InSies – All You Need For Inversion Software

We’re pleased to relay that InSeis – an excellent Kingdom add-on which allows two different seismic inversion techniques – is coming to Petrel Q1 2016. Coloured Inversion and Simulated Annealing Inversion are undeniably important techniques, which allow turning seismic signals into an acoustic impedance layer model with ease.

Simulated Annealing Inversion is aided with InSeis. Whereas Simulated Annealing Inversion is known as a very difficult procedure with it’s own complexities, generally it is a method of painstakingly calibrating absolute acoustic impedance into well log data, with the eventual aim of generating real, working petrophysical volumes.

 InSeis offers a robust algorithm that derives layered impedance models via very efficient searches of model space – in order to generate an optimised set of model parameters via simulated annealing. Essentially, with the aid of this fantastic little addon for Kingdom and soon Petrel Q1 2016, you’ll be able to use gloal optimisation techniques, which apply parameter constraints which use a macro-model and seismic amplitudes to reduce non uniqueness problems.

No small task – but be assured that InSeis supports multi-CPU processing, eliminating slowdown and making program lockups a thing of the past. Essentially, InSeis utilises however much power you decide to throw at it in as efficient a manner possible, and is very easy to parameterise and run. Furthermore, an added bonus in InSeis is that it has high-resolution output for relative impedance as well as absolute. Finally, it’s integrated within Kingdom and a breeze to link to Petrel.

Log and seismic matches are a vital part of simulated annealing inversion, and a trace near a well can be selected from a seismic data volume which uses a shortest distance criterion. InSeis tackles the well known problem of best matches not being shown from seismic traces which are generated from calibrated sonic logs. 

All in all, InSeis is a prgram thats robust and strong, and capable of deriving layered impedence models with little to no headaches.