What is Stochastic Modeling?

Stochastic modeling technology refers to the use of structural statistics of a geological body whose characteristics are known. Through some random algorithms, the distribution of this attribute in the unknown area is simulated to make it the same as the known statistical characteristics, so as to reach the simulated reservoir. Heterogeneity, modeling techniques for the purpose of predicting the distribution of well parameters.

Stochastic modeling

Subject: Oil and Gas

Use a statistical feature of a structure whose geological body has a known attribute, and use some random algorithms to simulate the distribution of this attribute in the unknown area, so that it is the same as the known statistical feature, so as to simulate the reservoir heterogeneity. Until the modeling technology for predicting the distribution of well parameters. It is a hot topic in reservoir description in recent years. There are many common algorithms for stochastic simulation: indicative point process method, Markov random function method, truncated Gauss method, two-point histogram method, mosaic process simulation method, probability field simulation method, index simulation method, and Boolean method. , Annealing simulation method, sequential Gauss method, sequential index simulation method, fractal random function method, LU decomposition method, turn-band method, etc. Different stochastic simulation algorithms can reflect different regional statistical parameters and spatial characteristics. In general, according to whether the stochastic simulation is faithful to the known sample values, it can be divided into two categories: conditional simulation and unconditional simulation. Conditional simulation means that the random simulation results must not only meet the geostatistical laws, but also keep the known sample values the same as the simulation values. Unconditional simulation only needs the simulation results to reflect known statistical laws. Conditional stochastic simulation is generally used in geological work. On the other hand, based on the content of stochastic simulation, stochastic modeling methods can be divided into two types, one is a discrete model and the other is a continuous model. Discrete models are used to simulate the distribution of reservoir geological bodies, such as sandstone bodies, shales, and rocks. The continuity model is used to simulate continuously changing geological phenomena, such as porosity, permeability, saturation, and mud content. Using truncation, the continuous model can also be applied to the simulation of discrete geological phenomena. Stochastic simulation technology has been used more and more widely in various aspects of reservoir description, such as the spatial distribution of lithofacies, porosity, sand body and fracture prediction, quantitative estimation of uncertainty of reservoir model, sampling design, flow simulation Sensitivity analysis and risk analysis in the process, etc. Its applications can be summarized as follows:

Parameter estimation: Perform local optimization estimation of various geological variables and reservoir parameters.

Reservoir heterogeneity research: Stochastic modeling technology is proposed for heterogeneity research. Stochastic technology is the bridge between observation points and unsampled points.

Comprehensive application of various data: Reservoir description involves multi-disciplinary and multi-type data information. How to systematically match and use various data information is very important. Geological statistics can do this job well.

Uncertainty description: The biggest feature of stochastic modeling is that it can give a variety of possible results, so it can objectively reflect the complexity of underground reservoirs and the limitations of people's recognition, and establish people's risk awareness.