SPIRIT

A knowledge-based well test interpretation tool

Client:

The Petroleum Science and Technology Institute, Enterprise, Shell, Amoco, Bow Valley, Elf-Enterprise.

Problem:

The motive behind the SPIRIT project was the need to improve the quality of well test interpretation as it is carried out within oil companies.

Reservoir engineers have access to well test interpretation (WTI) software, but problems exist in the use of these programs. In order to perform a full interpretation, an analytical model (called a WTI model) has to be selected, based on the well test pressure response data and the the known geological and engineering data about the well being tested. Selecting the most appropriate model is a complex task, requiring significant expertise which most engineers interpreting well tests do not have.

Approach:

SPIRIT was a collaborative project between the Department of Petroleum Engineering at Heriot-Watt University and the Artificial Intelligence Applications Institute of the University of Edinburgh (AIAI).

The project began in January 1991 and was separated into three phases. The first phase was a design phase in which the requirements of potential users were elicited and the overall architecture and functionality of SPIRIT was decided upon. The second phase was a research phase in which the implementation of various of the modules that make up SPIRIT was explored, as were ideas concerning the user interface. At the end of this second phase a working demonstrator was produced, which gave sponsors a feel for the potential capability of the final prototype. The third phase of the project entailed agreement on the design of SPIRIT and the implementation of the final prototype.

SPIRIT was delivered to PSTI, together with a technical description of the software, in April 1993. The next phase of SPIRIT was to take the software and introduce it into the sponsoring companies.

A fuller article about the SPIRIT project, was written as a chapter for the book "Artificial Intelligence in the Petroleum Industry", B. Braunschweig and R. Day (eds.), Editions Technip, 1995. ISBN 2-7108-0688-6

For more information contact AIAI at:
tel: +44 (31) 650 2732
fax: +44 (31) 650 6513
email: AIAI@ed.ac.uk

---

Return to AIAI information