摘要
论文采用神经网络模型对催裂化分馏塔的产品质量进行在线预测分析,采用CLISP与VisualC++语言混合编程的方法,开发基于网络环境下实时在线的故障诊断专家系统。这两种技术相结合,克服了各自的缺陷,实现了优势互补。结果表明该模型具有较高的精度,与化验值的拟合成度较好,给予了在线操作指导,减少了质量事故,提高了经济效益。
In this paper,the online estimation of FCC oil product qualities is realized by BP neural network and the real_time fault in the FCC Fractional Column is diagnosed by embedding CLIPS expert system into VC ++6.0.Combinating CLIPS with BP makes them overcome disadvantages and realize the complementarity of strongpoints.The results have been accurately proved by the experiments.The model can give online guiding operations,decrease quality accidents and increase economic profit.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第3期222-225,共4页
Computer Engineering and Applications
基金
天津市科技发展计划项目(编号:993109511)
天津市科学技术成果(编号:20040002)
