Research on Data Mining Model of Corrosion Influencing Factors Based on Dual Algorithm-Driven

Abstract To address the issue of unclear correlations between corrosion factors in industrial process pipelines, this study utilizes a dataset of corrosion-related indicators from the oil-gas pipeline of an atmospheric and vacuum distillation unit in a refinery. By integrating the correlation coefficient method with the Apriori association rule mining algorithm, a data mining algorithm based on strong association rules of corrosion influence factors in petroleum pipelines is proposed. Kolmogorov-Smirnov test is applied to the dataset to determine whether the data follow a normal distribution, and correlation analysis is used to identify strongly correlated datasets. The strongly correlated datasets are then encoded, and the Apriori association rule mining algorithm is employed to analyze the association strength between corrosion influence factors and corrosion rates, deriving strong correlation rules. A validation dataset is used to verify the obtained strong association rules. This study provides significant practical value for optimizing corrosion prevention measures in process pipelines and serves as a valuable reference for further research in related fields.