Search

This paper presents an approach for solving minimal time problems for biological batch models with switching controls. Multi-reaction processes operated in a sequence of operating modes are considered, where the control variable is the switch between different operating modes. In order to determine the optimal switching instants that minimize the final time to reach the target, a reachability analysis and a study of dynamical properties are first investigated. This leads to a reduced model and the characterization of reachable switching sequences. Optimization tools are then used to determine the optimal switching instants. Copyright © 2009 John Wiley & Sons, Ltd.

Three types of wastewater were considered: typical municipal wastewater and the effluent of a dairy industry, both polluted with organic carbon and nitrogen, and wastewater typical of a chemical industry, containing toxic or recalcitrant compounds (4-chlorophenol). Sequencing batch reactors (SBRs) were considered. Process experiments were performed to provide experimental data for parameter identification of the dynamical models of the SBRs. The processes were instrumented with sensors to measure O2, NH3, T, pH, OUR, and other off line measurements (e.g.: NO2 − and NO3 −, COD, phenol concentration, microbiological counts and activities) in order to study more accurately the behaviour of the biomasses. However, the general idea is to design a system with a minimal number of sensors.

This paper discusses several theoretical as well as practical challenging issues related to a recently developed time-optimal control strategy for the optimization of oxic/anoxic Sequencing Batch Reactors (SBR) for the removal of organic carbon and nitrogen from urban and industrial wastewaters and to its practical implementation.

International audience; Nitrification is usually the bottleneck of biological nitrogen removal processes. In SBRs systems, it is not often enough to monitor dissolved oxygen, pH and ORP to spot problems which may occur in nitrification processes. Therefore, automated supervision systems should be designed to include the possibility of monitoring the activity of nitrifying populations. Though the applicability of set-point titration for monitoring biological processes has been widely demonstrated in the literature, the possibility of an automated procedure is still at its early stage of industrial development. In this work, the use of an at-line automated titrator named TITAAN (TITrimetric Automated ANalyser) is presented. The completely automated sensor enables us to track nitrification rate trend with time in an SBR, detecting the causes leading to slower specific nitrification rates. It was also possible to perform early detection of toxic compounds in the influent by assessing their effect on the nitrifying biomass. Nitrifications rates were determined with average errors +/- 10% (on 26 tests), never exceeding 20% as compared with UV-spectrophotometric determinations.

International audience; The objective is the adaptation of an existing UV spectrophotometer for the on-line monitoring of nitrate, nitrite and soluble COD during cycles of SBRs. An on-line filtering system was also developed in order to provide particle-free fluids to the sensor. The modified UV-spectrophotometer sensor (called "STAC") was implemented into a pilot-scale bioreactor at INRA-Narbonne, treating a semi-synthetic dairy wastewater. After a calibration and validation steps, results obtained during five months have shown that the on-line spectrophotometer sensor was able to efficiently monitor on-line nitrate, nitrite in SBR content. Ranges of detection were between 0 to 40 mg/L for nitrate and 0 to 18 mg/L for nitrite, with an accuracy of about 10%. The on-line estimation of soluble COD was less efficient. The major drawback was the time-consumption for the cleaning of the homemade on-line filter.

This paper presents (i) a synthesis of the EOLI supervision system for optimizing both the biological reaction phases and the settling phase of Sequencing Batch Reactors (SBR) performing carbon and nitrogen removal from urban or industrial wastewaters; (ii) a summary of the integration capabilities of the integrated system developed by an industrial partner of the project to propose both the hardware and the software of the EOLI solution under a unified framework.

This paper presents two optimal control strategies developed for the control of two biological sequencing batch reactors (SBR), the first one performing both organic carbon and nitrogen removal from dairy wastewaters, while the second one treats toxic wastewaters. The proposed control laws are based on models that have been validated with real data acquired on two real biological SBR pilot plants. These models are presented in D. Mazouni et al., 16th IFAC World Congress, Prague, 5–8 July 2005. The validation results of these control strategies are presented and discussed.

This paper presents a new software developed in MATLAB for analyzing on-line data of an aerobic SBR, detecting faults and, in this case, proposing the most probable causes of fault. Process diagnosis is achieved using a statistical method divided in two main phases: off-line model building and on-line data diagnosis. The off-line model identifies the correct working conditions of the system (standard operative conditions). It includes the characterization of the deviation of the system from these standard conditions in the case of changing in the biomass properties or carbon and nitrogen load characteristics. The on-line diagnosis aims at collecting and analyzing all the available data available through industrial sensors, and at classifying the behavior of each treatment cycle. The diagnosis performance of the proposed method is tested using a data set of an aerobic SBR pilot plant.

This paper presents i) an extension of a simple model describing the dynamical behavior of important variables during the aerobic phase of a biological wastewater treatment plant and ii) a simple asymptotic observer based on this extended model for online estimating both soluble biodegradable carbon and nitrate+nitrite nitrogen concentrations. The proposed model is identified and the observer is validated with real data acquired on a real biological Sequencing Batch Reactor pilot plant.

In this paper, geometric tools are used for the attainability study of biological Sequencing Batch Reactors (SBR). It consisted in finding the set of initial conditions from which one can reach a final state target with admissible control. The possible solutions can be determined easily once the set of reachability is fixed. In the second step, optimal time control problem is considered. It consist in finding the switching instant between the different steps of the batch. Pontryagin's Maximum principle is used to solve a part of this problem. We propose asuboptimal solution allowing at most two commutation. A new problem is considered where the criteria is parameterized by the switching concentrations. A numerical solution is then proposed to solve this new minimal time problem.