Petrophysics Rocks

I've just returned from a wonderful vacation in New Zealand - what a fantastic place!

Whilst I was away a Subscriber asked me a couple of questions about my forthcoming GDPCmaker Suite Spreadsheets. and so I thought I'd reproduce them, and my responses here:

I WAS PARTICULARLY INTRIGUED BY YOUR "AN EXCEL EARTHMODEL." I AM LOOKING FORWARD TO SEEING A FULL COURSE ON YOUTUBE ABOUT HOW TO WORK WITH YOUR EXCEL SPREADSHEET. I HAVE A SMALL REQUEST REGARDING THIS COURSE. I WOULD BE VERY GRATEFUL IF YOU COULD EXPLAIN IN MORE DETAIL:

• WHAT SPECIFIC TASKS IS THIS SPREADSHEET DESIGNED FOR?


- Originally, I thought of this project as providing a public-domain teaching reservoir data set that I could easily use for a Petrophysics Training course. By creating an “Earth Model” (where we know the answer – because we have defined/fixed the Geology) – we can create several realisations (log sets) representing what “random logs” would look like in the same reservoir.


- In the real world we always have an issue with how many logs(wells) do we need to properly characterise/average a geologically variable reservoir. So, my thought was that, as a class, we could define a reservoir that was relevant to the participants, like fields they were currently working on, e.g. a deltaic/channel fining upward sequences system – variable Net/Gross – Channel/inter-channel areas? Each participant would get their own log data set to evaluate – and we would see how far-off the Field Average each Well was. Then we could combine the whole class Well dataset results, to investigate how many wells were required to match the original Earth Model “Truth” and evaluate….geological variance, tool measurement error and petrophysical model uncertainty.


- Also, sometimes when you look at real-world logs and observe strange tool responses (are they realistic?) – you just want to play with various bed/borehole geometry & variable bed-thickness – and see if you can match the actual problem log response….so it allows you to experiment with realistic scenarios.


- In the "olden days" we used to have to specify “dummy wells/values” to constrain our reservoir models/contour maps………e.g. force a channel boundary? NTG trend…these days “GEO modellers” do this directly in the property modelling – but I can’t help feeling there may still be a place for creating a set of pseudo control wells (high – side set? /low-side set? – to switch on and off, between sensitivities,….for a bit more repeatable and deterministic property control? – Perhaps it would be easier for the rest-of the geoscience team to share, communicate and understand what the black-box 3D model is doing….is it realistic?


• WHAT DATA IS INPUT INTO IT AND WHERE TO GET IT FROM (SOURCES)?


- You decide the reservoir geometries and properties. At the moment – I’ve set it up for you to define:

o Gross Reservoir thickness (e.g. 50-500 ft),
o how many Sand Beds you want (e.g. 5-50),
o Sand Facies Proportions (e.g. 70% Good Sands, 30% Bad Sands),
o Average Bed thickness by Facies (e.g. Good Sands c.12ft, Bad Sands 2ft),
o Average Clay/Shale trends by Facies (e.g. constant/fining up/coarsening up)
o Average PHIT/PHIE by Facies
o Average Poro-Perm by Facies
o Average SW from J-Functions (by Facies)

- So, you can modify any of these properties to design an Earth Model around your own example reservoirs and geometries (eventually I’ll release a metric version…..and perhaps a carbonate reservoir in the future - don't hold me to it!).


• WHAT SHOULD ONE PAY ATTENTION TO FIRST WHEN MODELLING?


- The idea is that you input this EARTHmaker Model into a LOGmaker Model and create your pseudo-wireline (and/or eventually LWD) logs…then you can petrophysically evaluate those in the CPImaker Model to see if you can get back to the EARTHmaker “Truth” – then ultimately you can put the CPImaker model into the FLOWmaker model and see what the expected fluid flow rates might be.

- The EARTHmaker will generally be at higher resolution than the LOGmaker results (which are affected by tool detector spacing and invasion profiles etc..) and the so the Reservoir Sums and Averages will change – so keep an eye on them…….expect the averages to stay consistent, or at least change consistently….if you lose Net Thickness – due to gamma ray resolution for example - you’ll probably be losing poorer log quality from the averages – so expect the average properties to improve?

– So understandable changes are good……..unexpected changes should be investigated…..part of the learning/understanding process!



• HOW TO CORRECTLY INTERPRET THE RESULTS OBTAINED?


- I guess the biggest issue one always encounters (chicken and egg) in knowing what to “pay attention” to? Do things look/feel right and how do you know?

- This is really the difference between Log Analysis (you DO STUFF/apply equations TO data and get WHAT result out?) and Petrophysics (where you UNDERSTAND the reservoir quality controlling factors and WHY you get the results you do?)

- Clearly with experience you can judge the coherence of the models you create – Do they look like the logs you’d expect to get? I hope you can use this tool to help gain/improve that experience, experiment with reservoir properties – what different log signatures do you get? - increase the invasion profile – flush away more hydrocarbons – do the logs responses change accordingly? Once you are in control of the rocks & fluids - you can experiment with the log response!

Skip the Carousel

1 2

6 months ago | [YT] | 8

Petrophysics Rocks

I recently had the wonderful opportunity to sit down and chat with Ohenewaa Amoako-Dadey for the Aberdeen Formation Evaluation Society's "Logs Have Voices Podcast" - We discussed careers in Petrophysics and, in particular, what new geoscientists and engineers could expect from a life in and around the squiggles! (see link below:)

open.spotify.com/episode/4dnFdMYQQuDPPNwPJIRRJJ?si…

Skip the Carousel

1 2

1 year ago (edited) | [YT] | 3