Knowledge of accurate rock strength is essential for in situ stress estimation, wellbore stability analysis, sand production prediction and other geomechanical applications. Reliable quantitative data on rock strength can only be obtained from cores. However, cores are limited, discontinuous and often biased. Consequently, rock strength evaluation is primarily based on log strength indicators, calibrated where possible against limited core measured values. There are a number of published log-core strength correlations that can be used for rock strength modelling. These empirical relationships are developed for specific rock type, age, depth range and field. Their general applications, therefore, need to be critically assessed on a case by case basis. This paper briefly: (i) outlines the best practice for obtaining quality rock strength data from core tests; (ii) presents common empirical rock strength equations for sedimentary rocks and (iii) discusses ways of improving rock strength estimates.While some equations such as porosity-based or sonic log-based rock strength models work reasonably well, rock strength variations within individual rock properties show considerable scatter, indicating that most of the empirical models are not sufficiently generic to fit all rocks in the database. Like any other physical rock properties, the variation in rock strength in a given sedimentary rock is controlled by mineralogy, sedimentology and micro-structure of the rock and simple log-derived rock strength models need further modification and classification incorporating these geological characteristics.This paper has shown that when sufficient core rock strength data exists, applications of computing techniques, such as fuzzy logic and cluster pattern recognition, coupled with sedimentary facies analysis and diagenetic classification can improve strength estimation. Semi-continuous impact energy logs using portable non-destructive testing tools can be correlated with petrophysical logs to generate mechanical facies and improved sampling for conventional rock testing.
The methodology for predicting sand production is in general constant across the industry. That is the determination of formation strength and field stresses and the application of them to a failure model. However, the variety of models available and their applicability and accuracy can be confusing with the results not always representing what is experienced under production conditions. This paper introduces a more holistic approach to sand production prediction which not only utilises numerical analysis, but also includes a qualitative approach using geological information. With respect to the numerical analysis the determination of various parameters used in modelling sand production and their effectiveness for different reservoir and production conditions is discussed. An overview of the various tests considered useful in calibrating these parameters is also presented. The geological approach discusses the impact of mineralogical, depositional, structural and diagenetic factors which can impact on the propensity for sand, but which are not fully taken into account by a purely numerical approach. The advantages and disadvantages of each of these aspects are presented and synergies identified between numerical and qualitative analysis. Case histories are presented which show that each of these methodologies, when used independently, can present dissimilar results. This may ultimately lead to recommendations which can result in costly operations and unnecessary equipment deployment. However, when numerical modelling and geological methods are used in conjunction with one another, the results present a more realistic and practical view of the formation sands and their potential for solids production during the life of the well. Introduction As an increasing number of fields around the world enter their mature stages of production, the impact of depletion and increasing water cut is having a dramatic effect on their propensity for sand production. It is estimated1 that by 2010, half of all production wells and one third of injectors will produce sand. There is continuing demand for improved technologies to mitigate solids production as the industry strives to extract hydrocarbons from deeper, hotter wells in increasingly hostile environments. This is being met by improved gravel pack placement techniques and advanced expandable technologies as well as the proven methods of sand control. If the factors which contribute to sedimentary rock strength were few and constant, then all sandstones would be of similar strength and a reliable and universally applicable methodology for sanding prediction would now be in place. This is, of course, not the case and is what makes the accurate prediction of sand behaviour somewhat unpredictable. However, there remains uncertainty as to the accuracy of geomechanics work utilised in qualifying and quantifying a given well's propensity for sanding. This process is widely acknowledged; that is, the determination of the formation strength and impact of the in-situ stresses identifying the conditions under which failure of the rock will occur. A typical sand prediction model structure is shown in Fig 1.
Knowledge of rock mechanical properties, including unconfined compressive strength (UCS), is essential for accurate geomechanical evaluations. Reliable quantitative data on UCS can only be derived at specific depths from laboratory tests on core samples, typically through destructive tests. Rock strength evaluation is primarily based on well log strength indicators, calibrated where possible against limited measurements on cores. A number of techniques have been developed to supplement plug-based destructive tests to indicate the rock's strength. Scratch and Schmidt hammer tests are examples of such techniques with the ability to provide continuous or semi-continuous, fine-scale measurements of rock mechanical properties. Both of these index tests are non-destructive; do not cause significant damage to the core and no special core preparation is required prior testing. In this paper we examine laboratory test results from both conventional triaxial testing on plugs and semi-continuous, non-destructive, index testing along core sections in conjunction with well log calibration, for a range of rocks from oil wells in different geological settings. The rocks represent a range of strengths from very weak to very strong (competent) rocks with UCS ranging between 1 and100 MPa. We show that the use of generalized (global) correlations of index testing versus UCS may provide reasonable estimates, however a local calibration based on triaxial tests would still be required for more accurate UCS profiling. Once calibrated to a few conventional plug tests, the continuous index test results can be correlated with petrophysical logs to generate a reliable strength log, coupled with sedimentological and diagenetic classification. We show that besides the direct application to geomechanical evaluation, the index tests can be used as a screening tool to optimize plug sampling to characterize the range of heterogeneity of the cored interval and also for filling the gaps in the core strength data where plugging is practically impossible or difficult, due to natural or coring-induced reasons.
Despite continued advances in sand production prediction software, numerical modelling does not always agree with observed sandface behaviour once a well is flowing. Where core is available, computer models should be calibrated to laboratory determined rock strengths, measured on plugs cut from the core. The number of core plugs is often restricted due to time limitations and availability of good quality core. These may not be fully representative of the entire reservoir section due to complex variations in rock composition. Sandstone failure and the onset of sand production can be more accurately predicted by integrating log derived rock strength, calibrated to rock mechanics laboratory test data, with petrographic observations made on the core; i.e. combining quantitative and qualitative methodologies. Such observations typically include mineralogy, rock texture, structure and reservoir quality. Laboratory numerical data are discrete, whereas log and observational data are continuous, and these do not always readily relate. This can make the integration of laboratory-to-field data a difficult process. In an attempt to more closely relate quantitative and qualitative core data, a Schmidt impact testing hammer has been used to make hardness measurements along an entire core. The Schmidt hammer was originally designed for the non-destructive testing of concrete hardness in the civil engineering industry and was later adopted to estimate rock strength in mining, quarrying and tunnelling applications. The Schmidt hammer allows very rapid (and inexpensive) core testing, and has the added advantage of being non-destructive. Rock mechanics parameters can be determined in dense arrays that reflect the real inherent heterogeneity of the core. Use of the Schmidt hammer has enabled extrapolation of laboratory determined rock strength across the entire core and has facilitated direct correlation of mineralogy and grain texture to sandstone failure and yield. Characterisation of overlying claystones has also provided essential data for the determination of stress shadowing effects. Further development and refinement of this technique should provide an additional useful tool in reservoir and overburden characterisation and predicting sand production. The technique does not require specialist rock mechanics knowledge and can be applied at the well site or core viewing room. Useful data can be obtained from core material ranging from fresh state to old archived material. The speed and practicality of the Schmidt hammer as a rock strength characterisation tool is seen to have great potential. Introduction Sanding prediction work often relies on numerical analysis (quantitative) without regard to reservoir quality, rock texture and micro-structure from core observations and petrology studies (qualitative). The two methods of approach compliment each other, and when combined, greatly enhance strength characterisation of sedimentary rocks. In practice however, the integration of numerical and observational data is often fraught with difficulties and pitfalls. The unconfined (or uniaxial) compressive strength (UCS) is the most often quoted rock strength parameter. It is also the most readily understood sanding indicator by non-specialists in the field of rock mechanics. The UCS test has many disadvantages, but nevertheless has widespread and valuable application to wellbore stability and sand prediction. If UCS data are not available, many process steps are required to calculate an approximated UCS (Fig. 1). Laboratory UCS data can be plotted on a core log for the correlation of rock strength to petrological features. However, in heterogeneous reservoirs of complex sedimentology and mineralogy, the often-infrequent sample interval of UCS data makes accurate correlation and extrapolation difficult.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.