Improved modelling and control of oil and gas transport facility operations using artificial intelligence

In recent years, the application of artificial intelligence (AI) based tech niques to a wide range of industrial processes has become increasingly comm on. One reason for this development is the level of maturity of both theory of AI concepts and its implementation into application tools for commercia l use. Another very important reason is the persistent drive of many indust ries to increase efficiencies and the realisation that this requires more e ffective processing of gained knowledge and information. In the oil and gas industry, due to high saturation levels of many production fields and the complex nature of processes, the need for increased efficiencies and highly effective processing of a large amount of information is particularly evid ent. Some organisations have recognised the opportunities offered by AI-bas ed techniques and started exploiting them in order to improve knowledge and information handling and process efficiencies. This paper discusses the ap plication of two AI-based techniques, fuzzy logic and artificial neural net works (ANNs), to specific problems related to the operation of oil and gas transport facilities. The background for the work, which is carried out in a co-operation between a university and a leading engineering service provi der, is described firstly. This is followed by a brief summary of the funda mentals of the AI techniques considered with respect to their use for indus trial purposes. Then, two case studies are presented. The first case study demonstrates the application of fuzzy logic to the control of a pump statio n in a pipeline system whilst the second case study shows the use of an ANN for the determination of important pipeline characteristics. Problem backg rounds, design procedures and outlines for the implementation of the used A I techniques are given. Finally, benefits of the adopted approaches are hig hlighted and the wider impact on both industry and research community is di scussed. (C) 2000 Elsevier Science B.V. All rights reserved.