|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
The drilling conditions described above have led to the following practices, which are reasonably uniform, in the geothermal drilling industry. Bits Because of the hard, fractured formations, roller-cone bits with tungsten-carbide inserts are almost universally used for geothermal drilling. The abrasive rocks mean that bit life is usually low (50 to 100 m), but many bits are also pulled because of bearing failures caused by rough drilling and high temperature. Polycrystalline diamond compact (PDC) bits have the dual advantages of more efficient rock cutting and no moving parts, but experience with PDC bits in geothermal drilling is both scant and unfavorable. Much research and development in hard-rock PDC bits is under way,  so it is possible that these bits will come into wider use in geothermal drilling.
Grant, Malcom A. (New Zealand Dept. of Scientific and Industrial Research) | Bixley, Paul F. (New Zealand Ministry of Works and Development) | Donaldson, Ian G. (New Zealand Dept. of Scientific and Industrial Research)
Geothermal wells exhibit a variety of internal flow effects caused by the flow of water, steam, or both between distinct permeable zones tapped by the well. These internal flow effects are described and it is shown how they may be recognized from downhole pressure and temperature profiles.
Pressure transients measured at depths other than that of the well's primary permeable zone can be corrupted by such flows. The effects of such flows on injection and discharge transients are discussed.
Two types of flow can occur in wells in geothermal reservoirs: (1) interzonal flow, in which fluid enters the well at one depth, flows up or down the wellbore, and exits at a second depth, and (2) internal convection in which fluid circulates within the wellbore. The first is more common.
A geothermal wellbore is a long, vertical or nearvertical pipe penetrating a reservoir of heated fluid. Most geothermal reservoirs consist of fractured rocks, and a well draws its fluid supply from one or a few fractures (also called "permeable zones," "aquifers," "productive horizons," or "feedpoints"). Normally no attempt is made to isolate individual feedpoints from one another, so multiple feeds are often exposed to the well over a vertical distance of 3,000 to 6,000 ft (1000 to 2000 m).
Over the depth of open hole the well is exposed to the reservoir, which may contain waters of different temperatures or steam/water mixtures. In addition, the reservoir pressure distribution is not static because the natural throughflow of the reservoir causes a nonstatic distribution of fluid. In Wairakei, New Zealand, the preexploitation vertical pressure gradient was, for example, about 7% above hydrostatic (for the temperatures involved).
The very high permeabilities encountered in good geothermal wells [permeability-thickness of the order 3 to 300 darcy-ft (1 to 100 darcy.M)] mean that comparatively small pressure differences from buoyancy fects or from the nonstatic reservoir profile may cause substantial flows within a wellbore. If a well has more than one significant feedpoint, it is impossible for it to attain both thermal and pressure equilibrium with reservoir, and fluid will flow between the permeable zones. Such flow up or down the wellbore distorts temperature profiles, so measured downhole data reflect not reservoir temperatures but the physics of heat and mass transfer within the wellbore. Pressure profiles measured downhole likewise reflect not the reservoir pressure profile but the fluid column occupying wellbore.
In the absence of these strong interzonal flow effects, the fluid-filled wellbore still represents an effect means of vertical heat transport by convective circulation within it.
Summary. During the last decade, the use of numerical modeling for geothermal resource evaluation has grown significantly, and new modeling approaches have been developed. In this paper, we present a summary of the current status in numerical modeling of geothermal systems, emphasizing recent developments. Different modeling approaches are described and their applicability discussed. The various modeling tasks-including natural-state, exploitation, injection, multicomponent, and subsidence modeling-are illustrated with geothermal field examples.
A number of different methods for modeling the behavior of geothermal reservoirs are currently available to reservoir engineers. These methods vary widely in complexity and cost of application. In the selection of the proper method for a particular study, one must consider the amount and quality of field data available and the objectives of the study. Geothermal systems are generally very complex, exhibiting such features as fracture-dominated flow, phase change, chemical reactions, and thermal effects. Modeling studies must be carried out to analyze data from geothermal wells accurately and to estimate the generating potential of a system. When a model of a geothermal potential of a system. When a model of a geothermal system is developed, the existing field data must be carefully evaluated, and the important physical processes that occur in the system identified. After a plausible conceptual model of the field is developed, one must choose a mathematical (numerical) model that can realistically evaluate the performance of the geothermal reservoir and reliably predict its future behavior. We have found that modeling the natural state of a field before modeling the field under exploitation can give very valuable reservoir information. It not only tests the conceptual model qualitatively, but also gives estimates of mass and heat flow in the system. Furthermore, it provides consistent initial conditions for the exploitation provides consistent initial conditions for the exploitation models. The primary objective for geothermal reservoir modelinc, is to provide answers to important reservoir management questions relating to well decline, well spacing, generating capacity (power potential) of the reservoir, injection effects, and potential subsidence and scaling problems. These questions must be addressed by use of a problems. These questions must be addressed by use of a proper exploitation model that has evolved from the proper exploitation model that has evolved from the conceptual model and the natural-state modeling studies. In this paper, we present a brief review of geothermal reservoir modeling, emphasizing recent developments. The different modeling, approaches are described, and their advantages and limitations are discussed. We briefly describe the governing equations for mass and heat flow and discuss phase transitions and solution techniques. Examples illustrate the different methodologies for modeling of natural state, exploitation, injection, multicomponent flow, and subsidence. Finally, we identify problems of current interest in geothermal reservoir modeling. Earlier summaries of geothermal reservoir modeling are given by Witherspoon et al. and O'Sullivan.
Physical Processes and Physical Processes and Conceptual Models
In contrast to oil and gas reservoirs, geothermal systems are very dynamic in their natural state. There is continuous transport of fluid, heat, and chemical species. Important physical processes in geothermal systems include mass transport, convective and conductive heat transfer, phase change (boiling and condensation), dissolution and phase change (boiling and condensation), dissolution and precipitation of minerals, and stress change caused by precipitation of minerals, and stress change caused by pore-pressure changes. Most of these processes are pore-pressure changes. Most of these processes are strongly coupled; for example, phase change disturbs chemical equilibria, often resulting in precipitation/ dissolution of minerals that in time can alter porosities and permeabilities of the subsurface rocks. This in turn can permeabilities of the subsurface rocks. This in turn can affect the mass transport in the system. In modeling geothermal reservoirs, one must carefully evaluate which physical processes need to be considered in a specific modeling study. This will depend on the objectives of the study and the complexity of the geothermal system. Most currently available geothermal simulators consider only single-component mass and heat transport. In recent years, several simulators capable of modeling the transport of a second component, either a noncondensible gas or a dissolved solid, have been developed. Conceptual models of geothermal systems vary greatly in complexity. Perhaps the "simplest" geothermal systems are those created by hot water upflow through a single fault or at the intersection of two or more faults (e.g., Susanville and East Mesa, CA. ) A rather complex porous-medium-type geothermal reservoir is the Cerro Prieto field, Mexico (Fig. 1). The lithology consists of interlayered shale and sandstone beds. The detailed lithology shown in Fig. 1 has been determined mainly on the basis of wireline well logs.
Numerical studies were performed to investigate the effects of localized feed zones on the pressure transients in two-phase reservoirs. Gravity effects were shown to affect the pressure transients significantly because of the large difference in the densities of liquid water and vapor. Production from such systems enhances steam/liquid-water counterflow and expands the vapor-dominated zone at the top of the reservoir. Subcooled liquid regions develop in the center of the reservoir as a result of gravity drainage of cooler liquid water. The vapor zone acts as a constant-pressure boundary and helps to stabilize the pressure decline in the system. The pressure transients at observation wells were shown to depend greatly on the location (depth) of the major feed zone; if this is not accounted for, large errors in deduced reservoir properties will result. At shallow observation points, pressures may actually increase as a result of enhanced steam upflow caused by production at a deep feed zone.
In two-phase geothermal reservoirs, heat is transported from the reservoir bottom to the caprock through counterflow of liquid and steam, often referred to as the heat-pipe effect. Because the heat content per unit mass of steam is much higher than that of liquid water at the same temperature, a mass-balanced counterflow of the two phases results in heat transport to the caprock from below. A given heat flow can cause two different thermodynamic conditions to exist in the reservoir: one where the vapor phase is nearly immobile and the pressure gradient is slightly less than hydrostatic, and another where the liquid phase is nearly immobile and the pressure gradient slightly exceeds vapor static conditions. In both cases, the product of the pressure gradient and the fluid mobility will be equal for the two phases, resulting in a mass-balanced counterflow. In this paper, a reservoir with a liquid-dominated heat pipe is considered; i.e., the initial pressure gradient in the reservoir is nearly hydrostatic.
Most geothermal reservoirs are located in fractured rocks, with fluids entering the wellbore at points where it intersects major faults and fractures commonly called feed zones. Most geothermal wells have one or two feed zones, the locations of which are inferred from drilling data (lost circulation, drilling rates), pressure and temperature profiles during heating of the well after drilling, and spinner tests during cold-water injection tests (most spinners are not operational at the high temperatures encountered in many geothermal wells). The pressure transients for systems with localized feed zones have received little attention, with the exception of studies addressing partially penetrating wells. The available results for partially penetrating wells are not readily applicable to problems involving fractured, two-phase reservoirs because it is assumed that the open well interval is at the top of the reservoir, and important gravity effects associated with two-phase reservoirs are therefore neglected.
Production results in pressure declines around well feed zones, which increase the vapor saturation. The increase in vapor saturation disturbs the stable counterflow mechanism. This enhances steam upflow and causes a steam zone to develop at the top of the reservoir beneath the caprock. Furthermore, gravity effects cause liquid downflow from shallow reservoir regions to the depth of the feed zone. These effects combine to produce complex pressure transients that will yield erroneous results when conventional analysis methods are used.
Studies have been conducted on pressure transients in two-phase reservoirs, generally neglecting gravity effects and localized feed zones. Various investigators have extended the single-layer pressure-transient theory to two-phase systems. These investigators incorporated the effects of the fluid enthalpy in their methods of analysis, but rigorous analysis is still not possible because of a lack of knowledge regarding relative permeabilities. Moench and Moench and Atkinson investigated pressure transients in two-phase fractured reservoirs with immobile liquid water. Cox and Bodvarsson investigated the effects of localized two-phase zones on pressure-transient data. They included gravity effects in some of the cases and illustrated that these can have large effects on pressure transients.
The main objective of this paper is to investigate the importance of gravity effects and localized fluid production (partial penetration) on the pressure behavior of two-phase reservoirs. The pressure transients at production wells are investigated, as well as the pressure behavior of observation wells. Of particular interest are the pressure transients at observation wells with feed zones at different depths. These data can help in the evaluation of steam/liquid-water counterflow and horizontal and vertical permeabilities. The global changes in thermodynamic conditions of two-phase reservoirs during localized production are also investigated to understand the long term behavior of such systems.
Fig. 1 shows the reservoir system considered. A single well penetrates a two-phase reservoir. Production from the well is assumed to be from either the top or bottom of the reservoir. The reservoir is 500 m [1,640 ft] thick, and the production interval is assumed to be 50 m [164 ft] thick. A 10-layer grid is used, each layer being 50 m [164 ft] thick. A constant mass flow rate of 15 kg/s [119,000 lbm/hr] is specified for the well.
Initially, two-phase conditions prevail everywhere in the reservoir system. Two-phase conditions with nearly uniform vapor saturation (Sv appx. 0.05) were achieved by maintaining an appropriate heat flow through the system. Constant heat flux is applied at the bottom, and the energy is transferred to the top of the reservoir by liquid/vapor counterflow. A constant heat sink is specified at the top of the reservoir, representing conductive heat losses. The heat flux used was 0.4 W/m2 [0.127 Btu/ft2-hr], which results in a vapor/liquid counterflow of about 2.4 x 10-7 kg/s.m2 [0.02 lbm/hr-ft2]. The initial pressure is practically hydrostatic with depth, and the initial temperature in the top and bottom layers is 245 and 287 deg. C [473 and 549 deg. F], respectively.
A porous-medium model is used in this work because it appears reasonable to attempt to understand porous-medium behavior before tackling the more complex case of a fractured reservoir. The porosity and horizontal permeability in the system are assumed to be 5% and 50 md, respectively; the vertical permeability is varied in the simulations. Linear relative permeabilities are used, with an immobile liquid saturation of 0.40 and immobile vapor saturation of 0.05.
The numerical simulator MULKOM was used in this work. The simulator has been validated against analytical solutions and the behavior of many geothermal fields. A description of the code was given by Pruess.
Production From Deep Feed Zones
A number of cases are simulated with fluids produced from the bottom 50 m [164 ft] of the reservoir.
Numerical simulation techniques are used to study the effects of noncondensable gases (CO2) on geothermal reservoir behavior in the natural state and during exploitation. It is shown that the presence Of CO2 has a large effect on the thermodynamic conditions of a reservoir in the natural state, especially on temperature distributions and phase compositions. The gas will expand two-phase zones phase compositions. The gas will expand two-phase zones and increase gas saturations to enable flow of CO2 through the system. During exploitation, the early pressure drop primarily results from "degassing" of the system. This primarily results from "degassing" of the system. This process can cause a very rapid initial pressure drop, on process can cause a very rapid initial pressure drop, on the order of megapascals, depending on the initial partial pressure of CO2. The flowing gas content from wells can pressure of CO2. The flowing gas content from wells can provide information on in-place gas saturations and provide information on in-place gas saturations and relative permeability curves that apply at a given geothermal resource. Site-specific studies are made for the gas-rich, two-phase reservoir at the Ohaaki geothermal field in New Zealand. A simple lumped-parameter model and a vertical column model are applied to the field data. The results obtained agree well with the natural thermodynamic state of the Ohaaki field (pressure and temperature profiles) and a partial pressure of 1.5 to 2.5 MPa [217 to 363 psi] is calculated in the primary reservoirs. The models also agree reasonably well with field data obtained during exploitation of the field. The treatment of thermophysical properties of H2O/CO2 mixtures for different phase compositions is summarized.
Many geothermal reservoirs contain large amounts of non-condensable gases, particularly CO2. The proportion of noncondensable gas in the produced fluid is an extremely important factor in the design of separators, turbines, heat exchangers, and other surface equipment. In the reservoir itself, the presence of CO2 significantly alters the distribution of temperature and gas saturation (volumetric fraction of gas phase) associated with given heat and mass flows. Therefore, when modeling gas-rich reservoirs it is essential to keep track of the amount of CO2 in each gridblock in addition to the customary fluid and heat content. Several investigators have considered the effects of CO2 on the reservoir dynamics of geothermal systems. A lumped-parameter model using one block for the gas zone and one for the liquid zone was developed by Atkinson et al. for the Bagnore (Italy) reservoir. Preliminary work on the Ohaaki reservoir was carried out by Zyvoloski and O'Sullivan, but these studies were limited because-the thermodynamic package used could only handle two-phase conditions. Generic studies of reservoir depletion and well-test analysis also were made in the previous works. The present study describes the effects of CO2 in geothermal reservoirs in a more complete and detailed way. We emphasize the potential for using the CO2 content in the fluid produced during a well test as a reservoir diagnostic aid, and as a means of gaining information about relative permeability curves.
The aim of the present study is to investigate the effects of CO2 on both the natural state of a reservoir and its behavior under exploitation. Several generic simulation studies are described. First, the effect of CO2 on the depletion of a single-block, lumped-parameter reservoir model is briefly examined. Secondly, the relationship between the mass fraction Of CO2 in the produced fluid and the mass fraction in place in the reservoir is studied. It is demonstrated that in some cases the in-place gas saturation can be determined for a given set of relative permeability curves. Finally, the effects of CO2 on the permeability curves. Finally, the effects of CO2 on the vertical distribution of gas saturation, temperature, and pressure of geothermal reservoirs in the natural state are pressure of geothermal reservoirs in the natural state are investigated. The numerical simulator with the H2O/CO2 thermodynamic package is applied to field data from the Ohaaki (formerly Broadlands) geothermal field in New Zealand.
Two simple models of the 1966-74 large-scale field exploitation test of the Ohaaki reservoir are presented. The first is a single-block, lumped-parameter model similar to those reported earlier by Zyvoloski and O'Sullivan and Grant. In the former work, a less accurate thermodynamic package for H2O/CO2 mixtures is used; the latter uses approximate methods to integrate the mass-, energy-, and CO2-balance equations. The second model described in the present work is a distributed-parameter model, in the form of a vertical column representing the main upflow zone at Ohaaki. This model produces a good fit to the observed distribution of pressure and temperature with depth in the natural state at Ohaaki and a good match to the observed response of the reservoir during 5 years of experimental production and 3 years of recovery.