Sandbody architecture analysis of braided river reservoirs and their significance for remaining oil distribution: A case study based on a new outcrop in the Songliao Basin, Northeast China

The reservoir architecture analysis of braided rivers, especially falling-silt seam forms, has played a key role in predicting remaining oil distributions. However, no studies have used architecture analyses that document braided river outcrops and researched the tapping of the few remaining oil distributions based on outcrops in the Songliao basin, northeast China. In this paper, the architecture characteristics and remaining oil distribution of braided river reservoirs are studied using a combination of an outcrop, modern deposition and subsurface well data. The new 8–13 m thick Lower Cretaceous Quantou Formation outcrop of the Songliao basin is a braided fluvial succession arranged in one large fining-upward cycle. Eight facies (Gt, St, Sm, Sh, Sp, Sw, Fl and Fm), four architecture elements (CH, DA, LV, and FF), and three orders of bounding surfaces (third-, fourth-, and fifth-order) are recognized. A new distribution pattern of falling-silt seams and a braided river architecture model are presented according to the analysis of the outcrop. In the mid-channel bar, the falling-silt seams thin from the mid-bar to the bar tail following the flow direction. Each falling-silt seam is oriented tangentially to the basal surface of the mid-channel bar, and the upper falling-silt seam extends farther than the lower one. In a Daqing Oilfield exploitation block in the Songliao basin, while channels and bars are the main reservoir units, they have different remaining oil distribution patterns. For bars, water injection wells located at the mid-bar, zonal injection technology, the drilling of horizontal wells, and proper well patterns are proposed. Fourth-order bounding surfaces, single braided channels, stacking patterns, and the lateral blocking of levees and floodplains are the key factors affecting the remaining oil distribution in channels.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.