotting neutron and density.  Here I have the lines from two charts on the same plot.  This allows me to compare the differences between charts and even service companies.  Below it is interesting to see how much a slight change in flushed zone water can have an effect on the density and neutron measurements for estimating porosity.  The slightly darker colors are from the more saline mud fluids.  

Neutron vs. Bulk Density


Thanks for reading!

Cheers,
Jon

Saturday, August 27, 2016

Xplotter - New Updates; multi-well cross plots and some nice color palettes

While starting a petrophysics consulting business I am making software on the side to support the consulting.  This is to combat the high cost of the major petrophysical software programs available and also have something professional and customizable.  Xplotter is a cross plot program that will read a CSV or LAS file directly and even allows adding in multiple LAS files to compare measurements in the same intervals.  This software is not yet ready for commercialization, but I am thrilled with how it is coming along.  Here are some of the recent updates and fun examples.


1)  Multi-well cross plotting & limit data to interval of interest

Some of the existing geologic software packages that include petrophysical packages still can't perform multi-well cross plotting.  This is vital when making a petrophysical model across a basin.  In this example I've pulled in three wells and have limited their data to the tops I picked for the Wolfcamp A.



2)  Fun with color palettes

I never meant to be an interior designer.  My clothes are very plain and I can't see the differences in shades of yellow paint.  But my color palettes for logging measurements is getting pretty complex and awesome.  Here are some fun examples.

Gamma Ray Uranium vs. Total Gamma Ray - colored by volume of calcite


Neutron vs. Bulk Density - colored by volume of calcite


Bulk Density vs. Neutron - colored by clay volume

UMAA vs. RHOMA - colored by clay volume


I call that last color palette "Coffee is good for Scientists".  Of course, I added some lines, text and title boxes outside of Xplotter.  I've worked out the code to be able to add comments to plots, move them around and just need to implement.

3)  Log/log & semilog

A while back I put in logarithmic capabilities.  I think that instead of writing a completely separate program for decline analysis, I will just include the calculations into Xplotter for decline and EUR calculations.  In the example below I am using public data from the Texas Railroad commision and looking at when different wells came online in the same section.I added the yellow and green boxes and text outside of Xplotter.



Thanks for reading!

Regards,
Jon





Saturday, August 20, 2016

OCGS Talk - August 31st - Permian Basin Value Model



OKLAHOMA CITY GEOLOGICAL
SOCIETY
TECHNICAL CASE STUDY

August 31, 2016

SPEAKER: Jon Reynolds,
Petrophysicist
Glacier Geosciences, LLC

“Permian Basin Reservoir Characterization:
Generating a Value Model with
Petrophysics, Geosteering and Production”

OCGS DEVON GEOSCIENCE CENTER
10 NW 6th Street
OKLAHOMA CITY, OK 73102

WEDNESDAY August 31, 2016
Social Hour 4:30 pm – 5:30pm
Presentation 5:30 pm – 6:30 pm

Reservations are appreciated by Monday August 29th, 2016

Abstract:
The oil and gas industry is optimistic about 2017. In order to fully prepare for potential acquisitions a good basis of comparison is nice to have in your back pocket. The Permian Basin has a wealth of public data for Midland and Delaware Basin to build a value model to compare against potential deals within the basins and for comparison against other basins. The data also covers an important time period in the development of shale oil that can be used as a teaching tool. In this talk we will answer technical questions such as:

“How much does lithology matter when you drill your horizontal in the Permian Basin?”
“Is there enough data in the Permian to build a multi-mineral petrophysical model without core data?”
“What data is available for free?”
“What would I need to purchase to infill and map the whole basin?”

For example, in the University of Texas Lands database:
     1) 29 gigs of LAS Files
     2) 10,000+ LAS (1,900 with triple combo)
     3) 4,200+ wells
     4) Not including the thousands of rasters

This talk will cover:
     1) Using the available public data to build a multi-mineral petrophysics model without core data for porosity, saturation and identify flow units
     2) Determine which intervals of the different formations were targeted using directional surveys and gamma ray logs in order to group like wells for production analysis
     3) Analyze production, petrophysics, geomechanics and stratigraphy to
optimize target zone selection

Bio: Jon Reynolds

During the last ten years Jon has worked for American Energy Partners,
Chesapeake Energy, Fronterra Geosciences and Baker Hughes. Currently he is an
independent Petrophysicist and formed his own company, Glacier Geosciences.
His background includes working unconventional oil and gas shale plays across
the United States, Argentina, Mexico and Australia. These include but are not
limited to: Wolfcamp, Bone Springs, Spraberry, Point Pleasant, Marcellus,
Haynesville, Woodford, Vaca Muerta, Bakken, Eagle Ford, Buda, Velkerri, Kyalla,
Barney Creek, Lawn Hill, and Pimienta. This includes full scale basin wide
reservoir characterization of mineralogy, porosity, organic maturation, pressure,
volumetrics and reserves. He has developed techniques to handle large data sets
of tens of thousands of wells at a time to the small scale, high resolution
stratigraphic modeling.

++++++++++++++++++++++++++++++++++++++++++++++++++++++

Drinks will be provided prior to the presentation.
Reservation is not required for the open house and presentation. An accurate
headcount will ensure a more enjoyable social hour. Thank you for your
consideration.
Please RSVP to Jennie.Reynolds@GlacierGeosciences.com
Name: ________________________________ Company: ____________________________

Wednesday, July 20, 2016

Xplotter - beta peak

Cross plots are an essential tool of the subsurface scientist.  We've all met our salty engineer that has spreadsheets jam packed with macros and scatter plots.  Your options are Excel, SpotFire, Orange or Tableau to name a few.  When using flat files; Excel has to be formatted and SpotFire requires a huge amount of meta data so you better not have a lot of columns in your excel sheet.  There isn't anything out there that just reads an LAS and makes a cross plot.  Until now.  And here it is below.


Not super impressive.  I bet you were expecting more.  But those images are further down in the article.  But please don't scroll down and cheat just yet.  You have to allow the author ramble and get on with their ideas to fully appreciate how we get there.


With any cross plot you want to select your curves for the difference axis.  I've made some pretty non-descriptive buttons you can click on to make your initial curve selection.


Select a few curve and you have your cross plot.  This one is super basic and boring because we are using my test well.  But once we right click on those buttons we can shift around scales and change the colors.  Crimson is available for my Boomer buddies and Orange is ready to select for my Cow-people.  




I've even made the option of reading in CSV files, saving and opening templates for efficiency.  


If we take a look at some actual data the plots start to look a bit more interesting.


And if log scale or semi-log scale is your fancy, we have that too for either axis.



But of course we want to filter to the interval of interest.  So to do that I made a handy right click menu that allows you to plug in depths.  


I bet you want me to hit that "Insert Bet Fit Line" option too, eh?  Sure.  Bonus pic!  Two features in one snapshot!  Depth limited and least squares regression best fit line!  And yes the best fit line is based on that depth interval.


And if you were to middle click (who uses middle click?  Must be an old Recall user) you will get the stats of the curves and best fit line.  


Can you copy those statistics?  Yes!  Can I help you sell your house?  Probably not, but I know a great sales lady who knows everybody.  

I love the purple but there is one last feature I've added recently that is pretty nifty.  There is an option to color your values based on a Z attribute value.  For this last illusion, I give you a neutron density plot colored by GR (0-50, 50-100, and 100-400).


And by the sweet gift of XML these templates are quickly saved for later.


Have a great week everybody!

Thanks to Guido van Rossum and my friend caffeine.

Saturday, June 25, 2016

Log Viewer Update - Graphical Curve Editing

Half or more than half of the time you take care of cleaning up data.  It is important that your tools work quickly and are easy to use.  Two days ago I added the ability to create, modify and delete formation tops.  Today this update to the Log Viewer allows graphical editing of any of the curves.  Below is a quick demonstration of how easy it is in my Log Viewer to correct measurements.

Let's revisit my Area 51 Roswell, New Mexico log.  In this example I'm going to draw a little bit on the Gamma Ray to make it look more real instead of just the interpolated numbers I generated.


 I've creatively named my curve editing mode.  This menu options toggles Curve Editing on and off.


Now, I just click where I would like to correct the log.  Once I have a couple dots I can right click and the viewer will interpolate in between my chosen points.  If I click somewhere I don't like, I just CTRL+Z to undo the last or keep hitting undo to remove as many as I want.  This is a feature you don't often get with software packages.  


Once happy with my edits I right click and I get my newly edited curve.  This is starting to look more geologic or like an evil character's nose from a Disney film.


Finally I can save my work by just outputting a new LAS. 


Cheers,
Jon

#LogViewer #Python #CodingWhile2YearOldNaps

Thursday, June 23, 2016

Log Viewer Update - Tops Picking

Graphically creating, moving and deleting tops in a log viewer is a must.  Typing in depths is not acceptable.  Making it easy to use is even better.  For those Recall users out there, you know that there are more buttons than just the left mouse button.  Most time in software you left click on EVERYTHING.  This can create a LOT of extra clicks.  "Click this reticle icon to edit tops.  Click here to create a top.  Hover your mouse over the top, raise your right foot and change to a different program to delete tops."

I've made mine pretty simple.  Select the menu to edit tops.  Left click and drag to move existing tops.  Right click to create a new one.  Middle click to delete.  All popup windows appear underneath the mouse click.  A temporary gray line appears while you are moving an existing tops.

I've got my made-up Roswell, New Mexico log here for an example.  (Yeah, I looked up API number for the Chaves County)



So to do anything you just have to turn on Tops Editing.



I creatively just added <Tops Editing Mode> to the title to know whether or not the mode was enabled.  



Right click and type in the new top name.  That's it.  



To move the top requires a simple left click and drag.  While dragging there is a temporary gray line to help that the move is in the right spot.  


I can create.  I can edit.  Now to delete I could just middle click to remove a top.


All of this is done in memory.  So saving edits is done simply from the menu bar.  


This motivates me to build a cross section view so I can correlate across multiple wells.

Wednesday, June 8, 2016

Permian Basin - Defining Study Areas


Above is a map of well locations from UT Lands and some geologic features from the Bureau of Economic Geology.  Quickly looking at this three different area pop out immediately.

Area 1


In Reagan county there is a lot of activity with the exploitation of the Wolfcamp.  As the Ozona arch and Big Lake fault can provide some interesting geologic features, the better part of the Wolfcamp reservoir quality lies in the northeast section of data points.  Thankfully there are a lot of well control points to choose from for vertical and horizontal wells.

Area 2


Up north in Midland Basin is the heart of the Spraberry trend.  Wells along the Martin and Andrews county line will provide a wonderful analog in this area.

Area 3


Delaware Basin is the sleepy giant that has gotten a ton of press in the last two years with large IP rates from short laterals.  The basin has slightly higher pressures than Midland but the acreage blocks are more of a shotgun blast / military camouflage type pattern.  In picking well control points you need to stay away from that eastern basin outline.  Also, the faulting to the south presents some challenges as well.

All maps were put together using free software and freely available public data.  QGIS was used as a mapping software.  And UT Lands and BEG were used as data sources.  

#QGIS  #DataMining  #PermianBasin