M. Kraaijveld’s research while affiliated with Shell Global and other places

We develop a neural network workflow, which provides a systematic approach for tackling various problems in petroleum engineering. The workflow covers several design issues for constructing neural network models, especially in terms of developing the network structure. We apply the model to predict water saturation in an oilfield in Oman. Water saturation can be accurately obtained from data measured from cores removed from the oil field, but this information is limited to a few wells. Wireline log data are more abundantly available in most wells, and they provide valuable, but indirect, information about rock properties. A three-layered neural network model with five hidden neurons and a resilient back-propagation algorithm is found to be the best design for the saturation prediction. The input variables to the model are density, neutron, resistivity, and photo-electric wireline logs, and the model is trained using core water saturation. The model is able to predict the saturation directly from wireline logs with a correlation coefficient (r) of 0.91 and an error of 2.5 saturation units on the testing data.

A well-known method to determine the hydrocarbon saturation distribution in a reservoir model is by using a saturation-height function derived from capillary pressure measured on core samples. This approach fails, however, in complex formations and does not use information from wireline logs. In this paper we use an artificial neural network to develop a saturation-height function for the complex Gharif Formation in Oman to predict the hydrocarbon saturation. Different neural network models were developed using different input variables. The optimal model was able to generate the saturation-height function with an error of 0.046 (fraction Of Pore volume, PV) using wireline logs, including the logarithm of resistivity, cation exchange capacity and porosity. This is a considerable improvement over conventional methods based on capillary pressure. The neural network model was then used to predict the saturation in the formation as a function of depth, and robust results were obtained.

The application of artificial neural networks (ANNs) in the petroleum industry is widely increasing after major developments in ANN design. In this study, ANNs were used to develop a model for predicting water saturation in shaly formations using wireline well logs and core data. A general workflow (methodology) was constructed in order to cover different design issues in the ANN modelling. The workflow presents how the neural network results can be interpreted in terms of investigating the contribution of input variables and comparing the results with other regression models. In addition, the workflow focuses on the relevance of the statistics of the data and the importance of determining the uncertainties in the original data before using it in the model. The data for this study was taken from a sandstone formation in Oman. The input variables to the model were density, neutron, resistivity and photo-electric wireline logs. A three layered feed-forward neural network model with five hidden neurons and Resilient Back-propagation (PROP) algorithm was found to be the best design. The neural network model was able to predict the water saturation directly from wireline logs with a correlation factor of 0.91 and a root mean square error (RMSE) of 2.5%.