Stefano Marsili-Libelli

A modified Activated Sludge Model No. 3 (ASM3) with two-step nitrification-denitrification

A common limitation of the Activated Sludge Models (ASM) [Henze, M., Gujer, W., Mino, T., van Loosdrecht, M.C.M., 2000. Activated Sludge Models ASM1, ASM2, ASM2d, and ASM3. IWA Scientific and Technical Report No. 9. IWA Publishing, London, UK] is the representation of the nitrification dynamics as a single-step process and the consequent denitrification on nitrate alone. This generally acknowledged simplification may represent a serious limitation in specific applications where nitrites become important, either as a target final product or an unwanted intermediate. This paper proposes an enhancement to the basic ASM3 model, introducing a two-step model for the process nitrification and, consequently, considering denitrification on both nitrite and nitrate. After introducing the relevant process kinetics and adapting the stoichiometric matrix accordingly, the model implementation in the Matlab/Simulink(TM) platform is described with reference to the benchmark setting. To obtain a fast implementation, the process units (reaction tanks and secondary settler) have been implemented as DLLs linked to the Simulink blocks, whereas the model parameters and stoichiometric matrix remain accessible to the user. The new model is compared with the standard ASM3 and checked for consistency and mass conservation. It is also shown that with the default kinetic parameters nitrite may represent a considerable fraction of the nitrified effluent, thus revealing a design limitation in the benchmark sizing. In the last part, an optimization of the benchmark plant volumes has been attempted in order to minimize such violations, resulting in a moderate increase of the overall reaction volume. The pertinent software is freely available for research purposes.

Confidence regions of estimated parameters for ecological systems

The parameters of ecological models are usually estimated through numerical search algorithms. Determining confidence boundaries for the parameter values obtained in such a way is a problem of great practical importance. In this paper a method is proposed to estimate such regions in two ways, based on either the Hessian matrix or the Fisher Information Matrix (FIM). There is a conceptual difference in the two approximations: the FIM approach is based on the sensitivity trajectories, whereas the Hessian expansion depends on the shape of the error functional. From a comparison between the two approaches, a discriminating method is obtained to detect inaccurate estimation results. The Hessian and FIM approaches differ by the second derivative terms of the output function. This difference is used to assess the success of the estimation, because the two methods yield the same confidence estimate only if the search terminates at the optimal parameter value. The method is demonstrated with reference to a pair of widely used dynamics: the Monod kinetics and the Richards logistic function applied to algal growth. It is shown that in both cases this method compares favourably with the residual correlation analysis and appears to have more discriminatory power.

Nitrogen removal via nitrite in a sequencing batch reactor treating sanitary landfill leachate

The present paper reports the results of the application of a control system, based on artificial intelligence concepts, for the automation of a bench-scale SBR treating leachate generated in old landfills. Attention was given to the nitritation and denitritation processes in order to enhance the nitrogen removal efficiency. Nitrification and nitrogen removal were usually higher than 98% and 95%, respectively, whereas COD removal was approximately 20-30% due to the low biodegradability of organic matter in the leachate from old landfills; therefore, external COD was added to accomplish the denitrification process. Adjusting the length of the oxic phase, almost complete inhibition of the nitrite oxidizing organisms was observed. The results confirm the effectiveness of the nitrite route for nitrogen removal optimisation in leachate treatment. A significant saving of approximately 35% in external COD addition was achieved.

Fuzzy control of disturbances in a wastewater treatment process

This paper describes a real-time process control scheme to cope with the problem of input disturbances in wastewater treatment processes, based on a fuzzy inferential control system. This can detect the presence of a dangerous input conditions whenever either organic overload or inhibitory/toxic compounds are present in the raw wastewater. Based on this diagnosis, a set of fuzzy rules are implemented to divert the process flow and bring the system back into a safe state. The control system was designed and tested using a pilot plant, to which a toxic disturbance was applied. Its behaviour with and without fuzzy control are compared, and the process reliability improvement in the controlled case is demonstrated.

Parameter estimation of ecological models

Abstract Determining the parameter values of ecological models poses special problems caused by nonlinear model struture, parameter correlation and data scarcity. All these aspects justify a numerical approach, which is pursued here through an efficient numerical optimization procedure based on the flexible polyhedron search. This paper describes the algorithm and its application to the estimation of some well-known microbial kinetics encountered in many applications in ecology and biotechnology. After assessing the main features of the algorithm, an estimate of the parameter covariance matrix is derived and related to system sensitivity. In the last part of the paper extensive numerical tests are discussed to assess the algorithm performance in dealing with noisy measurements. The numerical results show that the method is computationally robust and performs well even with heavily noise-corrupted data. The algorithm was implemented using Borlands Turbo BASIC and the source code is available for research purposes.

Adaptive fuzzy pattern recognition in the anaerobic digestion process

This paper addresses the problem of coordinating the operation of several units of a biological wastewater treatment process in order to avoid process failure. The state of the process is monitored through a number of chemical and biological variables and this information is used to infer the plant loading state. After a brief introduction of the plant supervision scheme, the paper describes how the Fuzzy C-Means (FCM) classification algorithm can be used in this application and enhanced with an adaptive capability, enabling the supervisory control system to detect process departure from normal operating conditions.

Water quality modelling for small river basins

Water quality modelling in small rivers is often considered unworthy from a practical and economic viewpoint. This paper shows instead that a simple model structure can be set up to describe the stationary water quality in small river basins in terms of carbon and nitrogen compounds, when the use of complex models is unfeasible. In short rivers point and nonpoint sources play a key role in shaping the model response, being as important as the self-purification dynamics. Further, the varying river characteristics, in terms of morphology, hydraulics and vegetation, require the introduction of variable parameters, thus complicating the originally simple model structure. To determine the identifiability of the resulting model an identifiability assessment was carried out, based on sensitivity analysis and optimal experiment design criteria. The identifiable subset was determined by ranking the parameters in terms of sensitivity and computing the associated Fisher Information Matrices. It was found that the inclusion of the nonpoint sources as piecewise constant parameters affected the identifiability to a considerable extent. However, the combined parameter-sources calibration was made possible by the use of a robust estimation algorithm, which also provided estimation confidence limits. The calibrated model responses are in good agreement with the data and can be used as scenario generators in a general strategy to conserve or improve the water quality.

Fuzzy Prediction of the Algal Blooms in the Orbetello Lagoon

Abstract The Orbetello lagoon is a shallow brackish waterbody subject to intense and diverse eutrophication (phytoplankton, macroalgae and macrophytes). Periodically a large amount of algae must be artificially removed, their collection and disposal representing a considerable management cost. This paper describes the design of a bloom predictor based on the daily fluctuations of simple water quality parameters such as dissolved oxygen, oxidation–reduction potential, pH and temperature. The task of the fuzzy predictor is to recognise the possibility that a bloom of the macroalgae population is about to occur, based on the changing daily pattern of these variables. The fuzzy predictor is based on a set of fuzzy rules derived from experimental observations and expert knowledge. A 15 month period of hourly data was analysed and used to form the initial knowledge-base. The tests show that the inferential engine has good predictive capabilities, which could be improved when more data become available. The algorithm outlined in the paper has enough generality to be applied to any environmental classification problem where short-term time patterns and long-term statistical variability should be considered jointly.

Optimisation of sanitary landfill leachate treatment in a sequencing batch reactor

A bench-scale SBR was operated for almost three years in an attempt to optimise the treatment of leachates generated in old landfill. The results of the first two years were used to design a monitoring and control system based on artificial intelligence concepts. Nitrogen removal was optimized via the nitrite shortcut. Nitrification and N removal were usually higher than 98% and 90%, respectively, whereas COD (of the leachate) removal was approximately 30-40%. The monitoring and control system was demonstrated capable of optimizing process operation, in terms of phase length and external COD addition, to the varying loading conditions. Using the control system developed, a significant improvement of the process was obtained: COD and N load were increased (HRT decrease) and a significant decrease (approximately 34%) of the ratio of COD added to N leachate content was observed.

Shock load modelling in the anaerobic digestion process

Abstract A simplified mathematical model for the behaviour of anaerobic digesters under shock loading conditions is derived with a special emphasis on bicarbonate alkalinity. This study was motivated by the development of a novel bicarbonate alkalinity measuring device providing on-line information of this variable, which is particularly well suited for the early detection of organic overloading. The model includes a set of dynamic equations and an extended ion balance to describe the effect of alkali, volatile fatty acids and carbon dioxide. After the structural description, the model is calibrated using specific shock experimental data and its implications are discussed in terms of a variety of shock conditions. Lastly, the feasibility of bicarbonate dosing control is assessed through proportional-integral-derivative (PID) controller design and model simulation.