Random laws of kriging method for simulation of characteristic of corroded surface

Tóm tắt

The geostatistical method to simulate the corrodded surface of steel structures is one of the most accurate corrosion simulation methods, clearly simulating the correlation of corrosion depth and corrosion characteristics on the steel surface. This study applied geostatistical theory, specifically Kriging, to simulate 11 times based on experimental data. The data set has a length of 200mm, a width of 36mm, and a grid spacing of 1mm. The results of each simulation were compared with those of the specimen data set from the experiment. The Kriging method can not only create a spatial distribution of the corrosion surface with arbitrary length, width, and grid dimensions but also achieve high reliability of characteristic values such as average value, min, max, variance, STDEV, Lag distance, average semivariance, Pairs, Nugget variance (Co), Structural varianceSill (Co+C), Range a, ... as well as the histogram distribution rule. The changing trend of the corrosion surface over time is also accurately predicted. They are creating favorable conditions to determine the behavior of corroded steel structures over time. However, creating a spatial distribution surface by Kriging is still random when each surface simulation is not the same, making the simulation results challenging to apply to finding the first corrosion point in practice.