| 摘要 |
The invention is a computerized process for discovering hydrocarbon by applying event recognition to oil&gas wells logs. Presently, geologists must manually s earch through thousands of oil&gas well logs to find a combination of sensor readings (the "Even t") that indicate probable hydrocarbon production. This manual search has two major limitation s: 1. Time - it takes a long time to search through thousands, or ten's of thousands, of wells logs looking for the desired correlations between Events from producing well and possible Events in various non-producing wells. In larger regional projects, this can take up to a year in duration and success is not guaranteed; and 2. Size limitation - while humans are very good at determining "what" to loo k for, we are very poor at holding very large amounts of data in a correlative form in our current memory. This means that we must correlate in a linear manner and keep the correlations small because we lose correlative power very quickly. This inability severel y limits the effectiveness of our efforts when attempting to establish a "big picture" vi ew of a large geographic area. The combination of these two factors limits the speed at which hydrocarbon prospects can be generated which is, in fact, is the primary obstacle to accelerated hydrocarbon production. The invention speeds up the prospect generation process by automating the requir ed correlations, thereby substantially increasing the production and recovery of hydrocarbon.
|
