BAN Huilin,LI Zhongzhi,LI Binyong,et al.Soft Sensing Method of BOD5 in Sewage based on FRBPSO-RBF Neural Network[J].Journal of Chengdu University of Information Technology,2024,39(04):416-421.[doi:10.16836/j.cnki.jcuit.2024.04.004]
基于FRBPSO-RBF神经网络的污水BOD5软测量方法
- Title:
- Soft Sensing Method of BOD5 in Sewage based on FRBPSO-RBF Neural Network
- 文章编号:
- 2096-1618(2024)04-0416-06
- Keywords:
- radial basis function neural network; particle swarm optimization; soft sensing model; 5-day BOD biochemical oxygen demand soft sensing; sewage quality prediction
- 分类号:
- TP391
- 文献标志码:
- A
- 摘要:
- 污水处理过程中污水BOD5难以实时准确测量,故软测量方法逐渐被用于污水BOD5的预测,其中RBF神经网络软测量方法应用广泛,但存在训练过程易陷入局部极值等问题。为提高RBF神经网络的预测精度,提出了基于适应度排名的粒子群算法(fitness ranking based particle swarm optimization,FRBPSO),根据适应度排名与迭代次数确定惯性权重的大小,并根据粒子个体历史最优值的排名与迭代次数确定自我学习因子与社会学习因子的大小,并将FRBPSO算法引入RBF神经网络的参数训练中。基于13个基准测试函数与其他3个粒子群优化算法对比,实验结果显示FRBPSO算法的寻优能力相对较强。再将基于FRBPSO算法的RBF神经网络用于构建污水BOD5软测量模型,仿真结果表明,在测试数据中,FRBPSO-RBF软测量模型的平均绝对误差比PSO-RBF软测量模型、DAIW-RBF软测量模型、SCVPSO-RBF软测量模型分别降低了0.7178、0.2402、0.5851,平均绝对百分比误差分别降低了0.47%、0.15%、0.33%,均方根误差分别降低了0.0034、0.0015、0.0039。与其他3个基于PSO算法的BOD5软测量模型相比,FRBPSO-RBF模型具有较高的BOD5预测精度。
- Abstract:
- During the wastewater treatment process, it is difficult to accurately measure the BOD5 of wastewater in real time. Therefore, soft measurement methods are gradually being used to predict BOD5 in wastewater. Among them, the RBF neural network soft measurement method is widely used, but there are problems such as the training process being easily trapped in local extremes. In order to improve the prediction accuracy of the RBF neural network, a fitness ranking based particle swarm optimization algorithm(FRBPSO)is proposed. The size of the inertia weight is determined according to the fitness ranking and the number of iterations, and the size of the self-learning factor and social learning factor are determined according to the ranking of the individual’s historical optimal value and the number of iterations. Then, the FRBPSO algorithm is introduced into the parameter training of the RBF neural network. Based on 13 benchmark test functions and comparison with other 3 particle swarm optimization algorithms, experimental results show that the FRBPSO algorithm has relatively strong optimization ability. Then, the RBF neural network based on FRBPSO algorithm is used to construct a soft measurement model for wastewater BOD5. Simulation results show that in test data, compared with PSO-RBF soft measurement model, DAIW-RBF soft measurement model and SCVPSO-RBF soft measurement model, the average absolute error of FRBPSO-RBF soft measurement model decreased by 0.7178, 0.2402 and 0.5851 respectively; The average absolute percentage error decreased by 0.47%, 0.15%, and 0.33% respectively; The root mean square error decreased by 0.0034, 0.0015 and 0.0039 respectively. Compared with other three BOD5 soft measurement models based on PSO algorithm, FRBPSO-RBF model has higher BOD5 prediction accuracy.
参考文献/References:
[1] 张钧,杨文武.水和废水中五日生化需氧量测定全程质量控制[J].环境科学与管理,2011,36(7):139-145.
[2] 黄道平,刘乙奇,李艳.软测量在污水处理过程中的研究与应用[J].化工学报,2011,62(1):1-9.
[3] Perera Yasith S,Ratnaweera D A A C,Dasanayaka Chamila H,et al.The role of artificial intelligence-driven soft sensors in advanced sustainable process industries:A critical review[J/OL].https://doi.org/10.1016/j.engappai.2023.105988,2023-03-21.
[4] Wang Gongming,Jia Qingshan,Zhou Mengchu,et al.Artificial neural networks for water quality soft-sensing in wastewater treatment: a review[J]. Artificial Intelligence Review,2021,55(1):565-587.
[5] Qiao Junfei,Hu Zhiqiang,Li Wenjing.Soft Measurement Modeling Based on Chaos Theory for Biochemical Oxygen Demand(BOD)[J].Water,2016,8(12):581-581.
[6] 丁海旭,李文静,叶旭东,等.基于自组织递归模糊神经网络的BOD软测量[J].计算机与应用化学,2019,36(4):331-336.
[7] 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.
[8] 谢俊标,江峰,杜军威,等.基于改进人工鱼群算法与RBF神经网络的股票价格预测[J].计算机工程与科学,2022,44(11):2080-2090.
[9] 吴学梅,牟莉.基于改进RBF-BP神经网络的预测方法研究[J].国外电子测量技术,2022,41(2):105-110.
[10] Kennedy J,Eberhart R.Particle Swarm Optimization[C].Proceedings of ICNN’95-International Conference on Neural Networks.IEEE,1995:1942-1948.
[11] 徐浩天,季伟东,孙小晴,等.基于正态分布衰减惯性权重的粒子群优化算法[J].深圳大学学报:理工版,2020,37(2):208-213.
[12] 王生亮,刘根友.一种非线性动态自适应惯性权重PSO算法[J].计算机仿真,2021,38(4):249-253.
[13] Zheng Kaitong,Yuan Xianfeng,Xu Qingyang,et al.Hybrid particle swarm optimizer with fitness-distance balance and individual self-exploitation strategies for numerical optimization problems[J].Information Sciences,2022,608:424-452.
[14] Zhao Shicun,Wang Da.Elite-ordinary synergistic particle swarm optimization[J].Information Sciences,2022,609:1567-1587.
[15] Yu Fei,Tong Lei,Xia Xuewen.Adjustable driving force based particle swarm optimization algorithm[J].Information Sciences,2022,609:60-78.
[16] Moody J,Darken C.Learning with localized receptive fields[C].Proceedings of the connectionist models summer school.Yale Univ,1988:133-143.
[17] SHI Y,EBERHART RC.Parameters selections in particle swarm optimization[C].Evolutionary Programming VII,1998:591-600.
[18] Yao Xin,Liu Yong,Lin Guangming. Evolutionary programming made faster.[J].IEEE Trans.Evolutionary Computation,1999,3(2):82-102.
[19] 陈博文,邹海.总结性自适应变异的粒子群算法[J].计算机工程与应用,2022,58(8):67-75.
备注/Memo
收稿日期:2023-05-25
基金项目:四川省科技计划资助项目(2021JDRC0046)
通信作者:李中志.E-mail:lizz@cuit.edu.cn