Abstract:

The corrosion rate of the probe at the top of atmospheric tower in a domestic ten million ton refinery has been very high for a long time. The corrosion types of atmospheric tower top include HCl corrosion, H₂S corrosion, ammonium salt crystallization, erosion corrosion, etc. The influencing factors include temperature, flow rate, water injection, etc. The measured data include pH value, Fe²⁺, Cl^-, H₂S, NH⁺₄, naphtha distillation range, etc. The influence of various operating parameters and analysis detection items on the probe change trend is different, and there may be correlation between the values. Therefore, it is difficult to determine the main influencing factors and propose solutions. Based on the multiple data analysis function of Matlab, the principal components analysis(PCA) is mainly used here to analyze and sort the corrosion factors, find the corrosion contribution rate of 86.5%, and then provide theoretical support for the implementation of anti-corrosion measures. It provides a solution to determine the corrosion factors of atmospheric tower top under complex conditions.

Key words: Matlab, atmospheric tower top volatilization line, corrosion influent factors, probe, corrosion rate, PCA, feature vector, NH4Cl

摘要：

国内某千万吨级炼油厂常压塔塔顶探针出现长期腐蚀速率超高的情况。常压塔塔顶的腐蚀类型包括HCl腐蚀、H₂S腐蚀、铵盐结晶、冲刷腐蚀等，影响因素包括温度、流速、注水量等，检测的数据包括pH值，Fe²⁺含量，Cl^-含量，H₂S含量，NH⁺₄含量，石脑油馏程等。各操作参数、分析检测项目对于探针腐蚀变化趋势的影响程度不同，各数值间可能存在关联，因而，确定主要影响因素并提出解决办法存在较大的困难。基于Matlab(矩阵实验室)多元数据分析功能，应用主成分分析(PCA)对腐蚀因素进行分析和排序，找到86.5%的腐蚀贡献率，进而为防腐措施的实施提供理论支持。为复杂条件下常压塔塔顶腐蚀因素的判定提供了一条解决途径。

关键词: Matlab, 常压塔塔顶挥发线, 腐蚀影响因素, 探针, 腐蚀速率, PCA, 特征向量, NH4Cl