Online nonlinear sequential Bayesian estimation of a biological wastewater treatment process

Online estimation of unknown state variables is a key component in the accurate modelling of biological wastewater treatment processes due to a lack of reliable online measurement systems. The extended Kalman filter (EKF) algorithm has been widely applied for wastewater treatment processes. However, the series approximations in the EKF algorithm are not valid, because biological wastewater treatment processes are highly nonlinear with a time-varying characteristic. This work proposes an alternative online estimation approach using the sequential Monte Carlo (SMC) methods for recursive online state estimation of a biological sequencing batch reactor for wastewater treatment. SMC is an algorithm that makes it possible to recursively construct the posterior probability density of the state variables, with respect to all available measurements, through a random exploration of the states by entities called ‘particle’. In this work, the simplified and modified Activated Sludge Model No. 3 with nonlinear biological kinetic models is used as a process model and formulated in a dynamic state-space model applied to the SMC method. The performance of the SMC method for online state estimation applied to a biological sequencing batch reactor with online and offline measured data is encouraging. The results indicate that the SMC method could emerge as a powerful tool for solving online state and parameter estimation problems without any model linearization or restrictive assumptions pertaining to the type of nonlinear models for biological wastewater treatment processes.     

Innovative graphene microbial platforms for domestic wastewater treatment

Supplying clean water to fulfill human requirements is one of this century’s priorities. Global water resources are barely aligned with the rising demand, which is further aggravated by rising population, climate change and water quality problems. Consequently, there is a persistent need for innovative technologies to valorize unconventional water resources such as domestic wastewater. Graphene holds promising prospects in developing domestic wastewater treatment to qualitatively enhance treatment efficiency and quantitatively increase water supply. This review highlights the existing wastewater treatment processes along with their challenges according to South Australian wastewater treatment plants (WWTPs) which are representative of many modern WWTPs. The discussion will also cover the current and potential applications of graphene for domestic wastewater treatment, as well as obstacles and research priorities required for commercialization.     

Confidence scores for prediction models

In medical statistics, many alternative strategies are available for building a prediction model based on training data. Prediction models are routinely compared by means of their prediction performance in independent validation data. If only one data set is available for training and validation, then rival strategies can still be compared based on repeated bootstraps of the same data. Often, however, the overall performance of rival strategies is similar and it is thus difficult to decide for one model. Here, we investigate the variability of the prediction models that results when the same modelling strategy is applied to different training sets. For each modelling strategy we estimate a confidence score based on the same repeated bootstraps. A new decomposition of the expected Brier score is obtained, as well as the estimates of population average confidence scores. The latter can be used to distinguish rival prediction models with similar prediction performances. Furthermore, on the subject level a confidence score may provide useful supplementary information for new patients who want to base a medical decision on predicted risk. The ideas are illustrated and discussed using data from cancer studies, also with high-dimensional predictor space.     

Considering Variations in Waste Composition during Waste Input‐Output Modeling

Concentrations of pollutants vary in wastes from different sources. However, existing waste input‐output (WIO) models do not take these differing concentrations into account. This article proposes a new category of model, which we are calling a waste input‐output model at the substance level (WIOS model). The WIOS model considers variations in waste composition. These variations potentially affect the life cycle inventory of the waste treatment stage. The proposed model is expected to produce more accurate results than existing WIO models that do not consider variations in the composition of wastes. In addition, the proposed model provides a method to trace substances undergoing waste treatment. In this article, use of the WIOS model is illustrated by simulating the overall environmental loads of total organic carbon from wastewater treatment at a facility in Germany. The results show that variations in the composition of wastes entering treatment significantly affect the modeled estimates of total environmental loads caused by wastewater treatment. In addition, the results of the proposed model are different from results given by existing hybrid input‐output WIO models that do not consider variations in the composition of wastewater as it undergoes treatment.     

Parallel hybrid modeling methods for a full-scale cokes wastewater treatment plant

Parallel hybrid modeling methods are applied to a full-scale cokes wastewater treatment plant. Within the hybrid model structure, a mechanistic model specifies the basic dynamics of the relevant process and a non-parametric model compensates for the inaccuracy of the mechanistic model. First, a simplified mechanistic model is developed based on Activated Sludge Model No. 1 and the specific process knowledge of the cokes wastewater treatment process. Then, the mechanistic model is combined with five different non-parametric models--feedforward back-propagation neural network, radial basis function network, linear partial least squares (PLS), quadratic PLS and neural network PLS (NNPLS)--in parallel configuration. These models are identified with the same data obtained from the plant operation to predict dynamic behavior of the process. The performance of each parallel hybrid model is compared based on their ease of model building, prediction accuracy and interpretability. For this application, the parallel hybrid model with NNPLS as non-parametric model gives better performance than other parallel hybrid models. In addition, the NNPLS model is used to analyze the behavior of the operation data in the reduced space and allows for fault detection and isolation.     

Development of mechanistically based model for simulating soluble microbial products generation in an aerated/non-aerated SBR

Soluble microbial products (SMPs) are considered as the main organic components in wastewater treatment plant effluent from biological wastewater treatment systems. To investigate and explore SMP metabolism pathway for further treatment and control, two innovative mechanistically based activated sludge models were developed by extension of activated sludge model no.3 (ASM3). One was the model by combining SMP formation and degradation (ASM3-SMP model) processes with ASM3, and the other by combining both SMP and simultaneous substrate storage and growth (SSSG) mechanisms with ASM3 (SSSG-ASM3-SMP model). The detailed schematic modification and process supplements were introduced for comprehensively understanding all the mechanisms involved in the activated sludge process. The evaluations of these two models were demonstrated by a laboratory-scale sequencing batch reactor (SBR) operated under aerated/non-aerated conditions. The simulated and measured results indicated that SMP comprised about 83% of total soluble chemical oxygen demand (SCOD) in which biomass-associated products (BAPs) were predominant compared with utilization-associated products (UAPs). It also elucidated that there should be a minimum SMP value as the reactive time increases continuously and this conclusion could be used to optimize effluent SCOD in activated sludge processes. The comparative results among ASM3, ASM3-SMP and SSSG-ASM3-SMP models and the experimental measurements (SCOD, ammonia and nitrate nitrogen) showed clearly the best agreement with SSSG-ASM3-SMP simulation values (R = 0.993), strongly suggesting that both SMP formation and degradation and SSSG mechanisms are necessary in biologically activated sludge modeling for municipal wastewater treatment.     

Life Cycle Assessment of Water: From the pumping station to the wastewater treatment plant (9 pp)

Goal, Scope and Background The goal of this study is to determine the environmental impact of using one cubic metre of water in the Walloon Region. The whole anthropogenic water cycle is analysed, from the pumping stations to the wastewater treatment plants. The functional unit has been defined as one cubic metre of water at the consumer tap. This study was carried out in the context of the EU Water Framework Directive. It is part of a programme called PIRENE launched by the Walloon Region to fulfil the requirements of this Directive. Methods A model of the whole anthropogenic water cycle in the Walloon Region was developed. The model is mainly based on site-specific data given by the companies working in the field of water production and wastewater treatment. It was used to assess the environmental impact from the pumping station to the wastewater treatment plant using the Eco-Indicator 99 methodology. Eco-Indicator 99 has been adapted in order to better take into account environmental impact of acidification and eutrophication. Characterisation factors have been calculated for COD, nitrogen and phosphate emissions. From the reference model, different scenarios have been elaborated. Results and Discussion On the basis of the inventory, the environmental impact of five scenarios has been evaluated. Acidification and eutrophication is the most important impact category. It is mainly caused by the wastewater that is discharged without any treatment, but also by the effluent of the wastewater treatment plant. So, this impact category has the lowest environmental load when the wastewater treatment rate is high. For the other impact categories, the impact generally increases with the wastewater treatment rate. During wastewater treatment, energy and chemicals are indeed consumed to improve the quality of the final outputs, and thus to reduce the environmental impact related to acidification and eutrophication. A comparison between the scenarios has also shown that the building of the sewer network has a significant contribution to the global environmental load and that the stages before the tap contribute less to the environmental impact than the stage after the tap. Conclusions The three stages that contribute significantly to the global environmental load are: water discharge, wastewater treatment operation and, to a lesser extent, the sewer system. The results show that the wastewater treatment rate must be as high as possible, using either collective or individual wastewater treatment plants. Even a small water discharge without any treatment has a significant environmental impact. Operation of the wastewater treatment plants must also be improved to reduce the environmental impact caused by the effluent of the plants. For new wastewater treatment plants, building plants treating nitrogen and phosphorus should be encouraged. A sensitivity analysis was conducted and showed that the results of the study were not very affected by a modification of key parameters. Impact assessment using the CML methodology has confirmed the results obtained with Eco-Indicator 99.     

Biological nitrogen and phosphorus removal in membrane bioreactors: model development and parameter estimation

Membrane bioreactors (MBR) are being increasingly used for wastewater treatment. Mathematical modeling of MBR systems plays a key role in order to better explain their characteristics. Several MBR models have been presented in the literature focusing on different aspects: biological models, models which include soluble microbial products (SMP), physical models able to describe the membrane fouling and integrated models which couple the SMP models with the physical models. However, only a few integrated models have been developed which take into account the relationships between membrane fouling and biological processes. With respect to biological phosphorus removal in MBR systems, due to the complexity of the process, practical use of the models is still limited. There is a vast knowledge (and consequently vast amount of data) on nutrient removal for conventional-activated sludge systems but only limited information on phosphorus removal for MBRs. Calibration of these complex integrated models still remains the main bottleneck to their employment. The paper presents an integrated mathematical model able to simultaneously describe biological phosphorus removal, SMP formation/degradation and physical processes which also include the removal of organic matter. The model has been calibrated with data collected in a UCT-MBR pilot plant, located at the Palermo wastewater treatment plant, applying a modified version of a recently developed calibration protocol. The calibrated model provides acceptable correspondence with experimental data and can be considered a useful tool for MBR design and operation.     

Some techniques in building and validating stochastic models of human reproduction

Stochastic models of human reproduction are beginning to play significant roles in the evaluation of family planning programs. A class of stochastic processes called absorbing, agedependent, semi-Markov processes frequently arises in the construction of such models. The paper begins with a discussion of some technicalities regarding absorbing, age-dependent, semi-Markov processes. Then, an algorithm due to Littman, which makes possible the computerization of this class of stochastic processes, is presented. Briefly, Littman’s algorithm provides an efficient method for numerically solving systems of renewal type integral equations, provided the system does not contain a large number of equations. After setting down a concrete model for a large clinical trial of intrauterine devices conducted in Taiwan, the paper concludes with a discussion of a method for validating the model based on the data collected in the clinical trial. Presented at the Society for Mathematical Biology Meeting, University of Pennsylvania, Philadelphia, August 19–21, 1976.     

Hybrid neural network modeling of a full-scale industrial wastewater treatment process

In recent years, hybrid neural network approaches, which combine mechanistic and neural network models, have received considerable attention. These approaches are potentially very efficient for obtaining more accurate predictions of process dynamics by combining mechanistic and neural network models in such a way that the neural network model properly accounts for unknown and nonlinear parts of the mechanistic model. In this work, a full-scale coke-plant wastewater treatment process was chosen as a model system. Initially, a process data analysis was performed on the actual operational data by using principal component analysis. Next, a simplified mechanistic model and a neural network model were developed based on the specific process knowledge and the operational data of the coke-plant wastewater treatment process, respectively. Finally, the neural network was incorporated into the mechanistic model in both parallel and serial configurations. Simulation results showed that the parallel hybrid modeling approach achieved much more accurate predictions with good extrapolation properties as compared with the other modeling approaches even in the case of process upset caused by, for example, shock loading of toxic compounds. These results indicate that the parallel hybrid neural modeling approach is a useful tool for accurate and cost-effective modeling of biochemical processes, in the absence of other reasonably accurate process models.     

Applying Life Cycle Tools and Process Engineering to Determine the Most Adequate Treatment Process Conditions. A Tool in Environmental Policy (12 pp)

Background The analysis of a wastewater treatment technology, under a expanded boundaries system which includes both the technology and the inputs required for its operation, quantifies the overall environmental impact that may result from the treatment of a wastewater stream. This is particularly useful for environmental policy makers being that a expanded boundaries system tends to provide a holistic view. The former view can be highly enriched with the use of process engineering tools, such as mathematical process modelling, process design, performance assessment and cost optimised models. Main Features The traditional approach used to assess waste treatment technologies is contrasted with a life cycle analysis (LCA) approach. The optimal design of a granular activated carbon adsorption (GAC) process is used as a model system to demonstrate the advantages of LCA approaches over traditional approaches. Further sections of the paper describe a mathematical framework for the assessment of technologies, design considerations applied in the cost optimised carbon adsorption model, the use of LCA techniques to perform an inventory of all emissions associated to the process system and, some of its environmental impacts. Results Economic and environmental considerations regarding the optimum process design are introduced as a basis for decision towards the selection and operating conditions of wastewater treatment technologies. Moreover, the use of LCA has revealed that the environmental burden associated with the wastewater treatment may produce a higher environmental impact than one that can be caused by untreated discharges. Conclusion The paper highlights the string advantages that environmental policy makers may have by combining LCA and process engineering tools. Furthermore, this approach can be incorporated into other existing treatment processes or for process designers.     

Energy use and recovery strategies within wastewater treatment and sludge handling at pulp and paper mills

This paper presents an inclusive approach with focus on energy use and recovery in wastewater management, including wastewater treatment (WWT) and sludge handling. Process data from three Swedish mills and a mathematical model were used to evaluate seven sludge handling strategies. The results indicate that excess energy use in WWT processes counters the potential energy recovery in the sludge handling systems. Energy use in WWT processes is recommended to aim for sufficient effluent treatment, not for sludge reduction. Increased secondary sludge production is favourable from an energy point of view provided it is used as a substrate for heat, biogas or electricity production.     

What do we gain from simplicity versus complexity in species distribution models?

Species distribution models (SDMs) are widely used to explain and predict species ranges and environmental niches. They are most commonly constructed by inferring species' occurrence–environment relationships using statistical and machine‐learning methods. The variety of methods that can be used to construct SDMs (e.g. generalized linear/additive models, tree‐based models, maximum entropy, etc.), and the variety of ways that such models can be implemented, permits substantial flexibility in SDM complexity. Building models with an appropriate amount of complexity for the study objectives is critical for robust inference. We characterize complexity as the shape of the inferred occurrence–environment relationships and the number of parameters used to describe them, and search for insights into whether additional complexity is informative or superfluous. By building ‘under fit’ models, having insufficient flexibility to describe observed occurrence–environment relationships, we risk misunderstanding the factors shaping species distributions. By building ‘over fit’ models, with excessive flexibility, we risk inadvertently ascribing pattern to noise or building opaque models. However, model selection can be challenging, especially when comparing models constructed under different modeling approaches. Here we argue for a more pragmatic approach: researchers should constrain the complexity of their models based on study objective, attributes of the data, and an understanding of how these interact with the underlying biological processes. We discuss guidelines for balancing under fitting with over fitting and consequently how complexity affects decisions made during model building. Although some generalities are possible, our discussion reflects differences in opinions that favor simpler versus more complex models. We conclude that combining insights from both simple and complex SDM building approaches best advances our knowledge of current and future species ranges.     

Domestic Water Use in the United States: A Life-Cycle Approach

Water and energy are two primary natural resources used by building occupants. A life-cycle assessment (LCA) is performed for water-consuming plumbing fixtures and water-consuming appliances during their operational life for four different building types. Within the cycle studied, water is extracted from the natural environment, subjected to water treatment, pumped to buildings for use, collected for wastewater treatment, and discharged back to the natural environment. Specifically, the impacts of water use, electricity and natural gas generation, energy consumption (for water and wastewater treatment, and for water heating), and the manufacture of water and wastewater treatment chemicals are evaluated both quantitatively and qualitatively on a generalized national level in the United States of America.
It is concluded that water use and consumption within buildings have a much larger impact on resource consumption than the water and wastewater treatment stages of the life cycle. To study this more specifically, the resource consumption of four different building types-an apartment building, a college dormitory, a motel, and an office building-is considered. Of these four building types, the apartment has the highest energy consumption (for water and wastewater treatment, and for water heating) per volume of water used, whereas the office building has the lowest. Similarly, the calculated LCA score for the apartment building is typically greater than those of the other three building types.     

Identifying Amsterdam's nutrient hotspots: A new method to map human excreta at building and neighborhood scale

Recovering nutrients from human excreta and wastewater has been receiving increasing attention as a means to supplement or replace synthetic fertilizer production. Apart from technologies for nutrient recovery at centralized wastewater treatment plants, numerous decentralized, source‐separated sanitation systems, also known as new sanitation systems, have been developed to facilitate recovery. Decision‐making for the planning and implementation of new sanitation systems would benefit from a spatially explicit inventory of nutrient hotspots in urban areas. To provide visual representations of nutrient loads, we developed a methodology that combines spatial‐temporal modeling with geographic information system analysis, and used it for the city of Amsterdam. The methodology is new in the field of nutrient mapping, especially at the smallest geographical scale: building. Nitrogen, phosphorus, and potassium loads and hotspots are mapped at both building and neighborhood scale, drawing attention to the need for multiple scale analyses in decision‐making. This study concludes with a discussion on the potential to further develop the method proposed to include more detailed and verified data and to identify nutrient hotspots that are promising as nutrient recovery sites with new sanitation systems.     

Waste Treatment and Assessment of Long-Term Emissions (8pp)

Goal, Scope and Background The disposal phase of a product’s life cycle in LCA is often neglected or based on coarse indicators like ‘kilogram waste’. The goal of report No. 13 of the ecoinvent project (Doka 2003) is to create detailed Life Cycle Inventories of waste disposal processes. The purpose of this paper is to give an overview of the models behind the waste disposal inventories in ecoinvent, to present exemplary results and to discuss the assessment of long-term emissions. This paper does not present a particular LCA study. Inventories are compiled for many different materials and various disposal technologies. Considered disposal technologies are municipal incineration and different landfill types, including sanitary landfills, hazardous waste incineration, waste deposits in deep salt mines, surface spreading of sludges, municipal wastewater treatment, and building dismantling. The inventoried technologies are largely based on Swiss plants. Inventories can be used for assessment of the disposal of common, generic waste materials like paper, plastics, packaging etc. Inventories are also used within the ecoinvent database itself to inventory the disposal of specific wastes generated during the production phase. Inventories relate as far as possible to the specific chemical composition of the waste material (waste-specific burdens). Certain expenditures are not related to the waste composition and are inventoried with average values (process-specific burdens). Methods The disposal models are based on previous work, partly used in earlier versions of ecoinvent/ETH LCI data. Important improvements were the extension of the number of considered chemical elements to 41 throughout all disposal models and new landfill models based on field data. New inventories are compiled for waste deposits in deep salt mines and building material disposal. Along with the ecoinvent data and the reports, also Excel-based software tools were created, which allow ecoinvent members to calculate waste disposal inventories from arbitrary waste compositions. The modelling of long-term emissions from landfills is a crucial part in any waste disposal process. In ecoinvent long-term emissions are defined as emissions occurring 100 years after present. They are reported in separate emission categories. The landfill inventories include long-term emissions with a time horizon of 60’000 years after present. Results and Discussion As in earlier studies, the landfills prove to be generally relevant disposal processes, as also incineration and wastewater treatment processes produce landfilled wastes. Heavy metals tend to concentrate in landfills and are washed out to a varying degree over time. Long-term emissions usually represent an important burden from landfills. Comparisons between burdens from production of materials and the burdens from their disposal show that disposal has a certain relevance. Conclusion The disposal phase should by default be included in LCA studies. The use of a material not only necessitates its production, but also requires its disposal. The created inventories and user tools facilitate heeding the disposal phase with a similar level of detail as production processes. The risk of LCA-based decisions shifting burdens from the production or use phase to the disposal phase because of data gaps can therefore be diminished. Recommendation and Perspective Future improvements should include the modelling of metal ore refining waste (tailings) which is currently neglected in ecoinvent, but is likely to be relevant for metals production. The disposal technologies considered here are those of developed Western countries. Disposal in other parts of the World can differ distinctly, for logistic, climatic and economic reasons. The cross-examination of landfill models to LCIA soil fate models could be advantageous. Currently only chemical elements, like copper, zinc, nitrogen etc. are heeded by the disposal models. A possible extension could be the modelling of the behaviour of chemical compounds, like dioxins or other hydrocarbons.     

Cutting the Wastewater Peaks for Cyclic Batch Production Plants

This paper proposes systematic approaches for cutting the wastewater peaks generated in cyclic batch production plants in which wastewater generation rates and concentration are varied with time. To reduce the investment for the wastewater treatment unit or not to violate the discharge limits, the fluctuations of both flow rate and concentration of the wastewater generated need to be controlled or equalized. Buffer tanks (or basin) are commonly used to regulate the wastewater flow and concentration before the wastewater is sent to treatment units or discharge. The size of the buffer tank needs to be designed properly so that its size is minimized and still has enough capacity to handle various operating conditions. In this paper, conceptual approaches are presented to minimize the size of the buffer tank for either controlling wastewater flow rate or concentration. Mathematical models are proposed for controlling both flow rate and concentration.     

Removal of natural and xeno-estrogens during conventional wastewater treatment

The ecological impacts of natural estrogens and xenoestrogens in treated wastewater include altered sexual development and sex ratios among continuously exposed organisms. The primary sources of estrogenic activity in wastewater are natural estrogens such as estrone, 17β-estradiol and estriol and synthetic compounds like 17α-ethinylestradiol, alkylphenols and alklphenol ethoxylates. Precursors in raw wastewater can yield estrogenic intermediates during wastewater treatment. All these compounds can be destroyed by biochemical processes, albeit at significantly different rates or under different conditions. That is, estrogenic compounds can be, but are not always, destroyed by conventional wastewater treatment processes, suggesting that conventional processes can be optimized for removal of estrogenic activity from wastewater. Sorption to sludges derived from wastewater treatment affects the fates of hydrophobic xenoestrogens such as nonylphenol, in part because the biodegradability of sorbed contaminants is limited. It may also be possible to tailor sludge stabilization processes to remove trace contaminants, including estrogens. For example, there are significant differences in the efficiencies of aerobic and anaerobic digestion for destruction of alkylphenols and probably other estrogenic compounds with aromatic moieties. Because advanced wastewater treatment is not economically feasible for most communities, there is ample incentive to develop accurate relationships between operational parameters and removal of estrogenic compounds during secondary wastewater treatment.     

Critical review of activated sludge modeling: State of process knowledge,modeling concepts,and limitations

This work critically reviews modeling concepts for standard activated sludge wastewater treatment processes (e.g., hydrolysis, growth and decay of organisms, etc.) for some of the most commonly used models. Based on a short overview on the theoretical biochemistry knowledge this review should help model users to better understand (i) the model concepts used; (ii) the differences between models, and (iii) the limits of the models. The seven analyzed models are: (1) ASM1; (2) ASM2d; (3) ASM3; (4) ASM3 + BioP; (5) ASM2d + TUD; (6) Barker & Dold model; and (7) UCTPHO+. Nine standard processes are distinguished and discussed in the present work: hydrolysis; fermentation; ordinary heterotrophic organisms (OHO) growth; autotrophic nitrifying organisms (ANO) growth; OHO & ANO decay; poly‐hydroxyalkanoates (PHA) storage; polyphosphate (polyP) storage; phosphorus accumulating organisms PAO) growth; and PAO decay. For a structured comparison, a new schematic representation of these processes is proposed. Each process is represented as a reaction with consumed components on the left of the figure and produced components on the right. Standardized icons, based on shapes and color codes, enable the representation of the stoichiometric modeling concepts and kinetics. This representation allows highlighting the conceptual differences of the models, and the level of simplification between the concepts and the theoretical knowledge. The model selection depending on their theoretical limitations and the main research needs to increase the model quality are finally discussed. Biotechnol. Bioeng. 2013; 110: 24–46. © 2012 Wiley Periodicals, Inc.