Abstract:There are various rock types in tight calcarenaceous sandy conglomerate reservoirs which includes calcarenaceous sandstone, lithic sandstone, carbonaceous sandstone, conglomerate, and mudstone. However, there is little difference in logging response between different rocks types and they are difficult to be identified. The identification of different rock types of tight calcarenaceous sandy conglomerate formation has become a key factor restricting the exploration and development of the reservoir. In order to solve the problem, the log response of different rock types and their differences internal cause were analyzed in detail based on multi-level cross-plots, and the distribution areas of different rock types on the cross-plots were determined. Firstly, the rock types of calcarenaceous sandy conglomerate reservoirs were divided into four categories: conglomerate, sandstone, coal and mudstone, and the resistivity, gamma and other logging methods were used to identify these four rock types. Secondly, according to the gravel content, the conglomerate was divided into two small classes: sand-bearing conglomerate and sandy conglomerate, and the resistivity, neutron and other logging methods were used to subdivide the conglomerate. Thirdly, sandstone was divided into carbonaceous sandstone, coarse-medium-grained calcarenaceous sandstone, fine-grained calcarenaceous sandstone, calcarenaceous siltstone&conventional lithic sandstone according to its composition and granularity, and the resistivity, gamma and other logging methods were used to subdivide the sandstone. Finally, the identification method of four categories and eight small classes was established based on multi-level cross-plots, and the identification results were in good agreement with core and thin sections names. This method provides a sound basis for clearly understanding the logging characteristics of each rock type of tight calcarenaceous sandy conglomerate formation, establishing the lithology profile, and determining the favorable reservoir distribution.