Browse Prior Art Database

System and Method for Adaptive Hierarchical Meta Level Modelling of SAGD Processes

IP.com Disclosure Number: IPCOM000245857D
Publication Date: 2016-Apr-13
Document File: 2 page(s) / 28K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a system and method for meta-level modeling of Steam Assisted Gravity Drainage (SAGD) operations, formed based upon similarity of the underlying system to analougous systems for which models have been established. Starting from an accostumed meth-level model, as more production data arrive, the model gradually evolves and attunes into the specific circumstances of the modelled system and thereby enhance prediction fidelity.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

System and Method for Adaptive Hierarchical Meta Level Modelling of SAGD Processes

Steam Assisted Gravity Drainage (SAGD) is a non-conventional oil production methodology in which steam is injected through injector wells and drained oil is extracted through producer wells. The operation involves several controls, such as the rate of steam injected, steam allocation to various sites, gas casing pressure, extracted emulsion pressure, etc.In addition, the apparatus is instrumented with sensors providing partial information regarding the state of the system. Examples of observable factors include emulsion rate, temperature profile along the well, sub-cool profile, bottom hole pressure, gas blanket pressure, and steam injection surface pressure.

Determining the optimal set of controls that maximize the yield of the system (e.g., maximize emulsion, minimize cumulative steam to oil ratio, maximize Net Present

Value, etc.) while honoring operational constraints (e.g., upper bounds upon bottom hole pressure, minimum temperature, sub-cooling, etc.) is a challenge.Determination of the optimal set of controls requires a predictive model that links a given set of controls with the anticipated quantities of interest (typically a subset of the observable factors).Development of a predictive model is particularly challenging as the underlying physics is complex, involving heat transfer (diffusion and advection) and flow in a porous medium, as well as chemical processes.

Physics-based predictive models require the prescription of a large number of nuisance parameters (e.g., porosity, permeability, heat coefficient throughout the entire field) for

which there are no definitive means to determine in realistic settings. Consequently, ad-hoc, often generic, values are typically prescribed. Such values may not properly distinguish between the specific characteristic of each individual well, and therefore fail to provide high fidelity results.

As data-driven (i.e., statistical) methods are agnostic to the underlying physical process, such methods' ability to offer reliable prediction is limited. In particular, as the underlying system is causal, it is likely to respond differently to the set of cont...