Healthy Research Rewards
ResearchHub is incentivizing healthy research behavior. At this time, first authors of open access papers are eligible for rewards. Visit the publications tab to view your eligible publications.
Got it
YC
Yan Chen
Author with expertise in On-line Monitoring of Wastewater Quality
Achievements
Open Access Advocate
Key Stats
Upvotes received:
0
Publications:
2
(50% Open Access)
Cited by:
0
h-index:
9
/
i10-index:
9
Reputation
Biology
< 1%
Chemistry
< 1%
Economics
< 1%
Show more
How is this calculated?
Publications
0

Learning a neural network-based soft sensor with double-errors parallel optimization towards effluent variable prediction in wastewater treatment plants

Dong Li et al.Aug 1, 2024
With the development of machine learning and artificial intelligence (ML/AI) models, data-driven soft sensors, especially the neural network-based, have widespread utilization for the prediction of key water quality indicators in wastewater treatment plants (WWTPs). However, recent research indicates that the prediction performance and computational efficiency are greatly compromised due to the time-varying, nonlinear and high-dimensional nature of the wastewater treatment process. This paper proposes a neural network-based soft sensor with double-errors parallel optimization to achieve more accurate prediction for effluent variables timely. Firstly, relying on the Activity Based Classification (ABC) principle, an ensemble variable selection method that combines Pearson correlation coefficient (PCC) and mutual information (MI) is introduced to select the optimal process variables as auxiliary variables, thereby reducing the data dimensionality and simplifying the model complexity. Subsequently, a double-errors parallel optimization methodology with minimizing both point prediction error and distribution error simultaneously is proposed, aiming to enhancing the training efficiency and the fitting quality of neural networks. Finally, the effectiveness is quantitatively assessed in two datasets collected from the Benchmark Simulation Model no. 1 (BMS1) and an actual oxidation ditch WWTP. The experimental results illustrate that the proposed soft sensor achieves precise effluent variable prediction, with RMSE, MAE and R