Gas Hydrate Reservoir Stochastic Simulation Research And Application Based On Multi-well Data

The research on natural gas hydrate has become a hot topic of energy scientificresearch in the world gradually. It is widely distributed in permafrost zone andsubmarine, its resource amount is huge, some experts boldly estimated that the naturalgas hydrate is2times of all the coal, oil and natural gas in carbon content. China hasvast sea and a lot of natural gas hydrate enrichment area. In the paper an appropriateporosity and saturation model was established in the natural gas hydrate formationbased on the existing data. By the method of stochastic simulation, the attributeparameters (porosity, saturation) on the outside of the borehole area were forecasted;the results can provide necessary basis and parameters for further study andexploration in the region.Based on the analysis of the basic physical properties of natural gas hydratephysical properties and area formation, the formation of natural gas hydrate in thestudy area has been clear; the porosity calculation method (density porositymethod after correction), saturation calculation method (optimization of the Archieformula and the dual water model) of the study area have been established on thebasis of the study of previous work. At the same time, this paper proposes the use ofprincipal component analysis method combined with layer difference method andcluster analysis method of wells in the study area for automatic hierarchicalprocessing; using principal component analysis method to comprehensive articles ofhigh resolution of logging curves characteristics and reduce the curve for the stratifiedseveral convenient calculation; clustering analysis can avoid some spikes caused bysmall delamination after using intra layer difference method. On the basis of singlewell zonal, do lateral comparison analysis with the hierarchical data for stochasticsimulation research of the study area level data.In the paper, the attribute data of the outside borehole were simulated using themethod of stochastic simulation in the Petrel software. Comprising of the advantagesand disadvantages and the applicability of the stochastic simulation method,sequential Gauss simulation was chosen as simulation method. Stochastic simulationis generally divided into three steps: data preparation, model establishment andsimulation; the data includes the basic information data, well logging data, loggingdata and hierarchical data etc. Use the processing level data to establish a model, thengrid the model, each grid plane represents100m*100m, in thickness we use layer thickness ratio method according to the classification accuracy in general about1m;before simulation, do statistical analysis and spatial correlation analysis on attributedata including the most value, variance, covariance and correlation coefficient and soon; spatial correlation analysis is to obtain a regional variation model. Finally, aftertesting through statistical comparison and sampling method reflected the dimensionaldistribution of the attribute data (porosity, gas saturation) of the whole area.On the basis of the existing data, this paper deals with the obtained attributedistribution basically reflects the distribution of the study area, the distribution ofporosity is more accurate, the distribution of containing natural gas hydrate saturationconforms to the basic in the trend, and the specific accuracy is not as high as porosity.The next step is to collect more geological, seismic data, improve the reservoirdatabases, and then the simulation would be more precise.