Well tests in wells with several layers

[vinomarky]

Welcome to the real world of imperfect and uncertain data

No, there is no magical software package that will auto-deconvolve multi-layer tests into unique single layer results. You will have to incorporate more information, some engineering judgement and accept the the result will still be uncertain

Additional information to incorporate;
Well logs - as a minimum thickness's of the various sands. Ideally the E&P companies have core and/or mobility tests, which have been used to tune well log analyses to (gu)estimate perms of those sands. In this case, you would be best served setting up a simple model to reflect the layered architectural, and history match while controlling the range of perms depending upon how confident you are of them (perhaps +/- 50%?). Note, if you are using numerical simulation to do this, understand that Peaceman well connection factor is based upon steady state equations, meaning you may need to significantly fine-up your gridding around the wellbore to represent transient effects well enough in your relatively short term build up tests.

Other information to consider;
Field wide information - geologically, are your sand units pervasive? High perm? relatively homogeneous?
If yes to all, then you will be looking for similar perms AND P*'s in all tests of similar vintage in a given layer (these questions may be true for one layer, but not necessarily all - in which case, it is still helping your understanding). If not, perhaps look for wells close to each other
Also, if looking at high kh contrast wells, think through the likely ranges and understand that kh as a measure of deliverability and Phi.h as a measure of storativity is important. The relative ratio of kh/Phi.h will generally control the degree of depletion seen in each layer (assuming no vertical connection in the reservoir, or significant aquifer support). The early extrapolated P* will probably be closer to the highest kh layer, while late time P* will point closer to the lower kh layer(s)

Essentially, you have more degrees of freedom than data points to constrain with your single well test results, and you will need to look for ways to collapse these freedoms through application of engineering understanding. Remember, an engineer who can only solve the obvious problems (data in + equation = answer) is really just an overpaid technical assistant - use your understanding to extrapolate, group, and selectively simplify.

You may well STILL not have enough information to reduce the levels of uncertainty enough, if which case, dont be shy to recommend additional single sand tests - is the client doing some infill wells? Then another test in an upcoming drill may be easy. It comes down to what are the implications of the questions you are trying to answer (how important is that reduced uncertainty?) If valuable, then additional tests should be easy to justify. If not, then accept the uncertainty - characterize it, and show the implications of the bookend assumptions along with a 'most likely', and plan for flexibility.