Diffentiation of wells in zones with residual reserves of oil, using neural network modelling

The grouping, based on a set of signs of a well stock of layer of the large-scale oil deposit of Western Siberia, using artificial neural network, was carried out. The initial set, including 555 objects, was used for grouping, 95% of objects were chosen from them as the training set and 5% - as test. For training of neural network 17 signs, characterizing both: geological and physical, technological parameters of layer were accepted. As a result of control and the subsequent training of neural network 4 groups of wells, the closest in geological and technological parameters were allocated. For each group of wells the parameters, characterizing uniqueness of the chosen group, are described. The binding to localization in the spatial relation of layer and to residual reserves of oil is given. Recommendations about involvement of residual reserves of oil in active development are offered for each group.
Materials and methods
Geological fields’ data, proxy model, artificial neural network, geological and statistical models, geological and hydrodynamic modeling.
Results
The carried-out grouping and identification of a producing well stock, according to geological and technological signs, allowed to estimate the structure of a producing well stock; to define a marginal well stock in the signs, allocated on set groups; to establish the possible reasons of extremely low outputs of wells. It is established that the condition of residual stocks at this stage of development is determined as technological and geological factor (high heterogeneity reservoir properties by the area and a section). In this regard, recommendations will be concerned with the active development of wells with high specific residual stocks, taking into account their geological and, whenever possible, technology factors. Work with the existing fund of the operating wells is considered to be the most perspective.
Сonclusions
Holding geological and technological actions in zones of the arrangement of marginal wells of the groups will allow to raise or restore the energy condition of a layer, to increase the efficiency of wells, to reduce the water content of the production, and, in general, to involve reserves of low-permeability layers in the development.