7.1 Determining Corrosion Rates (Short-Term vs. Long-Term)

7.1 Determining Corrosion Rates (Short-Term vs. Long-Term)

API 653 Core Principle: Accurate corrosion rate determination forms the mathematical foundation of storage tank mechanical integrity. Underestimating the degradation rate leads to unexpected wall puncture, environmental contamination, or catastrophic structural rupture; overestimating the rate causes premature, multimillion-dollar turnarounds and unnecessary metal replacement.

In storage tank integrity management, corrosion is rarely a perfectly linear, steady-state process over decades of operation. Feedstocks change, operating temperatures fluctuate, chemical inhibition systems fail, and bottom-water draw-offs vary in effectiveness. Consequently, API 653 Section 4 and Section 6 require tank inspection engineers to evaluate degradation over multiple time horizons, establishing both Long-Term and Short-Term corrosion rates to assess structural viability.

---

1. Mathematical Formulations: LTCR vs. STCR

Corrosion rates for aboveground storage tank components (shell courses, floor plates, annular rings, and fixed roofs) are calculated by tracking metal loss over measured operating intervals.

                    CORROSION RATE TIMELINE & DEFINITIONS

   t_initial (Baseline / New)            t_previous (Last Outage)       t_actual (Current)
       |
       |=====================================
       |                                    |
       |<------- Long-Term Interval ------->|<--- Short-Term Interval --->|
       |         (e.g., 20 Years)           |      (e.g., 5 Years)        |
       |                                    |                             |
       +------------------------------------+-----------------------------+
     Year 0                               Year 15                       Year 20

Long-Term Corrosion Rate (LTCR)

The Long-Term Corrosion Rate reflects the time-averaged degradation of a specific inspection point or component across its entire service life (or from initial installation to the current inspection):

$\text{LTCR} = \frac{t_{\text{initial}} - t_{\text{actual}}}{\text{Time elapsed between } t_{\text{initial}} \text{ and } t_{\text{actual}}}$

where:

• $t_{\text{initial}}$ = nominal initial thickness or baseline thickness at commissioning (inches or mm)
• $t_{\text{actual}}$ = actual thickness measured during the current inspection (inches or mm)
• $\text{Time elapsed}$ = operating duration between $t_{\text{initial}}$ and $t_{\text{actual}}$ (typically in years)

Short-Term Corrosion Rate (STCR)

The Short-Term Corrosion Rate measures degradation occurring exclusively between the two most recent inspection cycles:

$\text{STCR} = \frac{t_{\text{previous}} - t_{\text{actual}}}{\text{Time elapsed between } t_{\text{previous}} \text{ and } t_{\text{actual}}}$

where:

• $t_{\text{previous}}$ = thickness measured during the immediate prior inspection (inches or mm)
• $t_{\text{actual}}$ = actual thickness measured during the current inspection (inches or mm)
• $\text{Time elapsed}$ = operating duration between the two consecutive inspections (years)

Unit Conventions

In API 653 practice, corrosion rates are expressed in inches per year (in./yr) or mils per year (mpy), where $1\text{ mil} = 0.001\text{ in.}$ (or in metric units as millimeters per year, mm/yr): $1.0\text{ mpy} = 0.001\text{ in./yr} \approx 0.0254\text{ mm/yr}$

---

2. Engineering Criteria for Selecting Between STCR and LTCR

API 653 mandates that the corrosion rate utilized in remaining life calculations must accurately project future metal loss. An inspector or storage tank engineer cannot simply pick the lower rate to artificially extend run-length. The selection between STCR and LTCR is governed by strict engineering criteria:

+-------------------------------------------------------------------------+
|                 CORROSION RATE SELECTION DECISION LOGIC                 |
|                                                                         |
|   Compare STCR vs. LTCR:                                                |
|                                                                         |
|   [ STCR > LTCR ] ---> Did operational / process changes occur?         |
|                        |                                                |
|                        +-- YES: (Sour crude, temperature increase,      |
|                        |         water draw failure, biocide lapse)     |
|                        |         ====> MUST USE STCR (Governing Rate)   |
|                        |                                                |
|                        +-- NO:  Investigate UT measurement error,       |
|                                 probe calibration, surface roughness.   |
|                                 If valid, USE STCR conservatively.      |
|                                                                         |
|   [ STCR < LTCR ] ---> Did genuine mitigation occur?                    |
|                        |                                                |
|                        +-- YES: (New internal lining applied, CP grid   |
|                        |         energized, inhibitor program active)   |
|                        |         ====> STCR may be justified.           |
|                        |                                                |
|                        +-- NO:  Instrument error or non-representative  |
|                                 sampling. USE LTCR (Conservative).      |
+-------------------------------------------------------------------------+

When STCR Must Govern (Accelerated Conditions)

STCR must be selected whenever operating conditions have become more aggressive, causing recent metal loss to outpace historical averages. Key operational drivers include:

1. Crude Slate Shifts to Sour Service: Transitioning from sweet crude to sour crudes containing active hydrogen sulfide ($H_2S$), elevated Total Acid Number (TAN, naphthenic acids), or mercaptans aggressively attacks internal shell and floor surfaces.
2. Increased Water Draws / Stagnant Water Bottoms: Crude oil tanks routinely settle an emulsion of water, salts, and sludge at the bottom. If water draw-off schedules are missed or draw lines clog, acidic, anaerobic water pools against bottom plates, accelerating pitting corrosion.
3. Operating Temperature Elevation: In heavy oil, asphalt, or fuel oil storage, increasing bulk temperature to reduce viscosity dramatically escalates chemical reaction kinetics (governed by the Arrhenius relationship) and accelerates oxidation in the vapor space.
4. Microbiologically Influenced Corrosion (MIC) Upsets: Ineffective biocide dosing or prolonged water contact allows anaerobic Sulfate-Reducing Bacteria (SRB) or Acid-Producing Bacteria (APB) to proliferate beneath bottom sludge, driving localized pitting at rates exceeding $30\text{ to }50\text{ mpy}$ ($0.030\text{ to }0.050\text{ in./yr}$).
5. Vapor Space Oxygenation: Breaching tank nitrogen blanketing or installing faulty pressure/vacuum conservation vents introduces moist ambient oxygen, driving rapid general rusting of roof plates and the upper shell course.

Engineering Rule: If STCR is substantially higher than LTCR and can be linked to an operational or environmental change, STCR is the only acceptable rate for remaining life calculations. Using LTCR in this scenario dilutes the real, ongoing damage mechanism with historical benign data.

When LTCR Governs (Steady-State Reliability)

In tanks operating in continuous, unchanged service with stable process conditions, the LTCR is typically preferred. Ultrasonic thickness measurements have inherent tolerances (typically $\pm 0.005\text{ to }\pm 0.010\text{ in.}$ due to surface roughness, paint thickness, couplant variations, or transducer calibration). Across a short 2-year or 3-year interval, an instrument reading discrepancy of $0.010\text{ in.}$ can yield a false "spike" or "negative" STCR. The LTCR integrates measurements across 15 to 30 years, damping instrument noise and providing a reliable steady-state rate.

3. Determining Corrosion Rates When Historical Data Is Missing

When inspecting newly acquired facilities, older legacy tanks, or tanks where prior inspection dossiers have been lost, the inspector cannot calculate LTCR or STCR directly. API 653 establishes clear, prescriptive protocols to resolve missing baseline data.

Protocol 1: Similar Service Evaluation (API 653 Annex H)

In accordance with API 653 Section 6.3.3.2 and Annex H, corrosion rates may be estimated based on the documented operating history of a comparable tank operating in "Similar Service." Annex H defines rigorous criteria that must be satisfied to validate similarity:

• Product Stored: Same chemical composition, gravity, vapor pressure, and corrosivity class (e.g., low-sulfur diesel vs. high-sulfur diesel).
• Operating Temperature: Storage temperature profile must be comparable within defined engineering bands.
• Bottom and Foundation Design: Similar pad material (crushed stone, oiled sand cushion, concrete slab) and drainage geometry.
• Cathodic Protection Status: Both tanks must possess comparable CP systems (e.g., both with functioning ICCP ribbon grids meeting -850 mV CSE instant-off criteria).
• Internal Protective Linings: Same lining chemistry (e.g., 20-mil thin-film amine epoxy vs. bare steel) and age condition.
• Water Draw-Off Frequency: Comparable water-settling and drainage operational discipline.

Protocol 2: Published Industry Data and Conservative Defaults

If no similar service tank exists within the owner-user's fleet, the inspector may consult recognized industry corrosion databases (e.g., API RP 571, NACE / AMPP publications, or refining corrosion surveys). In the absence of definitive data, conservative default rates must be assigned (e.g., $5\text{ to }10\text{ mpy}$ for unlined crude tank bottoms, $2\text{ to }5\text{ mpy}$ for refined products).

Protocol 3: Prescriptive 5-Year Inspection Cap

To safeguard against inaccurate estimations, API 653 imposes a strict statutory constraint:

• API 653 Section 6.3.3.2.a: When a tank's corrosion rate is unknown and cannot be reliably established from similar service, an external ultrasonic shell thickness examination shall be conducted at an interval not to exceed 5 years.
• This mandatory 5-year inspection provides the necessary intermediate thickness reading, establishing a genuine two-point baseline from which an actual short-term rate can subsequently be calculated.

---

4. Statistical Treatment of Floor and Shell Thickness Data

Ultrasonic thickness (UT) and Magnetic Flux Leakage (MFL) surveys generate tens of thousands of data points across a storage tank. Interpreting this massive volume of data requires sound statistical methodologies rather than simple single-point evaluations.

+-------------------------------------------------------------------------+
|                    SHELL VS. FLOOR UT EVALUATION LOGIC                  |
|                                                                         |
|   SHELL EVALUATION (Membrane Hoop Stress):                              |
|   * Critical Length L = 3.7 * sqrt(D * t2)                              |
|   * Average thickness along vertical profile within L must >= t_min     |
|   * Isolated pits permitted below t_min under Section 4.3.2.2 limits    |
|                                                                         |
|   FLOOR EVALUATION (Leak Prevention / Hydrostatic Bedding):             |
|   * MFL continuous volumetric floor screening                           |
|   * Prove-up UT on deepest indications                                  |
|   * Minimum Remaining Thickness (MRT) criteria per Table 4.4            |
|   * Extreme Value Analysis (EVA / Gumbel) for uninspected areas         |
+-------------------------------------------------------------------------+

Shell Thickness: Critical Length (L) Averaging

Unlike pipe or pressure vessels where local thinning can trigger burst failures, tank shells resist hydrostatic hoop tension across broad circumferential bands. Therefore, API 653 Section 4.3.2.1 permits thickness averaging over a defined vertical length:

• Critical Length Formula: $L = 3.7 \sqrt{D \times t_2}$ where $D$ is tank diameter in feet, and $t_2$ is the minimum recorded thickness in the thinned area in inches (or $L = 0.0119 \sqrt{D \times t_2}$ in metric units with $D$ in meters and $t_2$ in mm). Length $L$ must not exceed 40 inches.
• Vertical Profile Averaging: Thickness readings are taken at multiple vertical intervals along the centerline of the corroded zone across length $L$. The mathematical average of these readings ($t_{\text{avg}}$) must equal or exceed the minimum required thickness $t_{\text{min}}$ calculated for that shell course.
• Isolated Pitting Exception (Section 4.3.2.2): Widely scattered pits are ignored if:
  1. No pit depth leaves less than half the minimum required thickness ($t_{\text{pit}} \ge 0.5 \times t_{\text{min}}$).
  2. The sum of pit dimensions along any vertical line does not exceed 2 inches in an 8-inch segment.

Floor Thickness: MFL and Extreme Value Analysis (EVA)

Tank bottom evaluation is governed by leak prevention rather than membrane stress. Bottom plates are inspected using automated Magnetic Flux Leakage (MFL) or phased array crawlers, followed by manual UT prove-up of flagged defect indications.

• Minimum Thickness Criteria: Floor plates must satisfy the Minimum Remaining Thickness (MRT) rules of API 653 Section 4.4.5 and Table 4.4 at the next scheduled turnaround (e.g., $0.100\text{ in.}$ for unlined floors without an RPB).
• Extreme Value Analysis (EVA / Gumbel Distribution): Automated scanners cannot access floor steel located beneath internal heating coils, roof support columns, water draw sumps, and shell lap joints. In modern Risk-Based Inspection (RBI per API RP 581), statistical Extreme Value Analysis (EVA) using the Gumbel extreme value distribution is applied to the population of measured pit depths. EVA calculates the statistical probability that an unmeasured pit in an inaccessible floor area exceeds the penetration threshold before the next scheduled outage.

---

5. Summary Comparison of Corrosion Rate Methodologies