Abstract

Macroscopic fractures in the subsurface are difficult to characterize because cores and image logs from vertical wells rarely intersect vertical fractures with sufficient frequency to obtain representative samples. Consequently, the capacity of fractures to transmit fluid is often a matter of speculation. Here we report application of new core analysis techniques (Laubach, 1997) that are designed to provide information about the attributes of macrofractures, particularly in cases where such fractures are not directly sampled. The objective in this study was to predict if macrofractures are open or mineral filled. Analytical techniques include photomultiplier-based electron beam-induced luminescence (scanned CL) to detect microfractures and standard petrographic techniques to estimate authigenic cement percent volume in the rock matrix. Observations from millimeter-sized areas of samples determine the timing of cement precipitation relative to fracture opening. Synkinematic cements that form during fracture growth tend to prop host fractures open without occluding them, whereas postkinematic cements that form after fractures have opened are the main cause of closed fractures. Thus high postkinematic cement volumes suggest occluded fractures and destruction of fracture permeability.

A test of this concept was carried out on naturally fractured Wolfcamp A2 sandstone cores from Chevron’s Mitchell 1B #7 and University 29 #1 wells, located in the Pakenham (Wolfcamp) field in the Val Verde Basin. These low-permeability sandstones are interpreted to be deep-water turbidites sourced from highlands associated with the Marathon Orogeny and deposited in the Val Verde foredeep. Production is primarily from high-density turbidites that pinchout against slope shales. Mitchell 1B #7 is a gas producing well and University 29 #1 is a non-producing well. Analysis was conducted on 42 ft of 142 ft of core from the A2 sandstone (with 7 visible fractures) in the Mitchell 1B #7 well and all 131 ft of A2 sandstone core (with 21 visible fractures) in the University 29 #1 well. Fracture height is typically limited by turbidite bedding thickness. Average fracture height from FMI logs is 1.5 ft and, using the technique of Narr (1996), average fracture spacing is 3.2 ft. Natural fractures are nearly vertical and typically strike about N 60 E, nearly orthogonal to present SHmax. Some of the fractures are lined by cement but are largely open, whereas others are completely mineral filled. We sampled core in the same depth intervals as the visible fractures, then used analysis of microfractures to obtain volume of postkinematic cement and thus a prediction of the ‘openness’ of the macrofractures.

Preliminary results show that, as predicted from microfracture analyses, open fractures are lined and bridged by synkinematic quartz cement and filled (closed) fractures are occluded by postkinematic ankerite, barite, and other minerals. Textures within macrofractures also preserve evidence of complex dissolution and reprecipitation events. Although comparisons are ongoing, initial results suggest that matrix postkinematic cement volumes accurately predict the location of filled fractures in most cases. These results show that microstructural analysis is a promising approach for diagnosing fracture openness and for helping to assess fracture permeability.