生物氧化预处理ORP实时动态精准预测研究

提出基于相空间重构理论和学习型狼群算法优化最小二乘支持向量回归机(LSSVR)的氧化还原电位(ORP)实时动态精准预测方法。以生物氧化预处理过程中采集到的ORP数据为研究对象,采用小波分析滤除ORP数据中的噪声,以相空间重构的ORP时间序列训练LSSVR预测模型。为了提高模型的泛化能力,提出学习型狼群算法优化LSSVR模型参数,并用时间窗口平移化方法和反馈校正法分别对模型和预测输出进行更新和修正。试验结果对比表明所提方法可行有效。

Real-time Dynamic Accurate Prediction of ORP in Biological Oxidation Pretreatment

In order to predict real-time change of Oxidation Reduction Potential (ORP) in production process, a real-time dynamic accurate prediction method was presented for oxidation reduction potential ORP based on theory of phase space reconstruction and optimized Least Squares Support Vector Regression (LSSVR) by learning wolf pack algorithm. Wavelet analysis was used to remove noise from ORP data, and ORP sequence was used to train LSSVR model based on phase space reconstruction. A learning Wolf Swarm Optimization Algorithm was proposed to optimize LSSVR parameters and time window translation method was used to update the model. Feedback correction method was used to revise model predictive output. The experimental results show that the proposed method has a better predictive effect.