In this study, Monte Carlo simulation and stochastic modelling analyses, in combination with geological-based concepts, are used to improve and predict the petrophysical parameters of the hydrocarbon-bearing reservoirs. The stochastic input is supplied in the form of probability density functions, correlation, and variance-covariance coefficients as determined from actual well logging data. The multivariate statistical analyses are used to produce correlated multivariate stochastic variables. Before going on the simulation model, two types of constraints (hard constraints and simulation constraints) are used to guide and control the simulation procedure. Monte Carlo simulation algorithm is then performed by running a 100 times iteration in front of each depth increment using correlated multivariate variables.
The present simulation model is applied in two different case studies (Northeastern offshore Nile Delta and Ras Fanar field). In the offshore Nile Delta, the model is applied to improve the petrophysical parameters of the gas-bearing reservoirs. All logging data are simulated so as to, the uncertainty in the interpretations due to some errors like that in the statistics of gamma ray counting and in the estimation of apparent porosity from dual neutron log, are greatly reduced. Different Monte Carlo-based distributions are used to derive the different parameters necessary for reservoir evaluation especially those of shale. The model is used also to predict the petrophysical parameters of wells providing that a satisfactory topological/geological neighbourhood hard constraint is well developed in the area. In Ras Fanar field, the simulation algorithm is used to solve the cycle skipping problem which is another important and common problem concerning with sonic log due to the characteristic acoustic properties of the Nullipore carbonate reservoir. The simulated petrophysical parameters are found to be more realistic and representative for the actual petrophysical parameters of the reservoirs than those estimated using conventional logging analyses. So, this simulation model can be used not only as enhancement procedure in similar areas where the information is scarce or of bad quality and even in new areas to secure optimum use of the available data, but also as a predictive tool to predict the missed sections of logs and/ or to predict the petrophysical parameters of a new exploratory well. |
