State and parameter estimation for wastewater treatment processes using a stochastic model

Comprehensive Review of Control and Operational Strategies for Partial Nitration/ANAMMOX System

“…68,69 Due to the high complexity of the wastewater treatment mechanism and properties, it is necessary to know the model accurately to estimate the parameters. 70 Two-input, twooutput models of the system are available in literature to develop the model based control strategies 71,72 and decentralized control strategies. 14,71 Nonlinear models for such processes are described by refs 34 and 55.…”

“…Consequently, control strategies that involve measurements of these variables inevitably result in poor system performance. However, sensors for the measurement of ammonium and DO concentrations are very accurate and commonly used in the wastewater treatment plant. , Due to the high complexity of the wastewater treatment mechanism and properties, it is necessary to know the model accurately to estimate the parameters . Two-input, two-output models of the system are available in literature to develop the model based control strategies , and decentralized control strategies. , Nonlinear models for such processes are described by refs and .…”

Efficient Biogas Production through Process Simulation

“…8,[10][11][12][13][14] A smaller number of models uses various mathematical methods and control engineering theories, such as generic algorithms, fuzzy logic, neural networks, or stochastic methods. [15][16][17][18][19] A subgroup of models can be characterized by its specialization on particular substrates, [20][21][22] reactor types, [23][24][25] and specific modeling purposes, [26][27][28][29] respectively. Generally, one has to weigh high modeling complexity against the necessary effort for model parameterization.…”

“…Whereas some models − are of comparably basic nature with respect to the number of modeled state variables and process steps, other models simulate the biogas process almost entirely, including a wide variety of intermediate products. − Another distinction can be drawn with respect to modeling techniques. A large group of models is mainly based on mass and energy balances as well as different growth and product formation/degradation kinetics. ,− A smaller number of models uses various mathematical methods and control engineering theories, such as generic algorithms, fuzzy logic, neural networks, or stochastic methods. − A subgroup of models can be characterized by its specialization on particular substrates, − reactor types, − and specific modeling purposes, − respectively.…”