The carbonate-evaporite depositional combination of Late Jurassic age is the top member of four upward-shoaling carbonate-anhydrite cycles in Upper Kimmeridgian age. The weak depositional contrasts in carbonate ramp setting make the lateral seal configurations subtle and tough to recognize. Multiple attribute analysis based on Artificial Neural Network (ANN) can delineate the internal character of the reservoirs and seals in a consistent way.
In order to recognize the sedimentary facies within this reservoir interval, multiple seismic attributes input to an unsupervised ANN. Unsupervised ANN is a powerful classification technique, which is implemented using a single layer perceptron network. The network is trained by comparing the neurons to the input vectors using competitive-learning techniques. Once a neuron migrates to the center of the class, the network stabilizes and training is finished. Without prior information, further sedimentary facies are recognized by integrating local geological knowledge.
The depositional environments in the study area are well characterized by unsupervised ANN and are consistent with the drilled wells and the geological model. Lagoonal deposits, ramp crest shoal and proximal deposits are recognized within the study area. The sedimentary facies recognition helps define potential areas for favorable prospect definition and hence prospect ranking.