应用BP网络进行重油化学组成的关联计算

CORRELATION OF CHEMICAL COMPOSITION OF HEAVY OILS USING BP NEURAL NETWORK

描述了应用具有高度非线性映射能力的BP神经元网络进行石油物性数据关联的基本方法，以重油的密度、碳氢比、粘度和康氏残炭等常规物性作为关联参数，对重油的化学组成进行关联，结果优于由回归分析所得的关联式。该方法具有统一的模型，自动形成物性与关联参数的相关关系，以及计算简单方便等特点，适合于处理组成极其复杂的石油体系。

A method is presented for the correlation of the physical properties of oils using BP neural network.The chemical composition of heavy oils are then correlated with the ordinary physical properties such as density,viscosity,C/H ratio and Conradson carbon residue etc.. The results are better than those obtained from regressive analysis method in literature. The BP neural network method is charaterized by its universal mathematical model,automatic formation of correlation relation between physical properties and correlation parameters, simplicity and convenience in calculation, and is appropriate for the oil systems with very complicated components.