Application Of Reservoir Uncertainty Modeling In Braided River-delta

Due to the incomplete information and reservoir heterogeneity,it is unable to accurately determine the distribution of reservoir and its properties.Stochastic modeling technique can characterize the uncertainty of geologic model by generate multiple realizations through changing the stochastic simulation path.Reservoir uncertainty modeling is a new technology developed on the basis of the stochastic modeling,and it paid more emphasis on the characterization and evaluation of uncertainty that resulted from data acquisition,modeling parameter and response of models.To improve the uncertainty appraisement of braided river-delta reservoir,taking a gas field in Donghai as an example,a quantitative method,using GIIP as the indexes and based on 3D-stochastic reservoir modeling,is put forward.Based on the modeling realizations of geological models which are singled out by experiments and describe the likelihoods of reservoir characteristic,the effects of uncertainty factors on modeling are estimated quantitatively and the statistic features of the main factors are figured out.Using the multinomial response surface of the main uncertainty factors and reserves as the substitute for modeling,probabilistic reserves are calculated by Monte Carlo method.This method is not only more objective and efficient,but also capable of identifying the key geologic uncertainties that influence the GIIP calculation and provides a new way to evaluate the uncertainty of GIIP.The main results were as follows:(1)Taking a gas field with relatively scarce well point data in Donghai as an example,this paper start from basic geological study,combing with the well point data and the seismic attribute inversion,proposed a probability constraint method based on multiple trend integration and a fine reservoir geologic modeling is set up.It also reveals the sweet-spot distribution and development pattern of reservoir physical property.(2)Choosing gas-water interface,cutoff of effective porosity,gas saturation,gas volume factor,sweet spot percent by volume and source direction as the key geologic uncertainties that influence the GIIP calculation.The various levels of factors are determined based on geological knowledge of the work area.With method of orthogonal experimental design,the work efficiency enhanced greatly.The result shows that: gas-water interface,cutoff of effective porosity,gas saturation are significant to the GIIP calculation.Also,the parameter combination is determined to obtain the optimal reserve.(3)The dominant nondeterministic variables that influence the GIIP calculation and their probability distributions are identified by the actual gas field data analysis,after which a modeling plan was designed in factor level using experimental design function of Minitab software.The experimental data and results were analyzed statistically by Plackett-Burman design,steepest ascent design and Box-Behnken design of response surface as a sequence.Multiple regression was performed on the nondeterministic variables and the GIIP calculated with corresponding models.Finally,the Monte Carlo simulation was run to obtain the GIIP probability curve and P10,P50 and P90 model parameters,which were used to establish the reference P10,P50 and P90 models.Cross plot of proxy equation-calculated vs.model-predicted GIIP is plotted,and it shows the two are excellent in fitting.