A wet-stock management and leak detection system for fuel tanks

One of the critical problems in wet-stock management is inaccurate (poor) tank calibration that masks the leakages. This study aims to prevent the masking effect and improve the leak detection for petroleum storage tanks. The common tank types used are simple cylindrical and cylindrical with dished heads. Through obtaining mathematical models for both tank types that convert the measured liquid height to accurate volume and taking into account the tank deformations this goal is achieved. The accounted deformations in tank parameters are radius, length, dished head depth, probe offset and tilt (axial and radial). The analysis of the real time data showed that the deformations can be categorized into two as uniform and non-uniform, both of which require different approaches. The simulations using the actual data gathered from different fuel service stations showed that the approach developed in this study is valid. Within the five sets of data collected from different tanks, the first tank is used only for its parameters to simulate the effects. The second one is classified as non-problematic that does not require any corrections and the uncorrected model data fits perfectly. Analysis of the third and fourth tanks showed uniform deformations, the deformation included tank models successfully predicted the real fuel volumes. The last tank was non-uniformly deformed. Although the parameterized models did not perform well for this case, inclusion of an error (or correction) function produced accurate results. The results showed that the variance in the uniformly deformed tanks reduced from -200 L – +20 L range to -20 L – +40 L range for and -300 L – +100 L range to -50 L – +20 L range for, which brings 81.3% and 81.1% improvement. In the non-uniform deformation case, by using only the parameterized models the variance changes from -800 L to +800 L range. However, inclusion of the error function reduces the variance to -15 L – +15 L range and brings a 97.7% improvement.