The main aim of well log interpretation is to determine the reservoirgeological and petrophysical properties downhole as well as properties fluidsand their distributions in a reservoir, and identification of hydrocarbon fromwell logs is the most important job. Neural networks, with their greatcapability in adaptation and expressing arbitrary complexity, seem especiallysuited to solve for the reservoir problems where there widely existuncertainty, fuzziness and nonlinearity. Much work has been published on theidentification of lithology and lithofacies from well logs using neuralnetworks. But very few, and almost no work could be found to deal withidentifying hydrocarbon from well logs in this neural network approach In thispaper, a supervised neural network (a multi-layer perceptron (MLP)) and twounsupervised neural networks (a self-organizing mapping (SOM) net and a fuzzyneural network (FNN) are used to identify oil from well log, and almost each ofthem can distinguish between oil, water, oil-water transitionand dry zonesdirectly from well logs, but with different accuracy. A case study from aChinese oil field is given to show their advantages and limitations, and theirimplementation is also given. The associated oil test results are available todemonstrate their dependability And it is shown that there is a very goodagreement between results oil test and oil identification from both LP and FNN.
Generally, the main task of well log interpretation is to determine thegeological and petrophysical properties of reservoir rock formation as well asproperties of formation fluids and their distributions. Obviously, well logsare the fundamental information source of well log interpretation, and theeventual aim of well log interpretation in oil exploration is to determine bothroperties of subsurface fluids and their distribution, which is not directlyfrom well logs and usually based on the predetermined geological and / orpetrophysical properties from well logs. During this conventional loginterpretation process, it often needs selection of empirical formula andvarious empirical parameters, which makes the actual log interpretationextremely complicated and subjective, and eventually lead to the extremedifficulty in enhancing the chance of success of well log interpretation.
Well log interpretation, in certain viewpoint, is to invert subsurfacegeological and petrophysical properties, as well as properties of subsurfacefluids and their spatial distributions from well logs. And well log responsesmeasured with wireline tools, which reflect the properties of the subsurfacereservoir, are in extremely complex relationship with various reservoirproperties, because of the high heterogeneity and anisotropy widely existent inreservoir medium. Thus the solution of the conventional log inversion is not ofsingularity. So, primarily, log analysts could simply interpret well logsqualitatively, aided with their field experience obtained from success and failby trial and error, that is, visual log interpretation of oil. Because ofadvances of logging technology and improvement of various experiment techniquesas well as introduction of advanced scientific technology such as computers, the log analysts began to quantitatively determine both the properties ofsubsurface fluids and their distributions from well logs, according to variousempirical formula obtained by experiments and various mathematical and physicalmodels established on basis of theoretical analysis, which eventually made welllog interpretation grow to be a largely integrated engineering science, thatis, the computerized og interpretation stage started when the identi
