电子测量与仪器学报

陈如清,于志恒.基于 TentFWA-GD 的 RBF 神经网络 COD 在线软测量方法[J].电子测量与仪器学报,2022,36(3):53-60

基于 TentFWA-GD 的 RBF 神经网络 COD 在线软测量方法

COD on-line soft measurement based on TentFWA-GD RBF neural network

DOI：

中文关键词: 农村生活污水处理 COD 软测量 RBF 神经网络烟花算法 Tent 混沌映射

英文关键词:rural domestic sewage treatment soft sensor of chemical oxygen demand RBF neural network fireworks algorithm Tent chaotic mapping

基金项目:浙江省基础公益研究计划项目(LGG18F030011)资助

作者	单位
陈如清	1.嘉兴南湖学院机电工程学院
于志恒	1.嘉兴南湖学院机电工程学院

Author	Institution
Chen Ruqing	1.College of Mechanical and Electrical Engineering, Jiaxing Nanhu University
Yu Zhiheng	1.College of Mechanical and Electrical Engineering, Jiaxing Nanhu University

摘要点击次数: 247

全文下载次数: 564

中文摘要:

针对污水处理过程 COD 难以实时准确测量的问题,提出了基于 TentFWA-GD 的 RBF 神经网络软测量方法。为解决现有 RBF 神经网络用于复杂工业过程软测量建模时存在网络参数难以确定及训练过程易陷入局部极值等问题,进一步提高 RBF 神经网络模型的预测精度与泛化能力,引入了 Tent 混沌映射对烟花算法(fireworks algorithm,FWA)进行改进,利用混沌运动的全局遍历性维持 FWA 的种群多样性并避免算法早熟收敛;将 TentFWA 算法与 GD 方法有机融合提出一种改进的 RBF 神经网络组合训练方法以改善网络的学习能力。将基于 TentFWA-GD 的 RBF 神经网络用于构建 4 个 Benchmark 函数拟合模型和农村生活污水处理过程 COD 在线软测量模型。仿真与应用结果表明,相对于其他神经网络模型,该模型具有较低的函数逼近误差和较高的 COD 预测精度。其中 COD 软测量模型训练结果的均方误差和平均绝对误差分别为 0. 18 和 0. 25,测试结果的两种误差分别为 0. 23 和 0. 36。

英文摘要:

With the goal to realize the real-time accurate measurement of chemical oxygen demand ( COD) in wastewater treatment process, a soft-measurement method based on TentFWA-GD RBF neural network (NN) was proposed. To solve the problems of network parameters settings and local optima existing in RBF NN based soft sensor modeling for complex industrial processes, as well as improve the model’s prediction precision and generalization ability, tent chaotic mapping was introduced in fireworks algorithm (FWA) to keep the population diversity and avoid the premature convergence by making use of the global ergodicity of chaos movement. Then a novel training method for RBF NN was proposed by combining the improved TentFWA with gradient descent (GD) method to enhance the learning ability. The TentFWA-GD RBF NN was applied to construct the fitting models of four Benchmark functions and the COD soft sensor model of rural domestic sewage treatment process. Simulation and application results showed that the model had lower function approximate error and higher COD prediction precision as compared with other neural network models. In COD soft sensor modeling, the mean square error and mean absolute error of the training results were 0. 18 and 0. 25, which of the test results were 0. 23 and 0. 36, respectively.

查看全文查看/发表评论下载PDF阅读器