Fast fuzzy modeling method to estimate missing logsin hydrocarbon reservoirs

In order to deal with huge amounts of computational complexities, conventional modeling systems always had to choose a tradeoff between accuracy and rapidity and usually one prevails over the other. Thus there is a need for a solution which provides acceptable accuracy and rapidity at the same time. In this research we propose a new fast fuzzy modeling method (FFMM) using Ink Drop Spread (IDS) and Center of Gravity (COG) operators. We applied this method to estimate missing logs of sonic and density. In the petroleum industry, characterization of pore–fluid pressures and rock lithology, along with estimation of porosity, permeability, fluid saturation and other physical properties are crucially important for successful exploration and exploitation. For many reasons, such as incomplete logging, inappropriate data storage and measurement errors, log suites are either incomplete or unreliable. By applying the proposed method, we estimated sonic and density logs. Correlation coefficients and MSEs for DT and RHOB logs were equal to 0.92, 21.07 and 0.85, 0.006 respectively. These results show that, despite the algorithmʹs very fast computation speed, its performance is comparable with that of methods like artificial neural network (ANN) and conventional Fuzzy Logic (FL); but the latter requires large amounts of storage and computing time, while the resources needed for the proposed method are limited to fixed values with respect to the number of data points.