基于SOA-ELM的水泥分解炉温度预测  

Temperature prediction of cement decomposing furnace based on SOA-ELM

作　　者：王盛慧[1] 王金有[1] 

机构地区：[1]长春工业大学电气与电子工程学院,吉林长春130012

出　　处：《中国测试》2017年第8期91-94,共4页China Measurement & Test

基　　金：吉林省科技厅项目(20150203003SF)

摘　　要：在变工况的水泥生产过程中,为预知风、煤、料的投入量,提出一种基于人群搜索算法(SOA)优化极限学习机(ELM)的水泥分解炉温度预测模型。采用现场数据,选取相关因素,用ELM建立预测模型,通过SOA对ELM的输入输出权值进行动态寻优,克服其初始权值的随机性,实现分解炉温度的预测。与未优化权值的ELM模型和利用粒子群算法(PSO)优化的ELM模型进行仿真对比,实验表明该SOA-ELM模型具有更佳的预测能力。在隐层节点数为9时,该模型的预测值与真实值的平均相对误差为0.004 5%。该模型的建立,可为后期的分解炉温度控制提供依据。A temperature prediction model of cement decomposing furnace based on extreme learning machine（ELM） which was optimized by seeker optimization algorithm （SOA） was proposed to predict the input quantity of wind, coal and cement during cement production under variable working conditions. It could overcome the randomness of initial weight value of ELM to achieve prediction of decomposing furnace temperature based on dynamic optimization via the input and output weight of SOA to ELM and the way of establishing prediction model by using ELM according to field data and relevant factors. Simulation comparison test between it with ELM with non-optimized weight value and ELM model optimized based on particle swarm optimization （PSO） was carried out. The test results show that the SOA-ELM model has better prediction ability. When the number of nodes in hidden layer is 9, the average relative error between the predicted value and the true value of the model is 0.0045%. The establishment of this model provides a basis for the temperature control of the decomposing furnace at the later stage.

关 键 词：水泥分解炉温度 预测模型 人群搜索算法 极限学习机 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]