No abstract

There has been a growing interest in geomechanics in the oil industry. From simple rock mechanical properties to fully coupled reservoir geomechanics simulation, real-time wellbore stability monitoring, stimulation job design, or sand management, the value of geomechanics has now been demonstrated and is applicable throughout the entire life of a field. Because it has become a key component for a number of departments, geomechanics is now considered a discipline by itself, and most of the major oil companies have established their own teams of rock mechanics experts. The mechanical earth model is the core of any geomechanical work. It is an integration process characterizing the field in terms of rock mechanics properties. Like any other type of project, it requires a minimum amount of specific information to be successful. Practically speaking, because data acquisition programs rarely focused on rock mechanics aspects in the past, it often appears that part of the key data required to perform a proper evaluation is not available. In such a case, one may have to use correlations to estimate the missing information or apply default parameters with their associated uncertainties. Ultimately, there is no equivalent to real data to ensure the accuracy of any study. The paper summarizes the data requirements for a reliable mechanical earth model and proposes a standard data acquisition program that would guarantee the availability of critical information. It will demonstrate that most of the data required are needed for other disciplines as well. The key is to make sure it is systematically acquired. Introduction Better understanding of the concepts and theory, technical advances, value added demonstration and recognition of the discipline have significantly changed the way geomechanics is perceived in the oil industry, resulting in an increased number of research and studies being initiated. A number of major oil companies have now established their own teams of experts to focus on rock mechanics related topics. From exploration to abandonment of the field, geomechanics relates to most of the technical aspects of the life of a reservoir. Pore pressure prediction, wellbore instability planning and management, casing design and well engineering, open hole completion stability, solids control, perforation design, reservoir stimulation, fault activation, subsidence and compaction, fully coupled reservoir geomechanics (etc…), it has been demonstrated and acknowledged that the range of application for the discipline is rather large. Rock mechanics is not new to the oil industry, but the concept of mechanical earth model (MEM) was only introduced in the late 1990s. The MEM consists in integrating data from various sources into a model that provides the rock mechanics parameters for a field, a well or a reservoir. Standard outputs from the model include elastic properties, rock strength and in-situ stress magnitude and direction. Those represent the main input to any subsequent failure analysis. With time, the MEM has become the core of any geomechanics work and potentially the most important step of the workflow.

Thin-bed sequences, especially those made up of alternating weak and strong sand layers with thickness of less than 1-ft, present a particular challenge in the generation of mechanical earth models (MEMs) for geomechanics design and analyses. Because these beds are thinner than the tool resolution, their heterogeneity and the variations between adjacent layers are often undetected by logging tools normally used to acquire geomechanics data. In other words, these tools cannot resolve the existence of alternating strong and weak sequences; they provide only averaged mechanical properties across these sequences. In one such case, an MEM and sanding study based on wireline sonic data indicated no tendency for sanding across a thick reservoir interval at the drawdowns required for economic oil production. The somewhat flat, log-derived curve of unconfined compressive strength (UCS) indicated that the entire reservoir was above the critical sand-free UCS state; however, laboratory data had revealed a large spread in UCS values, including rock strength values that are indicative of a significant sanding risk. In fact, the field had a history of solids production with depletion. Using the high-resolution images (with vertical resolution of 0.2 in.) provided by a microresistivity imaging tool across the same thin-bed intervals, and using alpha processing and other petrophysical log enhancement to further improve data interpretation, it was possible to generate a high-resolution (around 0.2 in.) porosity log. These downscaling techniques were combined with an empirical UCS-porosity relationship obtained from the available laboratory data. The resulting downscaled UCS log exhibited the wide range of strength values that were known to exist across the thin-bed sequence and were consistent with the history of sand production in the field. The remaining sanding evaluation for the well was completed using this integrated petrophysics and geomechanics approach. When the MEM and sanding analyses were extended to other wells in this field, the results matched well with the actual depletion curves and sanding history data. Introduction The Sarir field in Libya, operated by the Arabian Gulf Oil Company (AGOCO), has been experiencing progressively worse sand production problems in an increasing number of wells since 1984.[1,2] A geomechanics and sanding study initiated in 2004 investigated the probable cause of this problem and provided the information and interpretations needed to select an appropriate sand management solution, assist in future development decisions, allow appropriate completions planning, and optimize future reservoir management. To this end, the results of a data audit identified what geomechanics-related information existed that might be used to characterize previous sanding problems and diagnose their root causes. A laboratory testing programme using cores from the Sarir field obtained measurements of rock mechanics parameters relevant to stress modelling and sanding prediction. [3] Integrating laboratory results with other geomechanics information available from log, field, and drilling measurements provided an initial MEM. This MEM contained most of the information necessary for subsequent geomechanics analyses and sanding evaluations.[4,5] However, one significant finding was that routine UCS determinations from sonic logs in the Sarir field were unable to detect thin weak zones having thicknesses below the tool vertical resolution (i.e., 42 in.) of the sonic wireline tool. Therefore, petrophysical log-enhancement processing6,7 was conducted to obtain high-resolution, downscaled porosity logs using microresistivity image data. These data were then used to derive high-resolution (i.e., downscaled) UCS logs through UCS-porosity correlations. The UCS data from this process, which exhibited the same strength heterogeneities observed in the core data, provided the improved description of rock strength in the MEM that was needed in the subsequent sanding analyses of the thin-bed sequences.

No abstract

This paper was also presented as SPE 100948 at the 2006 SPE International Oil & Gas Conference and Exhibition in China held in Beijing, 5–7 December 2006. Abstract Sand production is a major concern for many operators. It can impact production, cause erosion in downhole and surface facilities, require additional separation and disposal, and lead to significant economic loss. On the other hand, precautionary but unnecessary sand prevention will mean unwarranted reduction in productivity. Reliable sanding prediction analysis thus provides a basis for designs that achieve appropriate sand management strategies and maximization of economic production, and overestimates or underestimates of sanding risk increase the chances of serious economical loss. This raises the question of how accurate and reliable sanding predictions might be achieved without overcomplicating the analyses and without requiring complex lab and field data that, in most instances, will be unavailable or the acquisition of which will incur unwanted delays and costs. This paper presents the case of a sanding study for the Messla field in Libya; a field that has produced oil for more than 30 years. This field experiences massive sanding from some wells but experiences no problems with other wells. This variation made the Messla field an ideal candidate for a detailed sanding and geomechanics investigation aimed at optimizing completions and production and at dramatically reducing the current sanding without having to enter into a lengthy data acquisition programme or time-consuming modelling. In this study, sanding prediction analyses were conducted using a technique that combines easily measurable lab data, log data, and analytical calculations with empirical methods that are supported by the results from previously run rigorous and advanced numerical code. The result of this integration is a sanding analyses tool that uses input parameters such as rock strength, geostresses, and particle size to:account for plasticity effects that modify the strength behaviour of sands surrounding openhole wells and perforations during drawdown to reduce uncertainty and conservatism such as seen in simple elastic models,account for scale effects associated with different perforation and borehole diameters,provide a significant improvement and predictive capability over simple empirical methods,provide the above accuracies without needing complex or extensive lab programmes to determine advanced rock mechanics properties. The application of this approach to the Messla field and a later comparison of the results to actual field data and observations validated the analyses and methods used. The application and comparison also disclosed that the approach was not only able to provide results that closely matched field experience but was also able to predict correctly, to the year, the onset of sanding in wells. This paper describes the methods employed in this investigation, provides details of the data acquisition and processing required, and demonstrates that accurate sanding predictions can be achieved by focusing effort on certain input data, targeting and reducing specific uncertainties, and by employing pragmatic models that do not rely on over-complicated measurements and analyses. Introduction The giant Messla field is located in the southeast portion of Sirte Basin in Libya, approximately 500 km southeast from Benghazi (see Figure 1). The field, operated by AGOCO, has been producing for over 30 years, and since the mid-1980s some wells have suffered massive sanding while others have not. 1 A geomechanics and sanding study was initiated in 2005 to investigate the reason for this variation, to evaluate the severity of sanding risk in other wells, and to provide the information and interpretations needed to design appropriate completions, maximize economical production and optimize future reservoir management.