The indentification for reservoir fluid based on the autocovariance function and implementation by matlab

In the log interpretation, it is an important task to research difference of the logging signal of different reservoir with mathematics comprehensively and accurately, so as to identify the fluid property of reservoir. The statistical trait of logging data can be reflected fully by its autocovariance function. In this paper, the sample autocovariance functions of the logging data of the oil-layer, gas-layer, water-layer and dry-layer were estimated, and were distinguished by discriminant analysis to identify the reservoir fluid. Software system for log interpretation was programmed by the software MATLAB. Basing on the result of the detection, and the analysis and calculation of logging data (SP, GR, CAL, RD), we processed 42 layers of three wells in an oil field, the interpretation agreement rate of SP and GR has been up to 100%, and the interpretation agreement rate of all has been up to 89%. This method can be considered to use for signal processing of other types.