Abstract: Coal-water slurry gasification is a very important technology in developing clean and efficient use of coal.Gasifier furnace temperature is one of the key variables which is closely related to the process safety,stability and long-term operation of the gasification.Thermocouple elements are easily ruined under complex industrial condition with high temperature,high pressure and high flow erosion.Thus it is difficult to maintain a long period of work.In this paper,aiming at an opposed multi-burner coal-water slurry gasification process,adopting fuzzy C-means clustering based multi-modeling method and least square support vector machines,a gasifier temperature soft sensor model is established.The actual operation’s validation results show that the predictive temperature of the furnace based on this soft sensor model has a pretty good predictive precision and generalization ability.

Key words: coal-water slurry gasification, fuzzy C-means clustering, least square support vector machines, multi-modeling, soft sensor modeling

摘要： 水煤浆气化是煤炭资源高效清洁利用的重要技术。气化炉反应温度是关系装置能否长周期安全稳定运行的关键参数,但是热电偶在高温、高压和气固物流冲刷环境下,使用寿命有限。本文以一多喷嘴对置式水煤浆气化炉为研究对象,在多模型建模方法的基础上,以数据点间的相似程度作为多模型子区间的划分手段,结合最小二乘支持向量机建立了基于模糊C均值聚类的气化炉温度软测量模型。实际工业运行数据验证结果表明,该软测量模型拟合精度较高,模型泛化能力较强。

关键词: 水煤浆气化, 模糊C均值聚类, 最小二乘支持向量机, 多模型, 软测量建模