API 653 and Tank Bottom Inspection and Evaluation

By Vince Carucci

API 653 states that “leaks in tank bottoms are not acceptable while tanks are in service.” It goes on to state that “periodic assessment of tank bottom integrity shall be performed.” API 653 does not mandate exactly how to perform this integrity assessment, what inspection techniques to use, nor the extent of that inspection. Several points are worth keeping in mind:

Bottom Plate Thickness Evaluation

API 653 specifies minimum acceptable thicknesses for the tank bottom plate and annular plate. It also indicates that the remaining thicknesses may be quantified using either a “probabilistic” method or a “deterministic” method. The bottom plate thicknesses are then compared to the required thicknesses to determine their acceptability.

The probabilistic method uses a statistical approach to extrapolate a relatively small amount of inspection data into a prediction of whether the tank bottom is too thin. This method has proven to be an accurate prediction technique. Unfortunately if it indicates that the bottom is too thin, a more extensive inspection must be done anyway to locate the thin areas.

The deterministic method uses more extensive inspection data to quantify the remaining thickness of the bottom. The required data include the following:

Current trends appear to be directed toward increased use of the deterministic approach.

Bottom Plate Thickness Measurement

From an evaluation standpoint, it would be ideal to have a complete thickness map of the bottom. However, it would be expensive to perform the ultrasonic inspection that is necessary to do that, and such an extensive survey is not necessary.

A bottom thickness inspection approach that is now commonly used is to employ Magnetic Flux Exclusion (MFE) testing equipment to inspect the tank bottom. We won't go into theory of how MFE equipment works here. However, the technique requires that the MFE equipment be set to a bottom thickness loss threshold (typically 30% to 40% of the original plate thickness) and then it is used to scan the entire bottom (like walking behind a lawnmower). The MFE equipment will identify any locations that are thinner than the threshold setting, but it will not quantify the thickness. Follow-up UT measurements are then made at the thin areas that were identified by the MFE. The thickness data are then used to perform a deterministic evaluation of the bottom.

The MFE testing with follow-up UT inspection is a cost-effective approach to bottom plate thickness inspection. However, several points must be kept in mind:


Integrity evaluation of the bottom plate requires a good understanding of inspection techniques, their limitations, and acceptance criteria. Since even a small leak is not acceptable and stress is not a primary consideration, proper interpretation and use of the inspection data are critical.