摘要
提出了将多传感器数据融合技术引入到工程系统的状态监测与故障预报领域,实现多模型监测,以提高故障预报的准确性和可靠性,并研究了将Bayes融合算法和自适应加权算法相结合实现多模型的决策融合。并针对一套石油化工工业的流化催化裂化装置(FCCU)仿真系统建立了多模型监测系统,融合的结果表明,对比于单一模型预报,通过使用融合算法实现多模型监测可以有效地提高预报的可靠性。
In this paper,one attempt to introduce a metaphor procedure to multi-sensor data fusion for engineering system condition monitoring.To implement the strategy, three different models are built to monitor a simulated process of Fluid Catalytic Cracking Unit FCCU system. By using an improved Bayesian fusion algorithm,better diagnostic conclusions are obtained which is more reliable than that from one-model diagnosis results.
出处
《工业控制计算机》
2005年第6期19-20,61,共3页
Industrial Control Computer
基金
国家自然科学基金资助(资金编号:60234010)
