A DATA ENHANCEMENT METHOD FOR SUPPORTING INTELLIGENT MANAGEMENT OF WWTPs UNDER DATA DEFICIENCY CONDITIONS
The intelligent management of wastewater treatment plants(WWTPs)requires high-quality and rich data support.However,in the current management mode of WWTPs,issues such as excessive aeration,excessive dosing,and monitoring have led to deficiencies in the quantity and quality of operation and maintenance data.The performance of various data-driven models supported by such defective data is not sufficient.How to improve the quality and quantity of data is crucial for the research and application of various artificial intelligence models.This study proposed a wastewater treatment data enhancement method based on generative adversarial networks(WP-GAN)to address the data defects,and validated the proposed method using a classic wastewater treatment artificial neural network model(W-ANN).The dataset used in the study was from a large urban WWTP with an anaerobic-anoxic-aerobic(A20)process.The measured data was amplified 5-fold through data enhancement treatment,and the W-ANN model was trained with pre-and post-enhancement data samples.It was found that the performance of the W-ANN model was significantly improved:the fitting degree increased from 20％to 65％,and the maximum simulation accuracy increased from 67.85％to 75.55％.The method proposed in this study is a universal data enhancement method to address data deficiencies,which can provide better data support for various data-driven models for the intelligent management of WWTPs.
data enhancementwastewater treatmentgenerative adversarial networksintelligent managementdata-driven models
NETL
NSTL
万方数据
