Project staff and support

Lian Duan (Intern, University of Oxford)
Ralph Stadie (Company Supervisor, Paradigm Geophysical)
Chris Farmer (Academic Mentor, University of Oxford)
Melvin Brown (Technology Translator, Industrial Mathematics KTN)

This Internship project was carried out at Paradigm Geophysical , in conjunction with the University of Oxford. It is part of the KTN's Industrial Mathematics Internships Programme, co-funded by EPSRC. Start date: February 2010; duration: 3 months, co-funded by EPSRC.

Paradigm software solutions provide the vision for business and engineering decision-makers in subsurface disciplines to proceed with less uncertainty and more confidence. Paradigm tools minimize risk and optimize responsible management of the earth's subsurface assets. The Intern worked with the R&D group within Paradigm and developed and implemented improved solution methods in the development version of a well testing package. The Intern therefore become familiar with the software design and development process.

The main objective was to modify an existing deconvolution algorithm used in the analysis of pressure transient data to smooth out the derivative response so that it appears more like a physical response (i.e. continuous) rather than a fabricated representation. The current method deconvolves pressure data influenced by multiple rate sequences into a pressure response due to a single constant rate. The second objective was to develop a confidence indicator for the calculated solution.

“The problem was by no means trivial and we were extremely pleased at the amount of progress that Lian made over a relatively short period of time. He also showed a great deal of initiative in looking at the problem using a number of different approaches and we are very satisfied with the results he achieved.” said industrial supervisor Ralph Stadie, Programme Manager, Paradigm Geophysical.

Project description

There have been many advances in well test analysis over the last thirty years of which decovolution is probably the most recent. In the context of the analysis of pressure transients recorded during the testing of a well, deconvolution refers to the extraction of an equivalent single rate pressure response from the actual well response and is usually a continuous variation in rate and pressure. Within its known limitations, it is a powerful technique since it allows the engineer to see the transient response of the well and reservoir without the complexity of overlapping pressure data in each flow period. However, a smooth and interpretable pressure response will be the key to the success in both the analysis of the deconvolution and the prediction of the well behaviour.

The project improved the existing deconvolution algorithm used in the analysis of the pressure transient data to smooth out the response so that it appears more like a physical response (i.e. continuous) rather than a non-realistic representation. This is also the main objective of this internship. In addition, a total production constraint has also been introduced to ensure that any rate adjustments conform to real-world expectations.

“This was a well-chosen project. It was a practical problem of importance to the company, it was challenging to Lian and it provided a model example that demonstrates the power of the Bayesian approach to inverse problems. The experience of software development in a commercial environment will be beneficial for Lian's future either as an academic researcher or as a software developer.” said academic mentor Chris L. Farmer, OCCAM, University of Oxford.

related resources:

»	Deconvolution of well test data
	Technical summary and references

other projects:

[Find other Energy and utilities projects]