Forecastining 2

[ppata]

Hi,
I'm a reservoir engineer and I need documentation for forecasting risk analisys (p10, p50 & p90). If you have some ideea about that(a link or a book or anything), please help.
I don't think you know how is when "they" ask you something without any help and the single software you have is ms excel.
I know what are the possibilities of excel but I'm affraid I'm not very familiar with the "p10, p50 & p90" style of working. In the past I used most of all the deterministic calculations and I need some definitions for p10, p50 & p90.

 

[zdas04]

ppata,
You'd really be suprised how many of us do "know how it is when "they" ask ... " it is a very common thing. We're grown-ups being paid for our abilities. If you're asked to do something outside your experience, then learn how to do it.

I'm assuming (since the only tool you have is Excel) that you're doing decline curve reserves analysis. There are 3 significant variables in that analysis: (1) rate of decline; (2) kick off point for your prediction; and (3) abandonment rate. A P10 analysis would (from convention) have you put values in all three of those variables that could be expected to happen 10% of the time (i.e., optimistic estimates of decline rate, starting point, and abandonment rate). A P90 has pessimistic values in all 3 variables. A P50 has your most likely estimates.

In other words, you run a range of each variable that you control and assign a probability that that will be the result. The sum of the reserves times the probability will be your P50.

A tank model is done the same way except the variables that you are estimating are different.



David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[DrillerNic]

If you can, get Crystal Ball, an excel add in. Then having built your spread sheet, using crystal ball you put ranges on each of the inputs in your spreadsheet (decline rate, starting oil rate, abandonment rate, oil price, say or area, reservoir thickness, net pay, porosity, etc), then tell Crystal ball to do a Monte Carlo simulation with (at least) 1000 runs, press go and nip off to get a coffee. When you come back, you'll be able to pick out the the P10 P50 and P90 numbers by plotting all 1000 results on a cumulative curve.

This is much the same as what Dave is suggesting above, but as P10 P50 and P90 are probabilites, you have to do a Monte Carlo analysis with lots of runs, rather than using your estimates of low mid and high values in a deterministic type calculation.

 

[ppata]

Me again
I talk with my boss about the Crystal Ball and @Risk, he told me that I must expect to start working with "Rose". Did you have an ideea about this software? I need some documentation (on .pdf or .doc) to make my homework on this subject.
Thank you for your replies and I hope you are "there".

A.