Dynamic multi-objective optimization control for wastewater treatment process under various operating conditions

Operating conditions in wastewater treatment processes (WWTP) are subject to dynamic changes. To enhance energy consumption (EC) and effluent quality (EQ) under varying operating conditions, this paper proposes the strategy of Dynamic Multi-Objective Optimization Control (DMOC). Firstly, the k-shape...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of machine learning and cybernetics Ročník 16; číslo 10; s. 8295 - 8309
Hlavní autoři: Deng, Xin, Xie, Xiaoyu, Huang, Linyu, Ning, Qian
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.10.2025
Springer Nature B.V
Témata:
Algorithms
Artificial Intelligence
Chemical oxygen demand
Complex Systems
Computational Intelligence
Control
Control systems design
Denitrification
Emission standards
Emissions
Energy consumption
Engineering
Greenhouse gases
Kalman filters
Kernel functions
Mechatronics
Multiple objective analysis
Nitrogen
Optimization algorithms
Original Article
Particle swarm optimization
Pattern Recognition
Robotics
Systems Biology
Variables
Wastewater treatment
Water quality
Multi-objective particle swarm optimization
Multiple operating conditions
Wastewater treatment process
Dynamic optimal control
K-shape clustering
ISSN:1868-8071, 1868-808X
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Operating conditions in wastewater treatment processes (WWTP) are subject to dynamic changes. To enhance energy consumption (EC) and effluent quality (EQ) under varying operating conditions, this paper proposes the strategy of Dynamic Multi-Objective Optimization Control (DMOC). Firstly, the k-shape algorithm is employed to categorize optimization cycles with similar influent parameters into the same operating condition. Subsequently, an adaptive kernel function is utilized to establish performance indicator models tailored to different operating conditions. Secondly, a Multi-Objective Particle Swarm Optimization algorithm based on classification memory (CM-MOPSO) is introduced to dynamically adapt to changes in performance indicators and preserve the optimal solution set for setpoints. This algorithm leverages past optimal solutions in similar environments to guide the evolution of the population during dynamic processes. The third aspect involves the design of a controller based on Extended Kalman Filtering (EKF) to effectively track the optimal setpoints. Finally, experiments are conducted on the Benchmark Simulation Model No. 1 (BSM1) to validate the proposed DMOC strategy. The obtained results demonstrate that DMOC not only optimizes EC and EQ simultaneously but also significantly enhances the control performance of the system.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1868-8071
1868-808X
DOI:10.1007/s13042-025-02723-9