Factor Analysis about Soil Corrosion of Buried Metal Pipe

Corrosion is one of the basic affecting factors that make metal-pipe disabled, and is under the influence of many factors. But it is hard to accurately get each factor's value, and dispersion and randomness of statistical data make it hard to apply traditional method in date processing, too. Projection Pursuit Method is an objective evaluation method. It, uninfluenced by human factors, entirely depends on the mathematic principle and considers index's correlation. In the paper Projection Pursuit Method is applied to evaluate soil corrosion of X70 pipeline from Changyang to Wuhan. Real-coded Accelerating Genetic Algorithm (RAGA) is utilized to optimize the projection direction of Projection Pursuit clustering (PPC) model, and multi-dimension data is transformed to low dimension subspace. Each index of factors is evaluated based on pursuing optimal projection direction and projection value. Then based on the above the optimal projection direction, weight vector can be calculated which objectively reflects the index's correlation. We can rank weight vector in order from big to small, and the results show that the soil electrochemistry property factors of corroding the X70 pipeline are CL^–, SO₄^2–, soil resistivity, HCO₃^–, pH value and volumetric moisture content in turn.