基于LSTM模型的污水处理厂出水总氮预测研究

doi:10.3976/j.issn.1002-4026.20240010

山东科学 ›› 2024, Vol. 37 ›› Issue (6): 116-124.doi: 10.3976/j.issn.1002-4026.20240010

基于LSTM模型的污水处理厂出水总氮预测研究

余铭铨¹(), 师浩铭²^,^*()

1.中国电建集团华东勘测设计研究院有限公司,浙江杭州 311122
2.浙江大学建筑工程学院,浙江杭州 310058

收稿日期:2024-01-16 出版日期:2024-12-20 发布日期:2024-12-05
通信作者: *师浩铭,男,硕士研究生在读,研究方向为基于机器学习的污水脱氮研究。E-mail:22112063@zju.edu.cn, Tel:18013962592
作者简介:余铭铨(1983—),男,硕士研究生,研究方向为环境工程设计与咨询、工程项目管理。E-mail:185700956@qq.com

Prediction of effluent total nitrogen in wastewater treatment using LSTM neural network

YU Mingquan¹(), SHI Haoming²^,^*()

1. Power China Huadong Engineering Corporation Limited, Hangzhou 311122, China
2. College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China

Received:2024-01-16 Online:2024-12-20 Published:2024-12-05

摘要/Abstract

摘要：

出水总氮质量浓度是评价污水处理厂生物脱氮效果的关键指标之一。为解决污水厂总氮排放易超标的问题,提出了一个基于长短期记忆网络(LSTM)的出水总氮实时预测模型。利用皮尔逊相关性分析来确定模型输入,并通过网格搜索算法优化模型超参数。将得到的LSTM模型应用于重庆市某实际污水处理厂预测出水总氮,并与传统的时序模型作对比,验证了该模型的可行性。结果表明: LSTM模型能够较好地预测出水总氮,其预测值与实际值的平均绝对误差为0.911 mg/L,均方根误差为1.074 mg/L,平均绝对百分比误差为11.28%,各项指标均优于循环神经网络(RNN)模型和自回归差分移动平均(ARIMA)模型。这一模型的构建可以为出水总氮的高效监测提供帮助。

关键词: LSTM模型, 皮尔逊相关性, 网格搜索算法, 出水总氮

Abstract:

The effluent total nitrogen (TN) is one of the key indicators for assessing the biological denitrification performance of wastewater treatment plants(WWTPs). To mitigate the prevalent issue of excessive TN discharges from WTTPs, we proposed a real-time prediction model based on long short-term memory (LSTM) networks. We performed Pearson correlation analysis to determine model inputs and used grid search algorithm to optimize model hyperparameters. Then, we used the proposed model to predict the actual effluent TN in a WWTP in Chongqing and compared its predictive performance with that of traditional time-series models. Results indicate that the proposed model can effectively predict effluent TN with an average absolute error of 0.911 mg/L, an average root mean square error of 1.074 mg/L, and an average absolute percentage error of 11.28%. All of these performance indicators surpass those of the recurrent neural network and ARIMA models. The proposed model can serve as the foundation for effective monitoring of effluent TN.

Key words: long short-term memory model, Pearson correlation, grid search algorithm, effluent total nitrogen

中图分类号:

TU992

余铭铨, 师浩铭. 基于LSTM模型的污水处理厂出水总氮预测研究[J]. 山东科学, 2024, 37(6): 116-124.

YU Mingquan, SHI Haoming. Prediction of effluent total nitrogen in wastewater treatment using LSTM neural network[J]. Shandong Science, 2024, 37(6): 116-124.

图/表 7

图1

图2

图3

图4

图5

图6

表1

参考文献 23

[1]	孙国庆. 智慧水务关键技术研究及应用[J]. 水利信息化, 2018(1): 46-49. DOI: 10.19364/j.1674-9405.2018.01.010.
[2]	蒋彬, 刘中亚, 陈垚, 等. 碳中和视角下污水处理现状与展望[J]. 工业水处理, 2022, 42(6): 51-58. DOI: 10.19965/j.cnki.iwt.2021-0905.
[3]	周青龄, 周琳, 桂双林. MBR工艺在生活污水处理中同步脱氮除磷效果[J]. 能源研究与管理, 2019(3): 36-38. DOI: 10.16056/j.1005-7676.2019.03.010.
[4]	李佟, 李军. 基于BP神经网络与马尔可夫链的污水处理厂脱氮效果模拟预测[J]. 环境科学学报, 2016, 36(2): 576-581. DOI: 10.13671/j.hjkxxb.2015.0559.
[5]	林佳敏, 陈金良, 林晶晶, 等. BP神经网络和ARIMA模型对污水处理厂出水总氮浓度的模拟预测[J]. 环境工程技术学报, 2019, 9(5): 573-578. DOI: 10.12153/j.issn.1674-991X.2019.03.261.
[6]	夏文泽, 冯骁, 王喆, 等. 基于新型联合循环神经网络(RNN)模型的出水总氮预测[J]. 净水技术, 2021, 40(8): 107-113. DOI: 10.15890/j.cnki.jsjs.2021.08.015.
[7]	崔玉波, 张万筠, 孙红杰, 等. 污泥处理湿地系统渗滤液中总氮和氨氮的预测模型[J]. 中国给水排水, 2015, 31(17): 72-75. DOI: 10.19853/j.zgjsps.1000-4602.2015.17.017.
[8]	黄学平, 吴留兴, 辛攀, 等. 基于3种机器学习模型的污水处理厂出水总氮预测分析[J]. 能源研究与管理, 2023, 15(2): 100-105. DOI: 10.16056/j.2096-7705.2023.02.015.
[9]	ZHAO Z H, WANG Z H, YUAN J L, et al. Development of a novel feedforward neural network model based on controllable parameters for predicting effluent total nitrogen[J]. Engineering, 2021, 7(2): 195-202. DOI: 10.1016/j.eng.2020.07.027.
[10]	HANSEN L D, STOKHOLM-BJERREGAARD M, DURDEVIC P. Modeling phosphorous dynamics in a wastewater treatment process using Bayesian optimized LSTM[J]. Computers & Chemical Engineering, 2022, 160: 107738. DOI: 10.1016/j.compchemeng.2022.107738.
[11]	PISA I, SANTIN I, MORELL A, et al. LSTM-based wastewater treatment plants operation strategies for effluent quality improvement[J]. IEEE Access, 2019, 7: 159773-159786. DOI: 10.1109/access.2019.2950852.
[12]	国家环境保护总局科技标准司. 城镇污水处理厂污染物排放标准: GB 18918—2002[S]. 北京: 中国标准出版社, 2002.
[13]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. DOI: 10.1162/neco.1997.9.8.1735. pmid: 9377276
[14]	SHAMS M Y, ELSHEWEY A M, EL-KENAWY E S M, et al. Water quality prediction using machine learning models based on grid search method[J]. Multimedia Tools and Applications, 2024, 83(12): 35307-35334. DOI: 10.1007/s11042-023-16737-4.
[15]	CHEN H L, CHANG X F. Photovoltaic power prediction of LSTM model based on Pearson feature selection[J]. Energy Reports, 2021, 7: 1047-1054. DOI: 10.1016/j.egyr.2021.09.167.
[16]	李如仁, 孙加瑶. 融合SBAS-InSAR与GS-LSTM的尾矿库沉降监测与预测[J]. 金属矿山, 2023(1): 102-109. DOI: 10.19614/j.cnki.jsks.202301011.
[17]	ZHANG Y T, LI C L, DUAN H P, et al. Deep learning based data-driven model for detecting time-delay water quality indicators of wastewater treatment plant influent[J]. Chemical Engineering Journal, 2023, 467: 143483. DOI: 10.1016/j.cej.2023.143483.
[18]	CHEN J C, CHANG N B, SHIEH W K. Assessing wastewater reclamation potential by neural network model[J]. Engineering Applications of Artificial Intelligence, 2003, 16(2): 149-157. DOI: 10.1016/s0952-1976(03)00056-3.
[19]	FARAMARZ B, MOHAMAD-JAVAD M, MILAD B, et al. Comparative study on total nitrogen prediction in wastewater treatment plant and effect of various feature selection methods on machine learning algorithms performance[J]. Journal of Water Process Engineering, 2021, 41:102033.
[20]	CHO K H, KIM J O, KANG S, et al. Achieving enhanced nitrification in communities of nitrifying bacteria in full-scale wastewater treatment plants via optimal temperature and pH[J]. Separation and Purification Technology, 2014, 132: 697-703. DOI: 10.1016/j.seppur.2014.06.027.
[21]	汤琪. 生物脱氮除磷新技术[J]. 重庆大学学报(自然科学版), 2006, 29(9): 138-143.
[22]	SHRESTHA N. Detecting multicollinearity in regression analysis[J]. American Journal of Applied Mathematics and Statistics, 2020, 8(2): 39-42. DOI: 10.12691/ajams-8-2-1.
[23]	KONTOPOULOU V I, PANAGOPOULOS A D, KAKKOS I, et al. A review of ARIMA vs. machine learning approaches for time series forecasting in data driven networks[J]. Future Internet, 2023, 15(8): 255. DOI: 10.3390/fi15080255.

模型	δ_MAE/(mg·L^-1)	δ_RMSE/(mg·L^-1)	δ_MAPE/%
ARIMA	2.853	1.788	29.06
RNN	1.094	1.188	13.45
LSTM	0.911	1.074	11.28

基于LSTM模型的污水处理厂出水总氮预测研究

Prediction of effluent total nitrogen in wastewater treatment using LSTM neural network

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 23

相关文章 1

Metrics

本文评价

推荐阅读 0