New Correlation for Calculating Porosity Using density log by applying artificial neural networks and uncertainty analysis: An application in the Mishrif Formation, South Iraq

Petrophysical properties consider a major objective to determine potential reservoir zones, estimate hydrocarbons present, and appraise recoverable hydrocarbons. prediction of the reservoir's petrophysical characteristics is one of the most Petroleum engineers’ challenges that overcome by advancing and using soft computing technologies in engineering matters. In present study a new correlation is developed to calculate porosity in term of RHOB well log, by using data of three wells in Mishrif Formation southern Iraq, constructed by utilizing two techniques are multi-layer regression (MLR) integrated with artificial neural network (ANN) technique. P90 porosity results mean 10 % uncertain porosity and 90 % confidence. New equation results are equal or near to P90 Porosity this led to the new equation minimized the systematic and deterministic error. Considering all it knows; the current study provides a unique framework for porosity estimate by utilizing the ANN model.