Abstract: Based upon the theories of artifcial neural netvork and fuzzy mathematics, a computer system for integrated pay zone interpretation is developed on Sun workstation. There are 46 parameters used to recognize the mature of fluids in reservoir, such as data from core analysis, cuttings, sidewall coring, well log, gas log and dilling. The ownership to oil, gas, water or dry beds is calculated using fuzzy statistics for category attributes, which is a quant itative description for the attribute. Zones containing different fluids are characterized by ten output patterns in the developed dynamice neural network. The structure of the network is adjusted to fit different geologic cases and data obtained by changing the input parameters of the input layer and the number of nodes in the middle layer. The system was applied to some areas in Liaohe depression, and showed remarkable consistent results with well tests in the field.

Key words: pay zone study, integrated interpretation, system, neural network, fuzzy mathematics