Keeping a Water Well Vertical: Deviation Causes, Anti-Deviation Assemblies, and Correction Procedures
Nov 20,2026
A crooked well is decided in the first tens of meters, not at pump installation. Four causes cover most drift - dipping beds, an uneven bottom, excess weight on bit, a soft string - and the counters are known: gauge-clearance stabilizers, a stiff or pendulum assembly, a stable bit crown with balanced cutters, surveys every 30-50 meters, and a fixed rule for when correction beats re-drilling.
Why Verticality Is the Hard Acceptance Gate
The numbers acceptance uses
Verticality has moved from a preference to a pass-fail line, and the shift is visible in the data: a 2025 Water Systems Council survey found 83 percent of practitioners in the western United States rating compliance as a significant factor in well work. The usual contractual numbers are simple. Total drift is capped at 1-2 meters per 100 meters of depth, and pump wells are often held to the tighter 1-meter figure because a submersible needs a straight section to hang in. Dogleg severity, the rate of angle change rather than the offset itself, is typically limited to 1 degree per 30 meters, because a smooth hole passes casing where a crooked one with the same total drift will not. Crews should read these limits from the contract before the rig moves in, not from the rejection letter afterward; the three numbers above are the ones every bid document should state.
What 4.2 meters costs
The scenario behind this article is a common one. A 200-meter water well reaches total depth on schedule; when the pump string goes in, the casing stops 4.2 meters short of bottom. A survey shows the bottom hole is displaced 4.2 meters from the collar - 2.1 meters per 100 meters, double the pump-well limit. The pump hangs above the target aquifer and yield drops; the gravel pack cannot be placed uniformly around a crooked casing; and the acceptance clause is triggered, so the well is filled and re-drilled. The direct cost of a re-drill at this depth - rig hours at a typical 220 USD per hour, cement, casing, and a second bit - lands in the 25,000-40,000 USD band and adds 14-21 days the schedule did not hold. The table below collects the limits and what each failure actually blocks.
Item | Typical limit | What exceeding it blocks |
|---|---|---|
Total drift | 1-2 m per 100 m of depth | casing to bottom, acceptance sign-off |
Pump-well drift | 1 m per 100 m | pump hanging depth, rated yield |
Dogleg severity | 1 deg per 30 m | casing passage, gravel pack uniformity |
Survey record | one reading per 30-50 m | traceable acceptance data |
This article scenario | 4.2 m over 200 m = 2.1 m/100 m | fill and re-drill, 25,000-40,000 USD |
Table 1. Verticality limits in common contract use and the item each exceeded limit blocks.
Four Causes of Deviation
Deviation is not random; four causes cover most wells. The first two are geological. When beds dip steeper than about 10-15 degrees, the anisotropy of the formation - it drills easier along the bedding than across it - pushes the bit toward the updip side on every revolution, and alternating hard-soft beds amplify the effect because each interface re-launches the bit with a small side force. The second is an uneven bottom: after a trip, if cuttings are not cleared and a ridge or rock island is left standing, the bit starts on a slope and the first meters set the direction. Both causes are visible before they cost anything - updip drift on the survey plot, and torque spikes on re-entry to an uncleared bottom - and both are cheaper to prevent with circulation and cleanup than to correct later.
Buckling and assembly stiffness
The other two causes live in the string. Excessive weight on bit buckles the compression section of the drill collars, and a buckled collar tilts the bit off-axis with a side force of its own; the working rule is to keep weight on bit below roughly 70-80 percent of the buoyed weight of the collars in compression. Stiffness is the fourth cause, and it is a sizing error: the bending stiffness of a collar grows with the fourth power of its diameter, so a 127 mm collar is about four times as stiff as an 89 mm one, and a string built from small collars flexes its way out of vertical under the same weight. The four causes and their field signatures are collected in the table below.
Cause | Mechanism | Field signature | First response |
|---|---|---|---|
Formation dip and anisotropy | side force toward updip per revolution, dip over 10-15 deg | steady drift in one direction on the survey plot | stiffer assembly, stable-crown bit |
Uneven bottom | bit starts on a rock ridge or island after a trip | torque spike on re-entry, drift from the casing shoe | circulate and clear bottom before drilling |
Excess weight on bit | collar buckling tilts the bit off-axis | weight above 70-80% of compressed string weight | reduce weight, add collar area |
Low assembly stiffness | small-diameter collars flex under load | drift grows with depth despite stable parameters | larger collars, closer stabilizer spacing |
Table 2. The four deviation causes, the mechanism of each, its field signature, and the first response.
Anti-Deviation Bottom Hole Assemblies
Packed hole versus pendulum
Three assemblies cover the anti-deviation toolbox. The packed-hole assembly is the default for dipping formations: a near-bit stabilizer running close to full gauge - clearance to the hole of 3 mm or less - followed by one or two string stabilizers 6-9 meters apart, makes the lower string a rigid beam that follows the hole already drilled. The pendulum assembly is the corrective one: the nearest stabilizer is placed 18-27 meters above the bit, leaving the collars between hanging free, so gravity acts on the unsupported length as a pendulum force that pulls the bit back toward vertical; it costs penetration rate, so crews run it with reduced weight and raised rotary speed. A single near-bit stabilizer on a short stiff string suits shallow sections where full packed-hole friction is not wanted. Choice is by formation dip, hole size, and how much deviation is already in the hole.
The bit end of the string
The bit is the first stabilizer in the string, and its geometry decides how much side force the formation can inject. A shallow-cone or short-parabolic crown resists steering - the formation has to push a wide, flat face to move it - while a deeper cone cuts aggressively and follows side forces just as readily, which is why deep-cone profiles sit in steerable assemblies and straight-hole runs take the flatter crowns. Gauge length does the tracking: 2-3 rings of gauge cutters, roughly 80-120 mm of cylindrical contact, keep the bit in the cylinder the previous run cut. In formations that shrink back onto the string, back-reaming cutters on the gauge ream the hole on the way out instead of letting a tight spot lever the assembly sideways. Cutter layout matters too: we balance the spiral layout at design so residual lateral force stays under 5-10 percent, and selection against the formation band - the November 3 chipping article covers that match - is part of the same stability question.
Assembly | Layout | Best interval | Watch-outs |
|---|---|---|---|
Near-bit stabilizer | one full-gauge stabilizer, hole clearance 3 mm or less | shallow sections, mild dip | alone it will not correct existing drift |
Packed hole | near-bit plus 1-2 string stabilizers, 6-9 m spacing | dipping or alternating beds | higher torque, needs a gauge hole |
Pendulum | nearest stabilizer 18-27 m above the bit | holes already drifting | lower weight, higher rpm, slower ROP |
Reduced-weight run | parameters only, no hardware change | crossing soft-hard interfaces | watch ROP and torque for interface signals |
Table 3. The three anti-deviation assemblies plus the parameter-only option, with layout, best interval, and watch-outs.
Parameter Control and Deviation Surveys
Weight below the buckling limit
Parameters are the cheapest anti-deviation hardware. Weight on bit held below the buckling limit removes the largest self-inflicted side force; the operating tables in the November 13 tricone article give the bands, and the same logic governs fixed-cutter bits. Interfaces deserve their own drill: when the log or the cuttings show a soft-hard boundary within 10 meters, crews cut weight on bit by 30-50 percent and rotary speed by 20-30 percent through the transition, because the interface re-launches the bit with a lateral kick each time the support under the cutting structure changes. A sudden rate-of-penetration jump - 50 percent or more within one meter - is the field signal that the bit has entered a soft layer and the parameters no longer match; treating that as a parameter event, not a bonus, is what keeps a straight hole straight.
Survey methods and intervals
What is not measured cannot be held. The single-shot survey is the workhorse: one instrument drop, one inclination and azimuth reading, taken every 30-50 meters, which on a 200-meter well is 4-7 readings - enough to catch drift while it is still tens of centimeters, not meters. The multi-shot records the whole interval in one run and serves as the acceptance record. Measurement-while-drilling shows the trend in real time but rents at daily rates only deep or high-risk wells justify. For budgets that rule out all three, the hydrogen-fluoride acid tube is the classic low-cost fallback: a small glass tube etched by the acid while the string is stationary, read to about a quarter of a degree, at roughly 10-20 USD per reading. The table below compares the four.
Method | What it gives | Typical interval | Cost band | Best for |
|---|---|---|---|---|
Single-shot | one inclination and azimuth reading | every 30-50 m | tens of USD per run | routine straight-hole control |
Multi-shot | full profile in one run | final acceptance record | hundreds of USD | documentation for the client |
MWD | real-time trend while drilling | continuous | 1,000-2,500 USD per day | deep wells, high-risk sections |
HF acid tube | inclination to about 0.25 deg | every 30-50 m | 10-20 USD per reading | low-budget and remote sites |
Table 4. Survey methods compared by output, interval, cost, and best use.
Correction or Re-drill Decisions
Re-drill versus directional correction
When the survey shows the hole outside the limit, two roads remain, and the choice is arithmetic. Directional correction means a bent-housing mud motor drilling a small arc back toward the collar; rental for the motor, survey tools, and the directional hand runs 1,500-3,000 USD per day, and most water-well corrections close in 2-3 days for roughly 4,000-9,000 USD. It works when the excess is moderate - drift within about 1-2 times the limit and casing not yet set - and the remaining depth justifies the arc. Fill-and-re-drill means cement to above the offending interval and a second attempt, costing the 25,000-40,000 USD band and 14-21 days from the opening scenario. The working criteria: correction when the excess is small and the hole is clean; re-drill when drift exceeds roughly 3-4 meters, when keyseats or collapse complicate the borehole, or when the correction arc would cross the pump section anyway.
Situation | Recommended road | Typical cost | Typical time |
|---|---|---|---|
Drift 1-2x the limit, casing not set | directional correction | 4,000-9,000 USD | 2-3 days |
Drift over 3-4 m, keyseat or collapse | fill and re-drill | 25,000-40,000 USD | 14-21 days |
Drift caught under 1 m by survey | parameter and assembly fix | operating cost only | same run |
Correction arc would cross the pump section | fill and re-drill | 25,000-40,000 USD | 14-21 days |
Table 5. The correction decision by situation, with the cost and time each road typically takes.
Writing verticality into the plan
Verticality is cheapest as an input, not a finding. Before the rig moves, the contract should state the three numbers - drift per 100 meters, dogleg per 30 meters, survey interval - because acceptance disputes at pump-installation time are decided by what was written at tender time. The BHA is chosen against the formation dip on the stratigraphic column, not against habit; the survey schedule is planned at the same 30-50 meter rhythm the limits assume; and the readings are kept, because traceable well-quality data is becoming a project requirement, exactly as the compliance survey at the opening of this article recorded. Suppliers share the input side: bit selection against the formation band, a stable crown with adequate gauge, and a balanced cutter layout are factory decisions we make and document for every straight-hole order, with the standard 15-20 day delivery for PDC bits. A straight hole is drilled by decision, from tender clause to cutter layout, not rescued afterward.
Deviation has four causes, countered by packed-hole and pendulum assemblies, stable bit crowns with balanced cutters, weight below the buckling limit, surveys every 30-50 meters, and a fixed correction versus re-drill rule.
Frequently Asked Questions
PREVIOUS:
Contact Us








