A mainstream viewpoint on sensitivity analysis in the wastewater modelling community is that it is a first-order differential analysis of outputs with respect to the parameters – typically obtained by perturbing one parameter at a time with a small factor. An alternative viewpoint on sensitivity analysis is related to uncertainty analysis, which attempts to relate the total uncertainty in the outputs to the uncertainty in the inputs. In this paper we evaluate and discuss two such sensitivity analysis methods for two different purposes/case studies: (i) Applying sensitivity analysis to a plant design (BSM1 plant layout) using Standardized Regression Coefficients (SRC) and (ii) Applying sensitivity analysis to help fine-tuning a fuzzy controller for a BNPR plant using Morris Screening. The results obtained from each case study are then critically discussed in view of practical applications of sensitivity analysis in day-to-day engineering projects. Abstract

Well-planned urban infrastructure should meet critical loads during its design lifetime. In order to proceed with design, engineers are forced to make numerous assumptions with very little supporting information about the development of various drivers. For the wastewater sector, these drivers include the future amount and composition of the generated wastewater, effluent requirements, technologies, prices of inputs such as energy or chemicals, and the value of outputs produced such as nutrients for fertilizer use. When planning wastewater systems, there is a lack of methods to address discrepancies between the timescales at which fundamental changes in these drivers can occur, and the long physical life expectancy of infrastructure (on the order of 25–80 years). To explore these discrepancies, we take a hindsight perspective of the long-term development of wastewater infrastructure and assess the stability of assumptions made during previous designs. Repeatedly we find that the drivers influencing wastewater loads, environmental requirements or technological innovation can change at smaller timescales than the infrastructure design lifetime, often in less than a decade. Our analysis shows that i) built infrastructure is continuously confronted with challenges it was not conceived for, ii) significant adaptation occurs during a structure's lifetime, and iii) “muddling-through” is the pre-dominant strategy for adaptive management. As a consequence, we argue, there is a need to explore robust design strategies which require the systematic use of scenario planning methods and instruments to increase operational, structural, managerial, institutional and financial flexibility. Hindsight studies, such as this one, may inform the development of robust design strategies and assist in the transition to more explicit forms of adaptive management for urban infrastructures.

Successful modeling of wastewater treatment plant (WWTP) processes requires an accurate description of the plant hydraulics. Common methods such as tracer experiments are difficult and costly and thus have limited applicability in practice; engineers are often forced to rely on their experience only. An implementation of grammar-based genetic programming with an encoding to represent hydraulic reactor models as program trees should fill this gap: The encoding enables the algorithm to construct arbitrary reactor models compatible with common software used for WWTP modeling by linking building blocks, such as continuous stirred-tank reactors. Discharge measurements and influent and effluent concentrations are the only required inputs. As shown in a synthetic example, the technique can be used to identify a set of reactor models that perform equally well. Instead of being guided by experience, the most suitable model can now be chosen by the engineer from the set. In a second example, temperature measurements at the influent and effluent of a primary clarifier are used to generate a reactor model. A virtual tracer experiment performed on the reactor model has good agreement with a tracer experiment performed on-site.

River water was treated by ultrafiltration at a relatively low transmembrane pressure (40 mbar). As observed before, flux stabilization occurred after several days of operation although no back-flushing or cross flow was applied. Interruptions in flux were applied by temporary offset of the transmembrane pressure. After restoration of the transmembrane pressure, the initial flux was higher than the stable flux level, and the flux recovery depended on the standstill time. Furthermore, if a short cross flow was applied after standstill, the flux was restored to an even higher level. In all cases, the flux decreased again during operation to reach finally the same stable level as before standstill. In order to evaluate the influence of intermittent operation as practiced for water treatment on a household level, daily interruptions of flux were applied. An optimum of total daily water production rate was obtained at 21 h of operation and 3 h of standstill per day. A model was developed which can describe the impact of intermittent operation on the flux depending on the duration of the standstill and operating periods. This enables the prediction of production capacity of the system operated intermittently. The flux increase during standstill could be explained by a relaxation and expansion of the biofouling layer, while the higher flux after forward-flushing was caused by this layer being partially sloughed off. Household water treatment with the process presented here will generally be operated on a discontinuous basis. The results show that such operation schemes do not compromise the permeability of the system, but actually lead to higher fluxes after standstill.

The hydraulics of water and wastewater treatment reactors has a major impact on their performance and control. The residence time distribution as a measure for the hydraulics represents macroscopic mixing in an integrated way with no spatial information. However, with regard to optimal sensor location for process control and for process optimisation measures, spatial information about macro-mixing is helpful. Spatially distributed measurements of reactive tracers can provide this information. In this paper we generally discuss how reactive tracers can be used to detect and characterize distinct large scale flow structures. It is shown that tracer substances are particularly suited if their reaction time scale is similar to the time scale of the large scale flow structure. For nitrifying activated sludge systems, ammonium is identified to be a suitable tracer. In a comprehensive experimental study at a real aeration tank, two distinct large scale flow features were identified by distributed ammonium measurements. Flow velocity measurements using acoustic Doppler velocimetry clearly supported the nature of these flow field anomalies. Ion-selective electrodes are a well suited device for ammonium measurements providing the temporal resolution that is needed for such an analysis.

A procedure is proposed which allows the detection of industrial discharge events at the inlet of a wastewater treatment plant without the need for measurements performed at the industry, for special equipment and for exact knowledge of the industrial sewage. By performing UV/Vis measurements at the inlet of a plant and analyzing them with a two-staged clustering method consisting of the self-organizing map algorithm and the Ward clustering method, typical sewage clusters can be found. In an experiment performed at a mid-sized Swiss plant, one cluster of a cluster model with five clusters could be attributed to an industrial laundry. Out of 95 laundry discharging events measured in a validation period, 93 were correctly detected by the proposed algorithm, two were false positives and five were false negatives.

Biocides leach from facades during rain events and subsequently enter the aquatic environment with storm water. Little is known about the losses of an entire settlement, since most studies referred to wash-off experiments conducted under laboratory conditions. Their results show a fast decrease of concentrations in the beginning, which subsequently slows down. The aim of this study is to develop a simple model to understand the mechanisms leading to these losses as well as to simulate losses under various rainfall and application conditions. We developed a four-box model based on the knowledge gained from fits of an exponential function to an existing experimental data set of a wash-off experiment. The model consists of two mobile stocks from which biocides are washed off during a rain event. These mobile stocks are supplied with biocides from storage stocks by diffusion-type processes. The model accurately reproduced the measured data of wash-off during single cycles as well as peak wash-offs over all cycles. Our model results for diuron losses showed that a large proportion (∼ 70%) of the applied biocides are still in the stocks even after a rain volume corresponding to several years (1100 mm y(-1), Swiss Plateau). Applications to realistic outdoor conditions showed that losses can not be neglected for urban environments and that knowledge about the amount of rainfall turned into runoff and the decay constants of the biocides in the facades are crucial. The model increased our understanding of the processes leading to the observed dynamic in laboratory experiments and was used to simulate losses for various rainfall and application conditions.

Based on three case studies it is demonstrated that the application of mathematical models in biological wastewater treatment has not yet reached its full potential. Model structure uncertainty and correlation of uncertain parameter values make up the first case. The combination of biokinetic and hydraulic models relates to the second case. The evolution of a full scale plant over its life expectancy is the frame for the third case. This paper was initially presented as a discussion starter and thus is raising questions rather than providing answers.

Due to the high energy input of aeration, the spatial distribution of air diffusers largely determines the flow field in aeration tanks. This has consequences on the efficiency of the aeration system, the performance of the aeration tank and on tank operation and control. This paper deals with these effects applying both Computational Fluid Dynamics (CFD) enhanced with a biokinetic model and full scale validation using velocity and reactive tracer measurements with high temporal and spatial resolution. It is shown that small changes in the diffuser arrangement drastically change the overall flow field. Using different aeration patterns in the same tank may lead to large scale instabilities in the flow field that lower plant performance and produce strong variations in concentration signals impeding their use for plant control. CFD is a valuable tool to analyze the interaction of flow field and aeration and their effects on plant performance and operation. But, in complex flow situations experimental validation is needed and strongly suggested.

This study demonstrates the usefulness of global sensitivity analysis in wastewater treatment plant (WWTP) design to prioritize sources of uncertainty and quantify their impact on performance criteria. The study, which is performed with the Benchmark Simulation Model no. 1 plant design, complements a previous paper on input uncertainty characterisation and propagation (Sin et al., 2009). A sampling-based sensitivity analysis is conducted to compute standardized regression coefficients. It was found that this method is able to decompose satisfactorily the variance of plant performance criteria (with R(2) > 0.9) for effluent concentrations, sludge production and energy demand. This high extent of linearity means that the plant performance criteria can be described as linear functions of the model inputs under the defined plant conditions. In effect, the system of coupled ordinary differential equations can be replaced by multivariate linear models, which can be used as surrogate models. The importance ranking based on the sensitivity measures demonstrates that the most influential factors involve ash content and influent inert particulate COD among others, largely responsible for the uncertainty in predicting sludge production and effluent ammonium concentration. While these results were in agreement with process knowledge, the added value is that the global sensitivity methods can quantify the contribution of the variance of significant parameters, e.g., ash content explains 70% of the variance in sludge production. Further the importance of formulating appropriate sensitivity analysis scenarios that match the purpose of the model application needs to be highlighted. Overall, the global sensitivity analysis proved a powerful tool for explaining and quantifying uncertainties as well as providing insight into devising useful ways for reducing uncertainties in the plant performance. This information can help engineers design robust WWTP plants.

Several strategic planning approaches have been proposed for dealing with future uncertainties in the urban water infrastructure sector. We identify three well established perspectives that address uncertainties in strategic decisions: an adaptive perspective, focusing on an incremental adaptation of existing structures as a reaction to unforeseen developments, a modeling perspective, focusing on an improved characterization of future context conditions and a managerial perspective, focusing on increasing the flexibility of the infrastructure organization. Despite their virtues, these approaches have definite weaknesses in their approach to uncertainty: they often consider a restricted scope of alternatives, they face substantial difficulties in predicting context conditions over time periods of decades and often consider objectives and tradeoffs only implicitly. We elaborate and illustrate with a case study a fourth perspective that may compensate for these specific weaknesses and complement the established strategic planning approaches. This perspective is based on a discursive, qualitative assessment of key elements in the strategic planning process among a selected set of local stakeholders and decision makers. We maintain that this approach leads to a more explicit and reflexive treatment of future uncertainty, conflicting objectives and a broadening of the considered alternatives and therefore to a more robust decision-making process.

Struvite crystallisation is a fast and reliable phosphor-us removal and recovery process for concentrated waste streams - such as hydrolysed human urine. In order to optimise P-elimination efficiency, it is beneficial to obtain larger particle sizes: they are easier to separate and less prone to wash-out than smaller particles. This paper presents the results of a study on the effect of process parameters on particle size in a single step struvite precipitation. Crystals formed in batch experiments with real hydrolysed urine were shown to have an average size of >90 mu m at pH 9 and 20 degrees C. This is reduced to 45 (mu)m when changing stirrer type. Particle size increases with lower supersaturation. The results showed that under otherwise constant conditions, particle size decreases with lower temperature and has a minimum between pH 9 and 10. Deviating trends are observed at pH <8. Struvite formation in a CSTR (continuously stirred tank reactor) process was shown to be a reliable stable process that does not require any pH control. A method based on conductivity measurement is presented to estimate ionic strength, which is needed for equilibrium calculations. (C) 2009 Elsevier Ltd. All rights reserved.

Reactor hydraulics is one of the key factors for plant performance and plant control. The residence time distribution is a good but limited indicator of reactor hydraulics. A more detailed view is obtained by direct observations within the reactor. Two different techniques (conservative tracer, reactive tracer) are discussed to detect major anomalies in the flow field of a wastewater aeration tank. Experiments with conservative tracers give valuable information over a very limited period of time making the analysis of the flow field difficult. On the other hand, reactive tracers can be monitored long term which helps identifying the flow pattern provided that a high spatial and temporal resolution of the measurements is applied. Experimental data is used to determine the structure and the parameters of a compartment model that corresponds well with the measurements.

This study quantifies the uncertainty involved in predicting micropollutant oxidation during drinking water ozonation in a pilot plant reactor. The analysis is conducted for geosmin, methyl tert-butyl ether (MTBE), isopropylmethoxypyrazine (IPMP), bezafibrate, beta-cyclocitral and ciprofloxazin. These compounds are representative for a wide range of substances with second order rate constants between 0.1 and 1.9x10(4)M(-1)s(-1) for the reaction with ozone and between 2x10(9) and 8x10(9)M(-1)s(-1) for the reaction with OH-radicals. Uncertainty ranges are derived for second order rate constants, hydraulic parameters, flow- and ozone concentration data, and water characteristic parameters. The uncertain model factors are propagated via Monte Carlo simulation and the resulting probability distributions of the relative residual micropollutant concentrations are assessed. The importance of factors in determining model output variance is quantified using Extended Fourier Amplitude Sensitivity Testing (Extended-FAST). For substances that react slowly with ozone (MTBE, IPMP, geosmin) the water characteristic R(ct)-value (ratio of ozone- to OH-radical concentration) is the most influential factor explaining 80% of the output variance. In the case of bezafibrate the R(ct)-value and the second order rate constant for the reaction with ozone each contribute about 30% to the output variance. For beta-cyclocitral and ciprofloxazin (fast reacting with ozone) the second order rate constant for the reaction with ozone and the hydraulic model structure become the dominating sources of uncertainty.

Reactor modeling is of major interest in environmental technology. In this context, new contaminants with higher degradation requirements increase the importance of reactor hydraulics. CFD (Computational Fluid Dynamics) may meet this challenge but is expensive for everyday use. In this paper, we provide research and practice with a methodology designed to automatically reduce the complexity of such a high-dimensional flow model to a compartmental model. The derivation is based on the concentration field of a reacting species which is included in the steady state CFD simulation. While still capturing the most important flow features, the compartmental model is fast, easy to use, and open for process modeling with yet unknown compounds. The inherent overestimation of diffusion by compartmental models has been corrected by locally adjusting turbulent fluxes. We successfully applied the methodology to the ozonation process and experimentally verified it with tracer experiments. The loss of information was quantified as a deviation from CFD performance prediction for different reactions. With increasing discretisation of the compartmental model, these deviations diminish. General advice on the necessary discretisation is given.

Based on the subjective experience of the author it is discussed how the nitrification processes served as an important basis for the development of today's understanding and mathematical models for many wastewater treatment processes (activated sludge, biofilm reactors) and self-purification processes in rivers. Besides being an important process for the protection of receiving waters, nitrification served as a proxy for the understanding of the behavior of a narrowly defined group of microorganisms growing on known substrates under environmental conditions. Until the upcoming of readily available microbial genetic techniques, nitrification was the single most studied microbial process in environmental engineering.

This study focuses on uncertainty analysis of WWTP models and analyzes the issue of framing and how it affects the interpretation of uncertainty analysis results. As a case study, the prediction of uncertainty involved in model-based design of a wastewater treatment plant is studied. The Monte Carlo procedure is used for uncertainty estimation, for which the input uncertainty is quantified through expert elicitation and the sampling is performed using the Latin hypercube method. Three scenarios from engineering practice are selected to examine the issue of framing: (1) uncertainty due to stoichiometric, biokinetic and influent parameters; (2) uncertainty due to hydraulic behaviour of the plant and mass transfer parameters; (3) uncertainty due to the combination of (1) and (2). The results demonstrate that depending on the way the uncertainty analysis is framed, the estimated uncertainty of design performance criteria differs significantly. The implication for the practical applications of uncertainty analysis in the wastewater industry is profound: (i) as the uncertainty analysis results are specific to the framing used, the results must be interpreted within the context of that framing; and (ii) the framing must be crafted according to the particular purpose of uncertainty analysis/model application. Finally, it needs to be emphasised that uncertainty analysis is no doubt a powerful tool for model-based design among others, however clear guidelines for good uncertainty analysis in wastewater engineering practice are needed.