本文利用模糊聚類的神經(jīng)網(wǎng)絡(luò)方法，采用數(shù)據(jù)分組訓(xùn)練、自動(dòng)確定模糊分類線、在線測(cè)量時(shí)分類中心自適應(yīng)修正，降低了計(jì)算量，提高了溫度精度，將該算法用于陽(yáng)極焙燒中溫度預(yù)報(bào)的仿真，仿真實(shí)驗(yàn)以已有的焙燒爐溫度場(chǎng)模型為基礎(chǔ)，仿真結(jié)果表明該方法能夠解決陽(yáng)極焙燒溫度難以在線測(cè)量的問(wèn)題。
關(guān)鍵詞：軟測(cè)量；神經(jīng)網(wǎng)絡(luò)；模糊聚類；陽(yáng)極焙燒；溫度控制
Temperature soft sensor method in anode baking ZhangGeng MiaoLi SunZhenWei (College of Electrical and Information Engineering, Lanzhou University of Technology, Lanzhou Gansu 730050 , China) Abstract：A neural network soft sensor modeling method based on fuzzy clustering is used. The training data set is separated into several clusters with different centers, the number of fuzzy cluster is decided automatically, and the clustering centers are modified using an adaptive fuzzy clustering algorithm in the online stage. This method can reduce the calculation remarkably and has good temperature prediction accuracy. The proposed method has been applied to the temperature estimation in an anode baking. Simulation results, which based on the existed model of the baking temperature,
show that the method can deal with the measuring problem of temperature in anode baking online.
Keywords：soft sensor; neutral network; fuzzy clustering; anode baking; temperature control