主成分回归法在炼油企业经营管理中的应用研究

Application of principal component regression( PCR) in business management of petroleum refining companies

生产装置数据信息有效地应用于管理和决策,有助于生产模式和效益增长方式的转变和重新构建,是炼油企业提高精细化管理水平的关键。常规算法在解决此类高维多重共线性问题时,模型的精度和适应性难以获得较理想效果。引入主成分回归(PCR)方法,以某炼化企业渣油加氢装置综合能耗与其原料和产品班平衡量差之间的关系为例,建立其关系模型,并将预测值与实际值进行分析比较。结果表明:所建模型具有较好的稳定性和精度,模型的相对误差小于0.1,并有较好的适应性和实用性。也表明PCR等数据挖掘技术和方法在炼化行业具有广阔的应用前景。

The effective application of data information of process units in the decision making and management will help to modify and restructure the modes of production and profits increase, which is very critical for the petroleum refining companies to improve their fine management. When the conventional methods are utilized to solve the high dimensional multicollinearity problems, it is difficult to get satisfactory results in mod- el accuracy and applicability. Therefore, the principal component regression （PRC） method is introduced and applied in a case study of the comprehensive energy consumption and product balance difference of a residue hydrotreating unit in a petroleum refinery. A relation model is established for the comprehensive energy consumption and product balance difference. The analysis and comparison of predicted data and actual data show that the model established has good stability, accuracy, adoptability, and applicability. The relative error of the model is less than 0.1. The data mining techniques and methods such as PCR, etc have a wide application in petroleum refining and chemical industries.