Abstract:Many reservoir parameters of interested reservoir interval,such as lithofacies,average porosity,effective thickness and sand(or shale)percentage content,can be derived from borehole data;they then form a characteristic space of multidimensional reservoir that involves several boreholes.The cluster analysis of these sampies is achieved by using the neural network of self-organized character mapping so as to find reliable borehole groups and the reservoir characteristic categories corresponding to the groups respectively.Then,the characteristic subset showing the classification ability of seismic traces is determined by doing iterative calculation for borehole-side seismic traces;and the clustering of the seismic traces can be achieved by using the subset.Finally,a mufti-layer feedforward neural network is used to map the seismic traces to the geological space consisting of reservoir parameters so as to know lateral reservoir distribution.The article describes the example of how we use the method to quantitatively analyse the reservoir parameter space in a three-dimensional seismic area in the northern Biyang depression.