Open AccessInfluent property forecasting of sewage treatment based on big data analysis approach
Author(s) -
Zhiwei Cheng,
Zhihong Xie,
Yong Deng,
Xiaodan Wang
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/687/1/012183
Subject(s) - adaptive neuro fuzzy inference system , property (philosophy) , fuzzy inference system , computer science , data mining , artificial neural network , time series , fuzzy logic , big data , artificial intelligence , mathematical optimization , machine learning , mathematics , fuzzy control system , philosophy , epistemology
Precise influent property forecasting is very important to maintain the stable operation of sewage treatment procedure. A big data analysis method of combining the wavelet packet transform (WPT) and adaptive network-based fuzzy inference system (ANFIS) is reported to solve this problem. In this approach, the WPT is used to decompose the influent property data in different cycles. The time sub-series, which are results of wavelet coefficients reconstruction, are employed to establish the forecasting system. The forecasting sub-results of each cycle are eventually integrated into an overall forecasting result. Furthermore, chaos theory is introduced to obtain the input structure of the multi-cycle regression models. The reported approach is verified by the historical influent property. A back propagation neural network and the standard ANFIS are used for a comparison test. The results demonstrate that the reported method has best ability in the peer models.