Development and Application of Digital Process Operator System in Urban WWTPs
In the context of China's economic and social development transitioning towards greening and low-carbon transformation,wastewater treatment emerges as a critical domain for pollution prevention and greenhouse gas reduction,facing dual challenges of enhancing treatment efficiency and reducing carbon emissions.This study aimed to establish an intelligent digital process operator system driven by production data,employing a variety of technologies including statistics,fluid dynamics theory,and machine learning.Focusing on a specific wastewater treatment plant(WWTP),this research utilized the XGBOOST machine learning algorithm to develop a comprehensive sewage volume prediction model that takes into account factors such as upstream pump stations,regional temperature,and weather conditions.Based on the operational principles and historical data of the wastewater treatment plant,process parameter prediction models for various process segments were developed by integrating"mechanistic"and"mathematical"approaches using machine learning and fluid dynamics theories.A digital process operator system for the wastewater treatment plant,comprising three-tier architecture of production site layer,control layer,and decision layer,was constructed and applied in the WWTP.The research findings demonstrate that the digital process operator system enables intelligent scheduling,process parameter prediction,and other intelligent decision-making functions for the wastewater treatment plant.Compared to the same period of the previous year without the system,the removal rates of total nitrogen(TN),ammonia nitrogen,and total phosphorus(TP)have increased,with an average decrease of 4.98％in electricity consumption and 4.11％in chemical usage,achieving the goal of pollution and carbon reduction,while enhancing the digitalization,greening,and intelligent level of the WWTP.
digital process operatormachine learningintelligent schedulingprocess parameter predictionpollution and carbon reduction
万方数据
维普
