- Introduction
- Vertical Wells
- Directionally Drilled Wells
- Application of Directionally Drilled Wells
- Common Types of Directionally Drilled Wells
- Directional Well Plan
- Directional Tools Used for Measurements
- Directional Survey Calculations
- Directional Survey Uncertainties
- Directional Well Plots
- Wells Without Directional Surveys
This chapter is from the book
This chapter is from the book
This chapter is from the book
Vertical Wells
Early in the development of the oil industry, all wells were drilled vertically because vertical wells are easier, cheaper, and safer to drill. That situation began to change in the late 1920s as operators struggled to develop oil fields that straddled the California coastline (LeRoy et al. 1977, Chilingarian and Vorabutr 1981). As recently as 1990, about 80% of all wells being drilled in the United States were vertical, and the remaining 20% were evenly split between deviated wells and horizontal wells (Baker Hughes Weekly Rig Count as of July 2018). By the middle of 2018, the picture had changed completely with almost 90% of the wells being drilled in the United States being horizontal and the remaining 10% being evenly split between vertical and directional wells (Baker Hughes Weekly Rig Count as of July 2018).
Determining the X, Y, and Z values for a vertical well is relatively straightforward. The X and Y values are taken from the surveyed location of the rig drilling the well. The Z is determined by direct measurement in the well. The situation for deviated and horizontal wells is more complicated and is discussed in more detail later in this chapter.
Surface Location Uncertainty
Although determining the X and Y location of a well is relatively straightforward, the available data is not always accurate. A recent attempt to locate several hundred wells in southwestern Wyoming revealed that about a quarter of them were mislocated in the state’s database by up to one mile (Joyce 2016). A number of studies encompassing 12,000 wells found that 40% of the surface locations were off by over 100 ft (Stigant 2012). There are numerous reasons why the surface location of wells can be incorrect in a database. In the early days of petroleum exploration, well locations were not always recorded or were recorded only very generally (Frischkneckt et al. 1985). Older survey methods (pre-1990) were not as accurate as modern methods using GPS satellites, especially in frontier areas and in difficult terrain such as the swamps of southern Louisiana. During the 1990s, the U.S. government deliberately degraded civilian accuracy of GPS data, but this practice was ended in May 2000. Current commercial GPS surveys are usually accurate within 1 ft or less.
Positional uncertainty of the surface location of carefully surveyed older wells is frequently ± 66 to 164 ft (± 20 to 50 m) (Williamson 1999). Even when the surface location of a well is accurately surveyed, the location may be incorrectly labeled on the drilling permit, scout ticket, or well header. Accurate surface locations may be entered incorrectly into computer databases. Finally, the surface location may be reported in the wrong geodetic datum and ellipsoid or the wrong map projection, causing the well to appear in the wrong location on a map (Williamson 1999, Jamieson 2012). A single point on the earth’s surface will have different latitude and longitude depending on the geodetic datum and ellipsoid referenced. Taking latitude and longitude from one datum and using it incorrectly with a different datum can produce hundreds of feet of error in the surface location (Stigant 2012, Jamieson 2012).
Jamieson (2012) lists several key facts pertinent to surface locations, including that latitude and longitude, elevations/heights, units, and headings are not unique unless qualified with a datum or reference; that every time data is handled or transferred, there is a chance that the references are not properly interpreted; and that most databases have incomplete metadata describing the references. Most geoscientists do not understand geodesy, the science of positioning on the earth’s surface. Unless you understand geodesy, you should always consult someone who does when setting up a mapping project (Jamieson 2012).
In many areas with a long history of drilling, it may be necessary to confirm the surface location of old wells. In some cases, it may be possible to do this from geographically registered air photos or even satellite images (Barnes 2005). In other cases, it may be necessary to use airborne magnetometers carried by either low-flying planes or helicopters to locate surface locations from the magnetic signatures of casing cut off below ground level as part of the abandonment process (Frischkneckt et al. 1985, Hammack et al. 2016). In difficult cases, it may be necessary to check surface locations on foot. In a few cases, it may not be possible to locate or confirm the location of old wells, especially if the well’s casing was cut off below the surface and pulled, leaving no equipment or surface magnetic anomaly to find.
Even when the surface location of a well is accurate, the data from that well can be assigned to the wrong wellbore. For example, data from a sidetrack may be incorrectly stored in a database as being from the original hole or from a different sidetrack. Log data can also be incorrectly stored under a different well with a similar name (Storey 2014). When data does not fit the map, it is important to go back and check that the data is assigned to the correct well at the right surface location.
Well Depth
Depth in a well is usually measured either with drill pipe, called driller’s depth, or with wireline, called logger’s depth. The reference elevation of the well has traditionally been the kelly bushing, a large piece of the drilling rig. Other common reference elevations are the rig floor (RF) or derrick floor (DF), the rotary table (RT), the bradenhead flange (BHF), and ground level (GL). In some offshore wells, a well may be started with one rig and finished with another rig. The same reference point on the different rigs may be at significantly different elevations relative to mean sea level (MSL). Thus, the zero depth for a shallow log run may be at a different elevation than the zero depth for a deeper log run.
We begin by discussing measured depth (MD), also known as along hole depth (e.g., Brooks et al. 2005) or true along hole depth (TAH) (Forsyth et al. 2013). For vertical wells MD is equal to TVD (which we discuss later in the chapter). In deviated wells, MD is not equal to TVD and must be corrected for the directional path of the well. Before we deal with the directional path of the well, however, it is important to understand the uncertainty in MD, as TVD can never be known more accurately than we know MD.
Depth is the most fundamental measurement made in a well, and yet it is a measurement for which there are few, if any, industry standards or accepted auditable practices (Forsyth et al. 2013, Bolt 2015). Bolt (2015) recently listed some of the costs of depth uncertainties (Table 3-1). These costs range from minor, thousands or tens of thousands of dollars, to major, tens of millions or hundreds of millions of dollars. Depth uncertainty is one of the largest uncertainties in assessing the value of oil and gas fields (Fig. 3-1), but it is rarely considered. It is important to get depth values as accurately as possible.
)
Figure 3-1 Depth is one of the key uncertainties in determining the value of oil and gas fields. (After Bolt 2015. Published with permission from Harald Bolt.)
Table 3-1 Potential cost of depth uncertainties (after Bolt 2015)
Event |
Cost ($) |
|---|---|
Lost time repeat surveys |
Thousands |
Interval relogging |
Tens of thousands |
Field depth resurveys |
Hundreds of thousands |
Field studies and sidetracks |
Millions |
Equity value discrepancies |
Tens of millions |
Incorrect reserves and geological models |
Hundreds of millions |
The largest investment decision that many geoscientists will ever work on is the development plan for a new discovery. These decisions can involve billions of dollars of investment capital in frontier discoveries, for example. Suboptimal development plans can result in the unnecessary spending of tens or hundreds of millions of dollars by installing larger facilities or drilling more wells than needed to develop a field. Alternatively, fields may be undercapitalized, requiring retrofitting or additional facilities at a much higher cost than if the initial field development plan was more appropriate. Worst of all would be the failure to develop a profitable discovery because the cost of the development was inaccurately assessed. Failure to accurately recognize field complexity and compartmentalization due to a lack of understanding of depth uncertainty associated with fluid contacts, for example, can contribute to very expensive business mistakes (Forsyth et al. 2013; Bolt 2015).
All structure maps are created relative to a horizontal reference plane, usually MSL. The elevation of MSL is constant over short distances but it is not constant over large distances. MSL in the United States and Canada used to be the National Geodetic Vertical Datum of 1929 (NGVD 29). NGVD 29 assumed that 26 tidal gages in the United States and Canada were the same elevation. More accurate surveying has revealed that these 26 gages actually differ in elevation by up to 5 ft (1.5 m) (FEMA 2014). NGVD 29 was supplanted with a new datum in the early 1990s, the North American Vertical Datum of 1988 (NAVD 88). Elevations relative to MSL for pre-1991 wells in the United States and Canada are likely to be relative to the NGVD 29 datum, whereas elevations relative to MSL for wells drilled after 1991 are likely to be referenced to NAVD 88. Although reference elevation differences for nearby wells referenced to the different datums are likely to be very small, it is possible to check elevations relative to the two different standards using a program (VERTCON) on the U.S. government’s National Oceanic and Atmospheric Administration (NOAA) website. Different areas of the world have different datums for MSL, but these differences are unlikely to affect any local mapping project.
Driller’s depth is measured with joints of drill pipe, which are generally measured on a pipe rack. Traditionally, joints of drill pipe have been measured from pipe shoulder to pipe shoulder using steel measuring tape under ambient surface conditions, although in recent years they are frequently measured using lasers (Jamieson, 2012). A joint of pipe will have a slightly different length when measured in North Dakota in the winter than if the same joint of pipe were measured in the Permian Basin in August. Driller’s depth is determined by summing the length of every joint of pipe in the drill string along with any bottomhole assembly (BHA) that is in the hole below the reference elevation of the well. The reference for the length of every joint of pipe run in the drill string is known as the driller’s tally. Traditionally, mudlogs, casing points, and total depth (TD) of a well have been based on driller’s depth.
Since Schlumberger introduced the resistivity log in France in 1927, most well depth measurements that are used in subsurface mapping are based on various types of wireline logs. These logs are generally measured with cable or wireline depths. Recently, more and more logs have been acquired using logging while drilling (LWD). LWD logs are measured, somewhat indirectly, using driller’s depth. Actual log readings are referenced to time. The depth of the bit is also referenced to time. Cross-referencing the time of a log reading and the depth of the bit at that time, along with knowledge of the position of tool sensors relative to the bit, allows the depth of log measurements to be determined.
In some situations, pipe-conveyed logs (PCL), also referred to as tough logging conditions (TLC or TCL) logs, are acquired using pipe-conveyed wireline tools. Some logs are also obtained on slickline, but these logs are almost invariably obtained within cased holes. Their depth is usually tied to an open hole log and are not considered further in this book. It is important that geoscientists and engineers making and working with subsurface maps understand that depth measurements of different types of logs, especially wireline and LWD logs, are made with different measuring tools and frequently give different depths for the same subsurface marker (Fig. 3-2). To produce accurate subsurface maps, it is important that mappers understand how depth is measured for different types of logs, how and why these depths may differ, and how these differences can be minimized.
)
Figure 3-2 Wireline and LWD logs from a single Gulf of Mexico well showing a hydrocarbon-bearing sand that is at different depths on the two different logs. The 20-ft depth difference occurs because wireline depth and LWD depth are measured using different tools. (Published with permission of J. Brenneke.)
Wireline Depth Measurements
Since the 1930s, most formation depths in wells have been measured using wireline tools. Wireline depths are measured by determining the position of various logging tools on the logging sonde and measuring the length of wireline cable lowered into the hole. In most cases the length of cable is measured using some type of wheel device. A number of factors can affect the measurement of cable length, but in modern wells, these factors are relatively minor and are not discussed in detail in this book. Interested readers are referred to extensive discussions of these factors in publications such as Bolt (2015).
The main factors affecting wireline length changes in the hole are cable stretch and thermal expansion/contraction. Most service companies have relatively accurate algorithms that allow depth correction for cable stretch (Bolt 2015, 2016), although these algorithms continue to be improved (e.g., Fitzgerald and Pedersen 2007). Wellbore deviation affects the stretch of wireline cables as friction on the borehole wall in deviated wells decreases cable tension (Fitzgerald and Pedersen 2007). Older algorithms are probably less accurate than more recent ones. The thermal expansion/contraction of modern logging cables is complicated, as they are composed of several different materials that react differently to temperature. The reaction to temperature may vary from one cable type to another. Some workers report that increased temperatures lengthen cables, whereas other workers report that increased temperatures shorten cables (Bolt 2015).
Wireline data may be measured either while running into the hole or when pulling out of the hole. Different companies prefer one method or the other, but the depths from the two methods are different (Pedersen and Constable 2006; Fitzgerald and Pedersen 2007). Most final logs are recorded while pulling out of the hole, although the depth on these logs may be adjusted to the depth recorded running into the hole.
LWD Depth Measurements
LWD depth measurements are based on driller’s depths, which in turn are based on the length of drill pipe and BHA in the borehole at any given time. The driller’s depth is based on three factors. First, the length of pipe and BHA is measured on the surface using steel tape measures or, increasingly, lasers. Second, the length of pipe and BHA is recorded in the driller’s tally book. Third, the length of drill pipe that is in the drill string but not yet in the hole is determined by the height of the traveling block or top drive above the reference location of the well.
LWD data is not really measured relative to depth. It is measured relative to time. The depth of the LWD data is then determined from the logger’s depth of the drill bit at the time the data was recorded and the distance of the LWD sensors from the drill bit.
The length of the drill pipe is measured under ambient surface conditions and no stress. Once in the hole, the drill pipe is subjected to forces different than ambient surface conditions and no stress. The two most significant forces with respect to drill pipe length are the tensional stress due to the weight of the drill string and thermal stress related to the in situ temperature of the pipe in the hole. The temperature experienced by the drill pipe is not the static temperature of the earth but the actual temperature in the borehole, which is affected by mud circulation and frictional heating from the drill bit. Both tensional stress and thermal stress cause the drill pipe to lengthen, and these forces tend to increase as a function of TVD. Lesser forces that affect the length of the drill pipe are the mud weight, which partially offsets the weight of the drill pipe; friction along the borehole wall, weight on bit (WOB), and borehole geometry, which also affect the tension on the drill string and thus its stretch; pump pressure, which can cause the drill pipe to balloon slightly, shortening the pipe; and torque, which can also shorten the pipe (Chia et al. 2006).
The length of each section of drill pipe is recorded in the driller’s tally book. Occasionally, a joint or stand is left out of the tally book due to human error. Also occasionally, one joint may be switched out for another of a slightly different length, perhaps because the first joint’s threads were scored or some other imperfection in the joint was recognized. If the length of the first joint is already recorded in the tally book, it might not be replaced with the length of the second joint. The final accuracy of driller’s depth cannot exceed the accuracy of the driller’s tally book.
Once a well is drilled as deeply as can be done with a given amount of drill pipe, the drill pipe is set in the slips at the rig floor and a new joint or joints of drill pipe is attached to the existing drill string. The top drive or traveling block is then attached to the top of the new joint or joints of drill pipe and drilling is resumed. The distance the top drive or traveling block moves down from its position in the top of the derrick after drilling resumes is used to determine the depth of the bit in the well.
As a well is drilled, the main forces acting on the drill pipe usually result in elongating the drill string. This elongation results in events logged in the well to be recorded at shallower depths than they really occur (Fig. 3-3).
)
Figure 3-3 Diagram explaining why LWD logs are almost invariably shallower than wireline logs in deeper wells. The length of the drill pipe is measured under ambient conditions at the surface. Tensional stress and thermal stress result in pipe stretch, which is generally not corrected for on LWD logs. (Published with permission of J. Brenneke.)
Pipe-Conveyed Log Depth Measurement
Pipe-conveyed logging (PCL) is usually performed when wireline logs are required but the conditions of the wellbore do not allow the wireline sonde to go down in the wellbore on its own. The wireline sonde is attached to the end of the drill pipe, and the wireline cable is run outside of the drill pipe through a side entry sub. The wireline operator plays out cable as the drill pipe is lowered into the hole. Prior to logging the interval in question, the PCL log is depth tied to an existing log, which may be either a wireline log or an LWD log. During the logging operation, the wireline operator attempts to synchronize the movement of the cable with the movement of the drill pipe. This synchronization is frequently not perfect, resulting in depth errors that are difficult to quantify (Wilson et al. 2004). LWD logs are also usually acquired with some WOB, whereas PCL logs are acquired with no WOB, resulting in some difference in driller’s depth between the two logs (Wilson et al. 2004). PCL logs are relatively uncommon and are not considered further in this book.
Relative Depth Uncertainty
Relative depth uncertainty is depth uncertainty within a single well. Both wireline and LWD logs generally have good relative depth uncertainty, so the distance between markers within a single well is usually relatively accurate. On some occasions, however, LWD relative depth uncertainty becomes less accurate (Fig. 3-4) (Storey 2013). The example in Figure 3-4 is somewhat extreme, but smaller variations are not uncommon (e.g., Forsyth et al. 2013, Fig. 5). The cause of these relative depth uncertainties in the LWD logs is not addressed by Storey (2013) or Forsyth et al. (2013) but could be related to poor calibration of the hanging block in the middle of a stand (R. Wylie, personal communication, 2018) or by variations in WOB. The relative depth uncertainty in the LWD data can be seen when both LWD and wireline logs are available but may go unrecognized if only LWD depth is available.
)
Figure 3-4 Relative depth uncertainty between an LWD log and a wireline log. The two gamma ray (GR) logs are on depth near both the top and the bottom of this section, but they are off depth by several meters near the center of the interval. The interval thickness between the D3 and D4 markers is 32.6 m (107 ft) based on the wireline log and 40.3 m (132 ft) based on the LWD log. This thickness difference, approximately 25%, would have a significant impact on volumetrics. (From Storey 2013. Published with permission of Oilfield Technology.)
It is important to check the WOB to minimize the possibility of relative depth uncertainty in an LWD log. If the WOB decreases while the drill pipe is rotating and the hanging block is stationary, the drill bit is probably drilling a new hole, and the LWD sensors are recording values from the new section. These readings may not be displayed because LWD logs generally display only the first readings from a given depth, and if the hanging block is not moving, the depth is assumed to be constant. Similarly, if the WOB is increasing and the hanging block is moving, the bit may not be moving as much as the hanging block. In extreme cases, if the drill string goes from extension into compression, the drill pipe could buckle or corkscrew, resulting in significant movement of the hanging block with no movement of the drill bit (Fig. 3-5).
)
Figure 3-5 Drill pipe in compression can produce sinusoidal buckling (Buckling-S) or helical buckling (Buckling-H), which can cause driller’s depth to be greater than TAH. In a deviated well, some parts of the drill string may be in compression while other parts are in tension. (From Jamieson 2012. Image courtesy of Pegasus Vertex, Inc.)
Absolute Depth Uncertainty
Although the relative depth uncertainty within a single well is generally small, the uncertainty in the absolute depth of any well is greater (Fig. 3-2). This uncertainty produces a depth uncertainty between wells. Because drill pipe stretch is rarely corrected for in LWD logs, LWD logs must almost invariably be shifted deeper to match wireline logs (Fig. 3-6).
)
Figure 3-6 (a) LWD depth shift required to match wireline log depth for deep, deviated wells in the Gulf of Mexico. Dot size is proportional to maximum well deviation that ranged from 5.38 deg to 48.47 deg. (b) LWD depth shift required to match wireline log depth for deep, near-vertical wells in the Gulf of Mexico. Dot size is proportional to hole diameter ranging from 5.75 in. to 14.75 in. (From Pineda and Bergeron 2017. Published with permission of Society of Petrophysicists and Well Log Analysts.)
When both wireline and LWD logs are run in a well, the LWD logs can be depth shifted to tie the wireline log. This depth shift is not constant, however, but varies with depth, and not always in a constant direction (Fig. 3-7). The real problem arises when only LWD logs are run in a well and it is compared to wells with wireline logs. There are no correction tables or rules of thumb that allow LWD logs to be depth shifted to mimic wireline depths, but in wells deeper than 15,000 ft, depth shifts of 20 ft to 50 ft are not uncommon.
)
Figure 3-7 Depth shift required to tie an LWD log to a wireline log in a deviated well from Kristin Field, Norwegian Sea. The required log shift reaches a maximum of 19.5 m (64 ft) between 5100 m and 5200 m MD (16,730 ft to 16,060 ft). This shift decreases to just over 14 m (46 ft) at TD. (From Pedersen and Constable 2006. Published with permission of Society of Petrophysicists and Well Log Analysts.)
LWD logs can be depth shifted to correct for drill pipe stretch if sufficient drilling data is available (Chia et al. 2006). These corrections allow the length of each joint of drill pipe to be corrected for mechanical stretch and thermal expansion and can dramatically reduce the depth differences between LWD and wireline logs (Fig. 3-8). The required drilling data are almost always available to the service company running the log but may not be available to other companies who wish to use the log in mapping.
)
Figure 3-8 (a) LWD and wireline GR logs from a deep, highly deviated well. The left side shows the curves differ in depth by 11 m (36 ft) as supplied in final form by the service company. The right side of the figure shows the depth agreement between the two logs after both the LWD and wireline depths have been corrected. (b) LWD and wireline GR logs from a well in a different field with a simpler geometric profile. Again, the logs on the left side of the figure are the final results supplied by the service company. Here the depth difference is 5.3 m (17 ft). The logs on the right side of the figure have been depth corrected. Here the depth correction to the wireline log is very small, but the depth shift to the LWD log is substantial. (From Chia et al. 2006. Published with permission of Society of Petroleum Engineers.)