Benchmark Simulation Model No. 1 Research Articles

An emergent approach for the control of wastewater treatment plants by means of reinforcement learning techniques

One of the main problems in the automation of the control of wastewater treatment plants (WWTPs) appears when the control system does not respond as it should because of changes on influent load or flow. To tackle this difficult task, the application of Artificial Intelligence is not new, and in fact, currently Expert Systems may supervise the plant 24h/day assisting the plant operators in their daily work. However, the knowledge of the Expert System must be elicited previously from interviews to plant operators and/or extracted from data previously stored in databases. Although this approach still has a place in the control of wastewater treatment plants, it should aim to develop autonomous systems that learn from the direct interaction with the WWTP and that can operate taking into account changing environmental circumstances. In this paper we present an approach based on an agent with learning capabilities. In this approach, the agent’s knowledge emerges from the interaction with the plant. In order to show the validity of our assertions, we have implemented such an emergent approach for the N-Ammonia removal process in a well established simulated WWTP known as Benchmark Simulation Model No.1 (BSM1).

Dynamic influent pollutant disturbance scenario generation using a phenomenological modelling approach

Activated Sludge Models are widely used for simulation-based evaluation of wastewater treatment plant (WWTP) performance. However, due to the high workload and cost of a measuring campaign on a full-scale WWTP, many simulation studies suffer from lack of sufficiently long influent flow rate and concentration time series representing realistic wastewater influent dynamics. In this paper, a simple phenomenological modelling approach is proposed as an alternative to generate dynamic influent pollutant disturbance scenarios. The presented set of models is constructed following the principles of parsimony (limiting the number of parameters as much as possible), transparency (using parameters with physical meaning where possible) and flexibility (easily extendable to other applications where long dynamic influent time series are needed). The proposed approach is sub-divided in four main model blocks: 1) model block for flow rate generation, 2) model block for pollutants generation (carbon, nitrogen and phosphorus), 3) model block for temperature generation and 4) model block for transport of water and pollutants. The paper is illustrated with the results obtained during the development of the dynamic influent of the Benchmark Simulation Model no. 2 (BSM2). The series of simulations show that it is possible to generate a dry weather influent describing diurnal flow rate dynamics (low rate at night, high rate during day time), weekend effects (with different flow rate during weekends, compared to weekdays), holiday effects (where the wastewater production is assumed to be different for a number of weeks) and seasonal effects (with variations in the infiltration and thus also the flow rate to the WWTP). In addition, the dry weather model can be extended with a rain and storm weather generator, where the proposed phenomenological model can also mimic the “first flush” effect from the sewer network and the influent dilution phenomena that are typically observed at full-scale WWTPs following a rain event. Finally, the extension of the sewer system can be incorporated in the influent dynamics as well: the larger the simulated sewer network, the smoother the simulated diurnal flow rate and concentration variations. In the discussion, it is pointed out how the proposed phenomenological models can be expanded to other applications, for example to represent heavy metal or organic micro-pollutant loads entering the treatment plant.

Systematic Approach to the Design of Operation and Control Policies in Activated Sludge Systems

This article describes the systematic design of a control structure for a biological wastewater treatment process as given by the test-bed Benchmark Simulation Model No. 1 (BSM1) and Activated Sludge Model No. 1 (ASM1). The objective of this work was to formalize and implement a systematic and yet simple procedure for the selection of control structures in wastewater treatment plants (WWTPs) and to show that the application of the proposed methodology agrees with the “empirical”’ findings regarding the operation of this process. The motivation underlying this endeavor was to search for a control configuration that leads to optimal economic operation while promptly rejecting disturbances at lower layers in the control hierarchy, thus avoiding violation of the more important regulatory constraints on effluent discharge. We started by optimizing a steady-state nonlinear model of the process for various important disturbances. The results confirmed that it is economically optimal to control the oxygen concent...

Benchmark Simulation Model No 2: finalisation of plant layout and default control strategy

The COST/IWA Benchmark Simulation Model No 1 (BSM1) has been available for almost a decade. Its primary purpose has been to create a platform for control strategy benchmarking of activated sludge processes. The fact that the research work related to the benchmark simulation models has resulted in more than 300 publications worldwide demonstrates the interest in and need of such tools within the research community. Recent efforts within the IWA Task Group on "Benchmarking of control strategies for WWTPs" have focused on an extension of the benchmark simulation model. This extension aims at facilitating control strategy development and performance evaluation at a plant-wide level and, consequently, includes both pretreatment of wastewater as well as the processes describing sludge treatment. The motivation for the extension is the increasing interest and need to operate and control wastewater treatment systems not only at an individual process level but also on a plant-wide basis. To facilitate the changes, the evaluation period has been extended to one year. A prolonged evaluation period allows for long-term control strategies to be assessed and enables the use of control handles that cannot be evaluated in a realistic fashion in the one week BSM1 evaluation period. In this paper, the finalised plant layout is summarised and, as was done for BSM1, a default control strategy is proposed. A demonstration of how BSM2 can be used to evaluate control strategies is also given.

Real-Time Control and Effluent Ammonia Violations Induced by Return Liquor Overloads

Whole plant modeling and control has received increasing attention in recent years and as a result the Benchmark Simulation Model No 2 (BSM2) platform was developed to compare control strategies. The objective of this paper consists of evaluating whether MLSS concentration control (with changing setpoints in summer and winter time to maintain nitrification capacity) could lead to effluent ammonia violations due to return liquor overloads and how such violations could be minimised by introducing alternative control systems. In this study three different control strategies have been implemented in the BSM2 platform, simulated and evaluated from a whole plant perspective. The results show that with the implementation of the MLSS controller a large overload is applied to the digester at the end of the winter period, but the large retention time of the anaerobic digester sufficiently dampens this overload. On the other hand, the flows can be treated more efficiently when storing the nitrogen-rich return flows during day time and releasing them at night using a simple timer-based controller.

Evaluation of plant-wide WWTP control strategies including the effects of filamentous bulking sludge

The main objective of this paper is to evaluate the effect of filamentous bulking sludge on the predicted performance of simulated plant-wide WWTP control strategies. First, as a reference case, several control strategies are implemented, simulated and evaluated using the IWA Benchmark Simulation Model No. 2 (BSM2). In a second series of simulations the parameters of the secondary settler model in the BSM2 are automatically changed on the basis of an on-line calculated risk of filamentous bulking, in order to mimic the effect of growth of filamentous bacteria in the plant. The results are presented using multivariate analysis. Including the effects of filamentous bulking in the simulation model gives a-more realistic-deterioration of the plant performance during periods when the conditions for development of filamentous bulking sludge are favourable: compared to the reference case where bulking effects are not considered. Thus, there is a decrease of the overall settling velocity, an accumulation of the total suspended solids (TSS) in the middle layers of the settler with a consequent reduction of their degree of compaction in the bottom. As a consequence there is a lower TSS concentration in both return and waste flow, less biomass in the bioreactors and a reduction of the TSS removal efficiency. The control alternatives using a TSS controller substantially increase the food to microorganisms (F/M) ratio in the bioreactor, thereby reducing both risk and effects of bulking sludge. The effects of ammonium (NH(4)(+)), nitrate (NO(3)(-)) and reject water control strategies are rather poor when it comes to handling solids separation problems.

Including the effects of filamentous bulking sludge during the simulation of wastewater treatment plants using a risk assessment model

The main objective of this paper is to demonstrate how including the occurrence of filamentous bulking sludge in a secondary clarifier model will affect the predicted process performance during the simulation of WWTPs. The IWA Benchmark Simulation Model No. 2 (BSM2) is hereby used as a simulation case study. Practically, the proposed approach includes a risk assessment model based on a knowledge-based decision tree to detect favourable conditions for the development of filamentous bulking sludge. Once such conditions are detected, the settling characteristics of the secondary clarifier model are automatically changed during the simulation by modifying the settling model parameters to mimic the effect of growth of filamentous bacteria. The simulation results demonstrate that including effects of filamentous bulking in the secondary clarifier model results in a more realistic plant performance. Particularly, during the periods when the conditions for the development of filamentous bulking sludge are favourable – leading to poor activated sludge compaction, low return and waste TSS concentrations and difficulties in maintaining the biomass in the aeration basins – a subsequent reduction in overall pollution removal efficiency is observed. Also, a scenario analysis is conducted to examine i) the influence of sludge retention time (SRT), the external recirculation flow rate ( Q r) and the air flow rate in the bioreactor (modelled as k La) as factors promoting bulking sludge, and ii) the effect on the model predictions when the settling properties are changed due to a possible proliferation of filamentous microorganisms. Finally, the potentially adverse effects of certain operational procedures are highlighted, since such effects are normally not considered by state-of-the-art models that do not include microbiology-related solids separation problems.

Multi-criteria analysis of wastewater treatment plant design and control scenarios under uncertainty

Wastewater treatment plant control and monitoring can help to achieve good effluent quality, in a complex, highly non-linear process. The Benchmark Simulation Model no. 2 (BSM2) is a useful tool to competitively evaluate plant-wide control on a long-term basis. A method to conduct scenario analysis of process designs by means of Monte Carlo (MC) simulations and multi-criteria evaluation is presented. It is applied to the open loop version of BSM2 and to two closed loop versions, one with a simple oxygen controller and the other one with an ammonium controller regulating the set-point of the oxygen controller (cascade controller). The results show a much greater benefit of the cascade controller compared to the simple controller, both in environmental and economic terms. From an optimal process design point of view, the results show that the volume of the primary clarifier and the anoxic fraction of the reactor volume have an important impact on process performance. The uncertainty analysis of the optimal designs, also performed with MC simulations, highlights the improved and more stable effluent under closed loop control.

Using bioprocess stoichiometry to build a plant-wide mass balance based steady-state WWTP model

Steady-state models are useful for design of wastewater treatment plants (WWTPs) because they allow reactor sizes and interconnecting flows to be simply determined from explicit equations in terms of unit operation performance criteria. Once the overall WWTP scheme is established and the main system defining parameters of the individual unit operations estimated, dynamic models can be applied to the connected unit operations to refine their design and evaluate their performance under dynamic flow and load conditions. To model anaerobic digestion (AD) within plant-wide WWTP models, not only COD and nitrogen (N) but also carbon (C) fluxes entering the AD need to be defined. Current plant-wide models, like benchmark simulation model No 2 (BSM2), impose a C flux at the AD influent. In this paper, the COD and N mass balance steady-state models of activated sludge (AS) organics degradation, nitrification and denitrification (ND) and anaerobic (AD) and aerobic (AerD) digestion of wastewater sludge are extended and linked with bioprocess transformation stoichiometry to form C, H, O, N, chemical oxygen demand (COD) and charge mass balance based models so that also C (and H and O) can be tracked through the whole WWTP. By assigning a stoichiometric composition ( x, y, z and a in C x H y O z N a ) to each of the five main influent wastewater organic fractions and ammonia, these, and the products generated from them via the biological processes, are tracked through the WWTP. The model is applied to two theoretical case study WWTPs treating the same raw wastewater (WW) to the same final sludge residual biodegradable COD. It is demonstrated that much useful information can be generated with the relatively simple steady-state models to aid WWTP layout design and track the different products exiting the WWTP via the solid, liquid and gas streams, such as aerobic versus anaerobic digestion of waste activated sludge, N loads in recycle streams, methane production for energy recovery and green house gas (CO 2, CH 4) generation. To reduce trial and error usage of WWTP simulation software, it is recommended that they are extended to include pre-processors based on mass balance steady-state models to assist with WWTP layout design, unit operation selection, reactor sizing, option evaluation and comparison and wastewater characterization before dynamic simulation.

An ASM/ADM model interface for dynamic plant-wide simulation

Mathematical modelling has proven to be very useful in process design, operation and optimisation. A recent trend in WWTP modelling is to include the different subunits in so-called plant-wide models rather than focusing on parts of the entire process. One example of a typical plant-wide model is the coupling of an upstream activated sludge plant (including primary settler, and secondary clarifier) to an anaerobic digester for sludge digestion. One of the key challenges when coupling these processes has been the definition of an interface between the well accepted activated sludge model (ASM1) and anaerobic digestion model (ADM1). Current characterisation and interface models have key limitations, the most critical of which is the over-use of X c (or lumped complex) variable as a main input to the ADM1. Over-use of X c does not allow for variation of degradability, carbon oxidation state or nitrogen content. In addition, achieving a target influent pH through the proper definition of the ionic system can be difficult. In this paper, we define an interface and characterisation model that maps degradable components directly to carbohydrates, proteins and lipids (and their soluble analogues), as well as organic acids, rather than using X c. While this interface has been designed for use with the Benchmark Simulation Model No. 2 (BSM2), it is widely applicable to ADM1 input characterisation in general. We have demonstrated the model both hypothetically (BSM2), and practically on a full-scale anaerobic digester treating sewage sludge.

Interaction between control and design of a SHARON reactor: economic considerations in a plant-wide (BSM2) context

The combined SHARON-Anammox process is a promising technique for nitrogen removal from wastewater streams with high ammonium concentrations. It is typically applied to sludge digestion reject water, in order to relieve the activated sludge tanks, to which this stream is typically recycled. This contribution assesses the impact of the applied control strategy in the SHARON-reactor, both on the effluent quality of the subsequent Anammox reactor as well as on the plant-wide level by means of an operating cost index. Moreover, it is investigated to which extent the usefulness of a certain control strategy depends on the reactor design (volume). A simulation study is carried out using the plant-wide Benchmark Simulation Model no. 2 (BSM2), extended with the SHARON and Anammox processes. The results reveal a discrepancy between optimizing the reject water treatment performance and minimizing plant-wide operating costs.

Benchmark simulation model no 2: general protocol and exploratory case studies

Over a decade ago, the concept of objectively evaluating the performance of control strategies by simulating them using a standard model implementation was introduced for activated sludge wastewater treatment plants. The resulting Benchmark Simulation Model No 1 (BSM1) has been the basis for a significant new development that is reported on here: Rather than only evaluating control strategies at the level of the activated sludge unit (bioreactors and secondary clarifier) the new BSM2 now allows the evaluation of control strategies at the level of the whole plant, including primary clarifier and sludge treatment with anaerobic sludge digestion. In this contribution, the decisions that have been made over the past three years regarding the models used within the BSM2 are presented and argued, with particular emphasis on the ADM1 description of the digester, the interfaces between activated sludge and digester models, the included temperature dependencies and the reject water storage. BSM2-implementations are now available in a wide range of simulation platforms and a ring test has verified their proper implementation, consistent with the BSM2 definition. This guarantees that users can focus on the control strategy evaluation rather than on modelling issues. Finally, for illustration, twelve simple operational strategies have been implemented in BSM2 and their performance evaluated. Results show that it is an interesting control engineering challenge to further improve the performance of the BSM2 plant (which is the whole idea behind benchmarking) and that integrated control (i.e. acting at different places in the whole plant) is certainly worthwhile to achieve overall improvement.

Application of multivariable statistical techniques in plant-wide WWTP control strategies analysis

The main objective of this paper is to present the application of selected multivariable statistical techniques in plant-wide wastewater treatment plant (WWTP) control strategies analysis. In this study, cluster analysis (CA), principal component analysis/factor analysis (PCA/FA) and discriminant analysis (DA) are applied to the evaluation matrix data set obtained by simulation of several control strategies applied to the plant-wide IWA Benchmark Simulation Model No 2 (BSM2). These techniques allow i) to determine natural groups or clusters of control strategies with a similar behaviour, ii) to find and interpret hidden, complex and casual relation features in the data set and iii) to identify important discriminant variables within the groups found by the cluster analysis. This study illustrates the usefulness of multivariable statistical techniques for both analysis and interpretation of the complex multicriteria data sets and allows an improved use of information for effective evaluation of control strategies.

Benchmark Simulation Model No 2 in Matlab-Simulink: towards plant-wide WWTP control strategy evaluation

In this paper, implementation of the Benchmark Simulation Model No 2 (BSM2) within Matlab-Simulink is presented. The BSM2 is developed for plant-wide WWTP control strategy evaluation on a long-term basis. It consists of a pre-treatment process, an activated sludge process and sludge treatment processes. Extended evaluation criteria are proposed for plant-wide control strategy assessment. Default open-loop and closed-loop strategies are also proposed to be used as references with which to compare other control strategies. Simulations indicate that the BM2 is an appropriate tool for plant-wide control strategy evaluation.

Plant-wide (BSM2) evaluation of reject water treatment with a SHARON-Anammox process

In wastewater treatment plants (WWTPs) equipped with sludge digestion and dewatering systems, the reject water originating from these facilities contributes significantly to the nitrogen load of the activated sludge tanks, to which it is typically recycled. In this paper, the impact of reject water streams on the performance of a WWTP is assessed in a simulation study, using the Benchmark Simulation Model no. 2 (BSM2), that includes the processes describing sludge treatment and in this way allows for plant-wide evaluation. Comparison of performance of a WWTP without reject water with a WWTP where reject water is recycled to the primary clarifier, i.e. the BSM2 plant, shows that the ammonium load of the influent to the primary clarifier is 28% higher in the case of reject water recycling. This results in violation of the effluent total nitrogen limit. In order to relieve the main wastewater treatment plant, reject water treatment with a combined SHARON-Anammox process seems a promising option. The simulation results indicate that significant improvements of the effluent quality of the main wastewater treatment plant can be realized. An economic evaluation of the different scenarios is performed using an Operating Cost Index (OCI).

Continuity-based model interfacing for plant-wide simulation: A general approach

In plant-wide simulation studies of wastewater treatment facilities, often existing models from different origin need to be coupled. However, as these submodels are likely to contain different state variables, their coupling is not straightforward. The continuity-based interfacing method (CBIM) provides a general framework to construct model interfaces for models of wastewater systems, taking into account conservation principles. In this contribution, the CBIM approach is applied to study the effect of sludge digestion reject water treatment with a SHARON–Anammox process on a plant-wide scale. Separate models were available for the SHARON process and for the Anammox process. The Benchmark simulation model no. 2 (BSM2) is used to simulate the behaviour of the complete WWTP including sludge digestion. The CBIM approach is followed to develop three different model interfaces. At the same time, the generally applicable CBIM approach was further refined and particular issues when coupling models in which pH is considered as a state variable, are pointed out.

WWTP dynamic disturbance modelling–an essential module for long-term benchmarking development

Intensive use of the benchmark simulation model No. 1 (BSM1), a protocol for objective comparison of the effectiveness of control strategies in biological nitrogen removal activated sludge plants, has also revealed a number of limitations. Preliminary definitions of the long-term benchmark simulation model No. 1 (BSM1_LT) and the benchmark simulation model No. 2 (BSM2) have been made to extend BSM1 for evaluation of process monitoring methods and plant-wide control strategies, respectively. Influent-related disturbances for BSM1_LT/BSM2 are to be generated with a model, and this paper provides a general overview of the modelling methods used. Typical influent dynamic phenomena generated with the BSM1_LT/BSM2 influent disturbance model, including diurnal, weekend, seasonal and holiday effects, as well as rainfall, are illustrated with simulation results. As a result of the work described in this paper, a proposed influent model/file has been released to the benchmark developers for evaluation purposes. Pending this evaluation, a final BSM1_LT/BSM2 influent disturbance model definition is foreseen. Preliminary simulations with dynamic influent data generated by the influent disturbance model indicate that default BSM1 activated sludge plant control strategies will need extensions for BSM1_LT/BSM2 to efficiently handle 1 year of influent dynamics.

Towards a benchmark simulation model for plant-wide control strategy performance evaluation of WWTPs

The COST/IWA benchmark simulation model has been available for seven years. Its primary purpose has been to create a platform for control strategy benchmarking of activated sludge processes. The fact that the benchmark has resulted in more than 100 publications, not only in Europe but also worldwide, demonstrates the interest in such a tool within the research community In this paper, an extension of the benchmark simulation model no 1 (BSM1) is proposed. This extension aims at facilitating control strategy development and performance evaluation at a plant-wide level and, consequently, includes both pre-treatment of wastewater as well as the processes describing sludge treatment. The motivation for the extension is the increasing interest and need to operate and control wastewater treatment systems not only at an individual process level but also on a plant-wide basis. To facilitate the changes, the evaluation period has been extended to one year. A prolonged evaluation period allows for long-term control strategies to be assessed and enables the use of control handles that cannot be evaluated in a realistic fashion in the one-week BSM1 evaluation period. In the paper, the extended plant layout is proposed and the new suggested process models are described briefly. Models for influent file design, the benchmarking procedure and the evaluation criteria are also discussed. And finally, some important remaining topics, for which consensus is required, are identified.

Towards a common benchmark for long-term process control and monitoring performance evaluation

The COST/IWA benchmark simulation model has been available for seven years. Its primary purpose has been to create a platform for control strategy benchmarking of biological wastewater treatment processes. The fact that the benchmark has resulted in more than 100 publications, not only in Europe but also worldwide, demonstrates the interest for such a tool in the research community. In this paper, an extension of the benchmark simulation model no. 1 (BSM1) is proposed. It aims at facilitating evaluation of two closely related operational tasks: long-term control strategy performance and process monitoring performance. The motivation for the extension is that these two tasks typically act on longer time scales. The extension proposed here consists of 1) prolonging the evaluation period to one year (including influent files), 2) specifying time varying process parameters and 3) including sensor and actuator failures. The prolonged evaluation period is necessary to obtain a relevant and realistic assessment of the effects of such disturbances. Also, a prolonged evaluation period allows for a number of long-term control actions/handles that cannot be evaluated in a realistic fashion in the one week BSM1 evaluation period. In the paper, models for influent file design, parameter changes and sensor failures, initialization procedure and evaluation criteria are discussed. Important remaining topics, for which consensus is required, are identified. The potential of a long-term benchmark is illustrated with an example of process monitoring algorithm benchmarking.