Publications

Fracture porosity improves the reservoir permeability and improves recovery of oil in place. Fracture porosity determination from log data is rarely used by petroleum engineers because of preferences for calculation by core laboratory test or formation imaging. Generally, fracture porosity can be estimated through the log data (density, neutron porosity and transit time). If one of these parameters is lacking, the estimation using log data becomes impossible. This paper discusses neural network modeling using input gamma ray, density, porosity neutron, deep resistivity, and transit time for the first phase and without transit time for the second phase. Well data were normalized to obtain proper relationship in crossplot analysis. Crossplot analysis shows the relationship between log data and fracture porosity log. From the result, it can be seen that density and p-wave velocity have more affect to fracture porosity. The neural network proves to be an effective technique to estimate fracture porosity in carbonate reservoirs. Fracture porosity obtained by missing sonic log has close result too. The coefficient relation between these two porosities is 0.91 for depth 17,120 ft to 17,256 ft.