Amerada Hess (UK) are taking quality in data management seriously. They consider that their exploration databases form an important strategic resource, contain millions of records compiled from numerous sources, over many years. Current attention to this asset has led to the conclusion that errors may exist in practically all of these datasets and to consider the limitations that such defects impose rather than ignoring them. Earlier last year Amerada Hess recruited Dr Paul Duller to co-ordinate their approach to data quality and lead a series of initiatives designed to improve the quality of data held their exploration databases. Commenting on this, Dr Duller said "A key element of our exploration success to date is access to accurate and reliable information and a clear understanding of the nature, origin and quality of data in use. Geological data by its very nature poses particular problems in terms of data quality, however quality assurance procedures can be applied to ensure the accuracy of the data and safeguard any associated exploration activity. A number of initiatives focusing upon data management and the quality of our seismic and well data are already underway."
Line-naming
As an example Duller cites a major clean up effort underway to standardize seismic line naming conventions. This problem arises from a historical laissez-faire approach to line naming conventions and poor data capture standards within the industry which has resulted in serious discrepancies between physical seismic data and their corresponding navigation database records. Although the industry is moving towards the increasing integration of these data types, the volume of data involved within a single companies archives (over 100,000 navigation lines and 200,000+ sections) places major logistic constraints upon how this problem can be resolved.
methodology
Amerada Hess have developed a structured methodology to reconcile this problem by matching navigation line names with their corresponding physical records, despite their apparent differences. This involves the generation of a line alias (a standardised and expanded form of the line name) and progressive, stepwise comparisons of records from both the physical and navigation datasets. Multiple match parameters (including the line alias) are used to provide additional levels of confidence in the match results, while user-defined transformations and filters serve to enhance the relative success rates achieved by this approach. Using this approach Amerada has been able to achieve in under six months, what would have taken years of painstaking manual cross-referencing using any other approach.
Click here to comment on this article
© Oil IT Journal - all rights reserved.