APPLICATION OF PREDICTION MODEL BASED ON RECURRENT NEURAL NETWORK IN MEMBRANE FLUX

Xinchang Wang, Chunqing Li

Abstract


In the process of sewage treatments using membrane bio-reactor, solid large particle suspensions and dissolved substances are continuously deposited on the membrane surface, which is a result of the formation of cake layer. As the thickness of cake layer continues to increase, the flux of the membrane continues to decline, which leads to membrane fouling. The degree of pollution directly determines the efficiency of wasted-water treatment, and so monitoring the membrane flux value to determine the degree of membrane contamination is the common method for the sewage treatment factories. In a certain period of time, using the existing membrane flux data to predict the membrane flux in the future, making the decision to clean or change the membrane at the most appropriate moment will save costs and energy. Recurrent neural networks have features that are sensitive to historical data because of their hidden-layer ring structure, which increases the correlation between data. Combined with the particle swarm optimization, it can obtain the characteristics of the global optimal value and optimize the initial weight of the recurrent neural network, which can reduce the number of iterations and obtain prediction results quickly. Based on this, establishing prediction model using particle swarm optimization to optimize recurrent neural networks to determine membrane fouling in this paper. Simulation results show that the prediction model proposed in this paper is better than recurrent neural network and traditional BP neural network.

Keywords


membrane fouling; recurrent neural network; prediction model; particle swarm optimization

Full Text:

PDF

References


Li Song, Liu Lijun, Liu Yingpeng. Chaotic time series prediction of improved PSO optimization BP neural network[J]. Computer Engineering and Applications, 2013, 49(06): 245-248+270.

ZHANG Hai, PU Jian, ZHANG Xiaocheng. Condenser vacuum model of Elman neural network based on particle swarm optimization algorithm[J]. Thermal Power Generation, 2010, 39(04): 7-11+35.

FAN Xiangxiang, LI Qi, ZHU Yajie, HOU Junxiong, FENG Jun. Research on Spatio-temporal Forecasting Model of Air Pollution Based on RNN[J]. Surveying Science, 2017, 42(07): 76-83+120.

Sutskever I, Martens J, Hinton G E. Generating Text with Recurrent Neural Networks[C]// International Conference on Machine Learning, ICML 2011, Bellevue, Washington, Usa, June 28 - July. DBLP, 2011:1017-1024.

Sutskever I, Vinyals O, Le Q V. Sequence to Sequence Learning with Neural Networks[J]. 2014, 4:3104-3112.

Pascanu R, Mikolov T, Bengio Y. On the difficulty of training Recurrent Neural Networks[J]. 2012, 52(3):III-1310.

Cho K, Merrienboer B V, Gulcehre C, et al. Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation[J]. Computer Science, 2014.

Pu Jian, Cheng Jing Based Turbine PSO-Elman neural network to predict the exhaust enthalpy calculation [J]. Nanjing Normal University (Engineering and Technology),2009,9(02):35-40.

Yu Jian, GUO Ping.Effects Elman neural network for network traffic prediction [J]. Computer Engineering and Design,2008,(17):4531-4534.


Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2018 International Educational Applied Scientific Research Journal