8.2 One-Foot Method, Variable Design Point & Reconstructed-Tank Shell Thickness

8.2 Variable Design Point & One-Foot Methods (API 650 & API 653)

API 653 Core Principle: In a vertical cylindrical storage tank under hydrostatic loading, the bottom shell-to-floor junction is clamped against radial expansion. This boundary restraint generates localized longitudinal bending moments and radial shears that alter circumferential hoop stress. While the One-Foot Method approximates this behavior at a fixed 12-inch elevation for tanks up to 200 feet in diameter, the Variable Design Point Method (VDM) mathematically determines the exact elevation of maximum stress, optimizing plate thickness and resolving borderline in-service fitness-for-service assessments.

Evaluating the stress distribution in an aboveground storage tank shell requires understanding cylindrical shell boundary mechanics. When a free, unconstrained thin-walled cylinder is filled with liquid, it expands radially in direct proportion to the hydrostatic pressure head: $w(x) = P(x) r^2 / (E t)$. However, an actual storage tank shell is welded to a rigid bottom plate assembly resting on an unyielding foundation. This structural boundary condition creates a significant mechanical discontinuity.

---

1. Structural Mechanics of Shell-to-Bottom Discontinuity

The bottom edge of Course 1 is fillet-welded on both sides to the tank bottom annular ring. Under the weight of the stored liquid, friction between the bottom plates and the foundation, combined with the membrane stiffness of the bottom sketch plates, almost completely arrests radial expansion at the shell-to-bottom joint ($w = 0$ at $x = 0$).

                    SHELL-TO-BOTTOM BOUNDARY INTERACTION

      Elevation (x)
           ^
           |        / Free Membrane Radial Deflection w(x)
           |       /  (Unconstrained Expansion: P * r^2 / E*t)
           |      / 
           |     /   
           |    /     +-----------------------------------------+
           |   /      | ACTUAL CONSTRAINED SHELL PROFILE        |
      12"  |--+-------| Peak Combined Hoop Stress Occurs Here!  |
           |  |       +-----------------------------------------+
           |  |
           |  | <--- Moment Discontinuity Zone (Radial Shear Q_0, Moment M_0)
           |  |
       0"  +==+================================================== Datum (Floor)
              ^ Clamped Shell-to-Bottom Fillet Weld (Radial Deflection w = 0)

The Beam-on-Elastic-Foundation Analogy

In classical elasticity theory (Timoshenko and Hetenyi), a vertical slice of a cylindrical shell acts as a beam supported continuously by an elastic foundation, where the elastic foundation modulus is provided by the circumferential hoop stiffness of the cylinder ($k = E t / r^2$).

The radial deflection $w(x)$ as a function of elevation $x$ above the bottom weld satisfies the fourth-order differential equation:

$\frac{d^4 w}{dx^4} + 4\beta^4 w = \frac{P(x)}{D_b}$

where $D_b = \frac{E t^3}{12(1-\nu^2)}$ is the flexural rigidity of the shell plate, and the characteristic damping attenuation parameter $\beta$ is defined as:

$\beta = \left[\frac{3(1-\nu^2)}{r^2 t^2}\right]^{1/4}$

Resulting Stress Regimes

1. At the Junction ($x = 0$): Radial expansion is completely restrained ($w = 0$), so circumferential membrane hoop strain is zero. However, the longitudinal bending moment ($M_0$) and radial shear force ($Q_0$) are at their absolute maximum.
2. Immediate Transition Zone ($0 < x < 12\text{ in.}$): As elevation increases, the clamping constraint of the bottom plate rapidly diminishes according to the exponential decay function $e^{-\beta x}$. Radial expansion develops, causing circumferential hoop stress to rise sharply.
3. Peak Combined Stress Zone: At a moderate elevation above the bottom joint, the sum of expanding circumferential hoop membrane stress and the decaying secondary bending stress reaches a mathematical peak. Beyond this point, hoop stress gradually decreases following the declining hydrostatic head $(H - x)$.

---

2. The One-Foot Method: Concept and Limitations

The One-Foot Method was developed historically as an empirical simplification of this complex boundary discontinuity. It assumes that the maximum combined circumferential hoop stress occurs precisely 12 inches (1.0 foot or 0.3 meters) above the bottom circumferential weld of each shell course.

Mathematical Formulation

Because the design point is fixed at 12 inches above the bottom girth weld, the effective liquid head acting at this point is simply $(H - 1)$:

$t = \frac{2.6 \cdot D \cdot (H - 1) \cdot G}{S \cdot E}$

   COURSE 1                     COURSE 2
   +----------------------+     +----------------------+ Top of Course 2
   |                      |     |                      |
   |                      |     |                      |
   |                      |     |                      |
   |                      |     |  * Design Point (x=1') -> Head = (H_2 - 1)
   |                      |     +----------------------+ Girth Weld 1-2
   |                      |     | Course 1             |
   |  * Design Point      |     |                      |
   |    (x = 1.0 ft)      |     |                      |
   +----------------------+     +----------------------+ Datum (Tank Floor)
     Head = (H_1 - 1)

Engineering Limitations of the One-Foot Method

Under API 650 Section 5.6.3.2 and API 653 Section 4.3.3.1, the One-Foot Method is subject to strict geometric boundaries:

• Diameter Limit: The One-Foot Method is valid only for tanks with nominal diameters up to 200 feet (60 meters).
• Conservatism on Large Tanks: When tank diameter exceeds 200 feet, or when the ratio of diameter to thickness ($D/t$) is very large, the characteristic decay parameter $\beta$ increases. This confines the rotational boundary restraint to a very narrow strip immediately adjacent to the bottom plate (often just 3 to 6 inches). Evaluating stress at 12 inches places the calculation point too far up the shell, resulting in an overly conservative shell thickness that wastes thousands of pounds of steel.
• Upper Courses Discontinuity: On upper shell courses with thinner plates and lower hydrostatic pressures, the 12-inch design point overestimates the bending interaction across horizontal girth seams between courses of differing thicknesses.

---

3. The Variable Design Point Method (VDM)

To eliminate the conservatism of the One-Foot Method on large tanks, Zick and McGrath formulated the Variable Design Point Method (VDM), codified in API 650 Section 5.6.4 and picked up by API 653 Sections 4.3.3.3 and 4.3.3.4.

Rather than fixing the calculation point at 12 inches, VDM mathematically computes the exact elevation $x$ above the bottom circumferential weld where circumferential stress reaches its true maximum.

Calculation for the Lowest Shell Course (Course 1)

For Course 1, the variable design point $x_1$ is calculated as the smaller of the values given by the following equations:

$x_1 = 0.61 \sqrt{r t_1} + 0.32 C h_1$

$x_1 = 0.26 h_1$

where:

• $r$ = nominal tank radius in inches ($r = 12 \times D / 2 = 6 D$)
• $t_1$ = thickness of Course 1 in inches
• $h_1$ = height of Course 1 in inches
• $C$ = bottom joint rotational stiffness factor, defined by:

$C = \frac{K^{0.5} (K - 1)}{1.0 + K^{1.5}}$

• $K = t_1 / t_2$ = ratio of the thickness of Course 1 to the thickness of Course 2

+-------------------------------------------------------------------------+
|                 VARIABLE DESIGN POINT ELEVATION (x_1)                   |
|                                                                         |
|   x_1 is governed by:                                                   |
|   1. Shell flexural parameter: 0.61 * sqrt( r * t_1 )                   |
|   2. Course height and bottom joint stiffness: 0.32 * C * h_1           |
|   3. Maximum geometric ceiling: x_1 <= 0.26 * h_1                       |
|                                                                         |
|   Once x_1 is determined (in inches):                                   |
|   Effective Head: H_eff = H - (x_1 / 12)                                |
|   t_min = [ 2.6 * D * ( H - x_1/12 ) * G ] / ( S * E )                  |
+-------------------------------------------------------------------------+

Calculation for Intermediate and Upper Courses

For upper courses (Course 2, Course 3, etc.), the design point $x_i$ is determined by evaluating the moment discontinuity across the horizontal circumferential weld separating the course under evaluation from the thicker course immediately below it:

$x_i = 2.10 \sqrt{r t_u} \cdot \left[ \dots \right]$

where $t_u$ is the thickness of the upper course, and $t_L$ is the thickness of the lower course.

---

4. API 653 Application: VDM as a Fitness-for-Service Reassessment Tool

API 653 splits the VDM rules across two consecutive paragraphs, and the distinction between them is exactly the kind of detail an open-book question probes:

API 653, 4.3.3.3 — permissive, for the smaller tanks. As an alternative to the one-foot method, the minimum acceptable shell plate thickness for tanks with diameters equal to or less than 200 ft may be calculated in accordance with the variable design point method of API 650, substituting "S x E" for "S", with $E$ and $S$ defined as in 4.3.3.1.

API 653, 4.3.3.4 — mandatory, for the larger tanks. The variable design point method shall be used for tanks greater than 200 ft in diameter, with all variables defined as in 4.3.3.1.

Exam scope note. API's API 653 Body of Knowledge lists "variable point method calculations" under the code calculations candidates are not expected to perform for the certification exam. Understand what VDM is, when 4.3.3.3 permits it and when 4.3.3.4 requires it, and why it differs from the one-foot method — but do not expect to grind through the Zick-McGrath equations under exam time pressure.

Strategic Fitness-for-Service Utility

During an internal inspection, ultrasonic thickness testing frequently reveals that an existing shell course has thinned to a value marginally below $t_{\text{min}}$ calculated by the standard One-Foot method (e.g., $t_{\text{act}} = 0.510\text{ in.}$ vs. $t_{\text{min}} = 0.530\text{ in.}$). Under the One-Foot method, the tank must be downrated or subjected to costly shell plate replacement.

By engaging a tank engineer to perform a VDM reassessment:

1. The variable design point $x$ is accurately calculated, which typically shifts the design point higher up the shell (e.g., $x_1 = 18\text{ to }26\text{ inches}$ instead of $12\text{ inches}$). This higher elevation reduces the effective liquid head: $(H - x/12) < (H - 1)$.
2. The recalculated $t_{\text{min}}$ via VDM is often 5% to 15% thinner than the One-Foot method result.
3. The corroded course is mathematically proven to be structurally acceptable ($t_{\text{act}} \ge t_{\text{min,VDM}}$), saving the facility tens or hundreds of thousands of dollars in unnecessary cut-outs, insert plates, or operational derating.

---

5. Comprehensive Comparison: One-Foot Method vs. VDM

---

6. Worked Comparison Example: Large Diameter Tank

Consider an existing crude oil storage tank:

• Nominal Diameter $D = 240\text{ ft}$ (Radius $r = 1,440\text{ in.}$)
• Maximum liquid height $H = 48\text{ ft}$
• Specific gravity $G = 1.00$
• Material: ASTM A283 Grade C, bottom course ($S = 23,600\text{ psi}$ from API 653 Table 4.1)
• Joint efficiency: Full radiography ($E = 1.00$)
• Course 1 height $h_1 = 96\text{ in.}$ (8 ft)

1. One-Foot Method Calculation (shown for comparison only):

Note that API 653 4.3.3.1 limits the one-foot method to tanks of 200 ft diameter or less, so at $D = 240$ ft this number has no code standing — 4.3.3.4 requires VDM here. It is computed purely to show the size of the conservatism the one-foot method would introduce: $t_{\text{min,1}} = \frac{2.6 \times 240 \times (48 - 1) \times 1.00}{23,600 \times 1.00} = \frac{624 \times 47}{23,600} = \frac{29,328}{23,600} = 1.2427\text{ in.}$

2. Variable Design Point Method (VDM) Calculation:

Assume Course 1 nominal thickness $t_1 = 1.15\text{ in.}$ and Course 2 thickness $t_2 = 0.95\text{ in.}$ ($K = 1.15 / 0.95 = 1.21$):

1. Calculate stiffness parameter $C$: $C = \frac{(1.21)^{0.5} (1.21 - 1)}{1.0 + (1.21)^{1.5}} = \frac{1.10 \times 0.21}{1.0 + 1.33} = \frac{0.231}{2.33} = 0.0991$
2. Calculate design point $x_1$: $x_1 = 0.61 \sqrt{r t_1} + 0.32 C h_1 = 0.61 \sqrt{1,440 \times 1.15} + (0.32 \times 0.0991 \times 96)$ $x_1 = 0.61 \sqrt{1,656} + 3.044 = 0.61 \times 40.69 + 3.044 = 24.82 + 3.04 = 27.86\text{ inches}$ Check limit: $x_1 \le 0.26 h_1 = 0.26 \times 96 = 24.96\text{ inches}$. The limit governs: $x_1 = 24.96\text{ in.} = 2.08\text{ ft}$.
3. Calculate effective liquid head: $H_{\text{eff}} = H - \frac{x_1}{12} = 48 - 2.08 = 45.92\text{ ft}$
4. Calculate required thickness via VDM: $t_{\text{min,VDM}} = \frac{2.6 \times 240 \times 45.92 \times 1.00}{23,600 \times 1.00} = \frac{28,654}{23,600} = 1.2141\text{ in.}$

Resulting Benefit: The VDM calculation demonstrates that the required shell thickness is reduced from $1.243\text{ in.}$ to $1.214\text{ in.}$, an immediate thickness reduction of nearly $0.030\text{ in.}$ that can qualify an in-service corroded shell without repair.

---

7. Reconstructed Tank Shells: A Different Calculation Family Entirely

Everything above deals with an existing tank evaluated under API 653 Section 4. A reconstructed tank — one that has been dismantled and re-erected, per API 653 Section 10 — is not evaluated with Section 4 stresses at all. API 653 Section 8, Design Considerations for Reconstructed Tanks, sends you back to API 650 new-construction rules, and the BOK lists this as its own calculation category.

Step 1 — Get the two allowable stresses (API 653, 8.4.2 and 8.4.3)

Note the contrast with an in-service evaluation: API 653 Table 4.1 stresses (0.80Y / 0.429T etc.) are expressly not for use for reconstructed tanks. Reconstruction reverts to the more conservative API 650 design basis.

Step 2 — Calculate $t_d$ and $t_t$ by the one-foot method (API 650, 5.6.3.2)

For tanks of 200 ft (61 m) diameter and smaller:

$t_d = \frac{2.6 \cdot D \cdot (H - 1) \cdot G}{S_d} + CA$

$t_t = \frac{2.6 \cdot D \cdot (H - 1)}{S_t}$

where

• $t_d$ = design shell thickness (in.); $t_t$ = hydrostatic test shell thickness (in.)
• $D$ = nominal tank diameter (ft); $H$ = design liquid level (ft), measured from the bottom of the course under consideration
• $G$ = design specific gravity of the liquid to be stored
• $CA$ = corrosion allowance (in.) — added to $t_d$ only, never to $t_t$ (the test is run on clean water with no future corrosion assumed)

Step 3 — Take the governing value

The required thickness of the course is the greater of $t_d$ and $t_t$, and in no case less than the nominal minimum thickness API 650 assigns by tank diameter:

Worked Example — Reconstructed Course 1

A tank is being re-erected at a new site: $D = 100$ ft, $H = 36$ ft, $G = 0.95$, corrosion allowance $CA = 0.0625$ in. The shell material gives $S_d = 23,200$ psi and $S_t = 24,900$ psi from API 650 Table 5.2a.

$t_d = \frac{2.6 \times 100 \times (36 - 1) \times 0.95}{23,200} + 0.0625 = \frac{8,645}{23,200} + 0.0625 = 0.3726 + 0.0625 = 0.4351\text{ in.}$

$t_t = \frac{2.6 \times 100 \times (36 - 1)}{24,900} = \frac{9,100}{24,900} = 0.3655\text{ in.}$

The design condition governs ($0.435 > 0.366$), and $0.435$ in. comfortably exceeds the 1/4-in. nominal minimum for a 100-ft tank. Course 1 must be at least 0.435 in. thick.

The classic trap: candidates drop the corrosion allowance out of $t_d$, or add it to $t_t$ as well. Adding $CA$ to $t_t$ makes the hydrotest thickness artificially govern and produces the wrong answer. $CA$ belongs to $t_d$ alone.