Aiming at the most familiar gas-water stratified flow in the highly deviated or the horizontal gas wells, based on the electromagnetic image logging technology, the identification method for stratified smooth flow and stratified wavy flow is researched. Firstly, the similar characteristics in the flow section of all the gas-water flow patterns in the horizontal or inclined pipe are extracted. According to the abstraction of the characters, four material distribution models are built up. Secondly, the electromagnetic image logging measurement in response to all the four material distribution models are simulated by the finite element method and the peculiarity related with the flow patterns of each response is analyzed. Thirdly, the character related to the models is computed as a parameter from the measurements. By using the parameter, the stratified model is separated from other models and water holdup rate is computed. Finally, a pattern recognition method is used to identify the transformation rules of the material distribution models in the flow section; the gas-water stratified flow is distinguished. With this method, the gas-water stratified flow patterns can be identified and the water holdup rate can be computed, and the method is quicker and more accurate than analyzing reconstructed images.
