Mechanistic Process Model Research Articles

A full-scale operational digital twin for a water resource recovery facility-A case study of Eindhoven Water Resource Recovery Facility.

Digital transformation for the water sector has gained momentum in recent years, and many water resource recovery facilities modelers have already started transitioning from developing traditional models to digital twin (DT) applications. DTs simulate the operation of treatment plants in near real time and provide a powerful tool to the operators and process engineers for real-time scenario analysis and calamity mitigation, online process optimization, predictive maintenance, model-based control, and so forth. So far, only a few mature examples of full-scale DT implementations can be found in the literature, which only address some of the key requirements of a DT. This paper presents the development of a full-scale operational DT for the Eindhoven water resource recovery facility in The Netherlands, which includes a fully automated data-pipeline combined with a detailed mechanistic full-plant process model and a user interface co-created with the plant's operators. The automated data preprocessing pipeline provides continuous access to validated data, an influent generator provides dynamic predictions of influent composition data and allows forecasting 48 h into the future, and an advanced compartmental model of the aeration and anoxic bioreactors ensures high predictive power. The DT runs near real-time simulations every 2h. Visualization and interaction with the DT is facilitated by the cloud-based TwinPlant technology, which was developed in close interaction with the plant's operators. A set of predefined handles are made available, allowing users to simulate hypothetical scenarios such as process and equipment failures and changes in controller settings. The combination of the advanced data pipeline and process model development used in the Eindhoven DT and the active involvement of the operators/process engineers/managers in the development process makes the twin a valuable asset for decision making with long-term reliability. PRACTITIONER POINTS: A full-scale digital twin (DT) has been developed for the Eindhoven WRRF. The Eindhoven DT includes an automated continuous data preprocessing and reconciliation pipeline. A full-plant mechanistic compartmental process model of the plant has been developed based on hydrodynamic studies. The interactive user interface of the Eindhoven DT allows operators to perform what-if scenarios on various operational settings and process inputs. Plant operators were actively involved in the DT development process to make a reliable and relevant tool with the expected added value.

Design Space and Control Strategy for the Manufacturing of Wet Media Milled Drug Nanocrystal Suspensions by Adopting Mechanistic Process Modeling.

Wet media milling is a fully industrialized technology for the manufacturing of drug nanocrystal suspensions. This work describes the development of an advanced control strategy and an associated design space for a manufacturing process at a commercial scale. Full-scale experiments and mechanistic process modeling have been used to establish a physically reasonable control strategy of factors relevant to the quality attributes of the nanocrystal suspension. The design space has been developed based on a mature mechanistic process model of the wet media milling procedure. It presents the process-product attribute relationship between a multidimensional range of measured process parameters and a range of the product-quality attribute mean particle sizes. The control strategy allows for simple, robust, and sound scientific process control as well as the operational flexibility of the suspension batch size. This is an industrial case study of control strategy and design-space definition with the crucial contribution of mechanistic process modeling for an intended commercial manufacturing process.

Continuous adaptation of a digital twin model for a pilot flotation plant

Model-based methods have a key role in achieving the technical, economical, and environmental performance improvements of the mineral processing systems. However, unmodeled process phenomena and disturbances leading to unreliable modeling results, may prevent the efficient online utilization of these methods at the operational level after the deployment of the model. This study demonstrates the feasibility of online adaptation of a dynamic, mechanistic process models in mineral beneficiation application at a pilot environment. At first, a digital twin of the grinding and flotation stages of a pilot-scale plant was developed. In the experimental campaign, a change from copper-zinc-pyrite ore to a mixture of pyrite-rich and non-sulfide gangue-rich material was carried out. Thus, during experiments, a notable change in the flotation performance was observed, which could not be replicated by a constant-parameter digital twin model. The proposed parameter adaptation framework, encompassing stochastic optimization in moving time windows, was found to be suitable for finding new optimal model parameters during the changing experimental conditions by using the elemental grades in different flotation stages. In addition, simulation studies are presented to highlight the challenges of digital twin parameter adaptation in mineral processing applications, where often only a sparse and disturbance influenced data from the key process variables are available. The adaptive digital twin allows applying the dynamic, mechanistic process models efficiently for predictive simulations in operational decisions leading to more sustainable and resource-efficient minerals processing.

Kinetic Modeling of a Poly(N-vinylcaprolactam-co-glycidyl methacrylate) Microgel Synthesis: A Hybrid In Silico and Experimental Approach

Synthesizing microgels with new functionalities is vital for applications in various disciplines; yet microgel synthesis kinetics are mostly unknown. We present a bottom-up approach to model the synthesis of N-vinylcaprolactam-based microgels functionalized with glycidyl methacrylate. The existing synthesis model requires parameter values for unknown reaction rates, which we estimate by a hybrid approach based on quantum chemical calculations and experimental data. Using quantum mechanics, we compute propagation rate constants and enthalpies of the underlying polymerization reactions. We estimate 7 out of 21 reaction parameter values using the reaction calorimetry measurements and a mechanistic process model. Our hybrid approach averages a coefficient of determination of 0.97 for the enthalpy transfer rate over time during microgel synthesis. Our approach illustrates that quantum chemistry methods and physical experiments can be integrated into models toward better understanding and designing of pVCL/GMA microgel synthesis processes.

Economic Model Predictive Control of a Recirculating Aquaculture System

The recirculating aquaculture systems (RAS) has been proposed to reduce the water consumption of commercial aquaculture. RAS removes organic and particulate waste from fish tank water while also aerating; as such the treated water can be recycled back into the tanks. In this work, we treat the RAS as a batch process such that economic model predictive control (EMPC) can be applied using a mechanistic process model. The EMPC, which considers fish production profits as well as material utility and electricity costs for RAS, is deployed for various water temperatures such that its effect on the fish growth and economics are quantified. Moreover, batch length is also determined through tracking of the process profit trajectory. The results show that the EMPC-operated RAS can substantially increase the fish sales price with time-varying control decisions. Moreover, the EMPC is shown to adjust its operating policy mid-batch to accommodate for temperature disturbances.

Improving models of coordination incentives for biodiversity conservation by fitting a multi-agent simulation model to a lab experiment

Coordination incentives (CI) like the agglomeration bonus that reward the spatial agglomeration (or other spatial patterns) of biodiversity conservation measures are gaining increasing attention. Experiments on CI, accompanied by statistical analyses, reveal insights into the behaviour of human subjects. However, the scope of statistical models is limited and one may, as in other sciences like physics or ecology, gain additional insights by fitting mechanistic process models to the experimental data. I present the first application of this type in the context of CI and fit a multi-agent simulation model to a seminal experiment on the agglomeration bonus. Comparing two basic approaches for the decision making of the model agents, reinforcement learning and using expectations about the future, reveals that the latter is much better able to replicate the observations of the experiment. Improved models of agent behaviour are indispensable in the model-based assessment of CI for the conservation of biodiversity.

Effect of hot metal charging on economic and environmental indices of electric arc furnace steelmaking in China

Hot metal (HM) charging electric arc furnaces (EAF) with good production indices and economic benefits have been widely used in China. In this study, the economic and environmental indices of EAF were analyzed based on a mechanistic process model, and the energy consumption, carbon emissions, and production costs with different hot metal ratios (HMRs) were obtained. There are strong correlations between HMR and the energy consumption, carbon emissions, and production costs of EAF. The Pareto optimal solution for the current state is 50% HMR-EAF, which produces carbon emissions of 1402.68 kg-CO2/t and a production cost of 3609.40 CNY/t. Considering the current situation in China, this process is more in line with the need to coordinate the economic and environmental indices of steel plants compared to direct reduced iron (DRI)-EAF, scrap-EAF, and blast furnace-basic oxygen furnace (BF-BOF) processes. Moreover, it has the potential to be more easily realized and promoted.

Towards Autonomous Process Control—Digital Twin for CHO Cell-Based Antibody Manufacturing Using a Dynamic Metabolic Model

The development of new biologics is becoming more challenging due to global competition and increased requirements for process understanding and assured quality in regulatory approval. As a result, there is a need for predictive, mechanistic process models. These reduce the resources and time required in process development, generating understanding, expanding the possible operating space, and providing the basis for a digital twin for automated process control. Monoclonal antibodies are an important representative of industrially produced biologics that can be used for a wide range of applications. In this work, the validation of a mechanistic process model with respect to sensitivity, as well as accuracy and precision, is presented. For the investigated process conditions, the concentration of glycine, phenylalanine, tyrosine, and glutamine have been identified as significant influencing factors for product formation via statistical evaluation. Cell growth is, under the investigated process conditions, significantly dependent on the concentration of glucose within the investigated design space. Other significant amino acids were identified. A Monte Carlo simulation was used to simulate the cultivation run with an optimized medium resulting from the sensitivity analysis. The precision of the model was shown to have a 95% confidence interval. The model shown here includes the implementation of cell death in addition to models described in the literature.

An integrated real-time optimization, control, and estimation scheme for post-combustion CO2 capture

This study presents a novel operational scheme for post-combustion CO2 capture (PCC) plants downstream from fuel-fired power plants. The approach is comprised of real-time optimization (RTO), nonlinear model predictive control (NMPC), and moving horizon estimation (MHE) layers. These layers are integrated to operate the system economically via a new economic function that accounts for the most significant economic aspects of PCC, including the carbon economy, energy, chemical, and utility costs. The proposed approach was employed on the case study of an MEA-based PCC absorber section, which uses a mechanistic process model to provide an accurate representation of the system. The NMPC layer is novel in its ability to enable flexible control of the plant by manipulating fresh material streams to impact CO2 capture and the MHE layer is the first to provide accurate system estimates to the controller with realistically accessible measurements. The proposed scheme was subjected to a cofiring scenario, whereby the switching between two fuels (i.e., biomass and coal) is reflected in the flue gas composition. In this scenario, a ∼19% steady-state cost improvement is observed with respect to the pre-disturbance cost. Moreover, the MHE was shown to cause an acceptable ∼0.5% of performance loss in the process economics through its effect on the NMPC. The scheme was also subjected to a ±20% diurnal variation in power plant load through steps in the flue gas flowrate and was found to provide consistent steady-state economic improvements (from ∼12% cost improvement to ∼17% loss abatement) for each of the disturbances observed. Furthermore, a price variation scenario highlighted the operational dependence of the system upon changes in economic incentives via the prices. When compared to a ‘no RTO’ case, the scheme was found to yield economic improvement ranging from ∼3% to ∼14% depending on the pricing. All scenarios in the case study displayed steady-state cost savings that exceeded the energy penalty imposed on the power plant by the PCC plant. This suggests the proposed scheme is an effective framework for the economic operation of a general class of PCC plants (i.e., with different solvents, process designs and control schemes, etc.) and can help enable the viability of PCC for the continued use of fuel-firing.

System model development and computer experiments for continuous API manufacturing

This work details development of a system/process model and usage of Computer Experiments methodology for continuous manufacturing of Active Pharmaceutical Ingredients (API). The process of interest has three interconnected continuous unit operations, with the first two being reactors and the last being an evaporator. The overall workflow detailed can be summarized as follows: a) develop mechanistic process models, b) verify/validate the model, c) convert model parameters to process parameters d) design computational grid to simulate the model – Latin hypercube design in this case, e) perform simulations, f) fit meta-model (surrogate model) to the output parameter/attribute of interest, g) analyze the meta-model using global sensitivity analysis, and h) slice and dice the model, e.g., study response surfaces of interest, for further analysis. This workflow has several practical advantages in process development, optimization, and troubleshooting, especially when the number of process inputs and/or parameters are large and/or output response surfaces are potentially complex. Phenomenological model development is detailed for biphasic reaction with large solvent volume change.

A flexible air separation process: 1. Design and steady‐state optimizations

AbstractThe flexible operation of energy‐intensive processes, such as cryogenic air separation, has economic potential due to increasing fluctuations of the electricity markets. Multiproduct air separation processes with high ratios of liquid product are very promising for flexible operation due to storable products. We present a process design with an integrated liquefication cycle and liquid assist operation, that facilitates a high liquid product ratio and a flexible process operation. We use a mechanistic dynamic process model in steady‐state process optimizations covering the wide operational range of the proposed process. The optimization results show that the power demand can be varied in a range from 3.5 to 28 MW without violating operational constraints by changing the nitrogen and oxygen production rates. Thus, the proposed process is a promising air separation candidate for flexible operation with respect to fluctuating electricity markets.

Full Scale Implementation of AirPrex Phosphorus Recovery and the Development of a Mechanistic Process Model in GPS-X™

Full Scale Implementation of AirPrex Phosphorus Recovery and the Development of a Mechanistic Process Model in GPS-X™

Workflow to set up substantial target-oriented mechanistic process models in bioprocess engineering

Abstract A multitude of new applications in bioprocess technology strongly depend on model-based methods as they feature prediction and control capabilities. The critical path is usually the availability of suitable models. In this work a workflow for the generation of substantial target-oriented mechanistic process models is presented. This workflow is based on backpropagation starting from a material balance for a certain target variable. Iteratively, necessary states as well as mechanistic links are included in the model using a model library reducing the computational effort. The parameters of these links are estimated using a simplex algorithm whose objective function depends on the target variable only. Practical identifiability analysis is used for the assessment of the need of further iterations and for validating the mechanistic model. To demonstrate the workflow, a model describing a mammalian cell culture process aiming at modeling viable cell count is used as an example. The generated model satisfies the predefined requirements and is very simple, consisting of three states and seven model parameters only. The presented workflow is simple, generic, transparent, so that also applications in a regulatory environment should be possible. It also provides additional process knowledge that can be used in bioprocess development and optimization.

Studies on the enzymatic synthesis of N-acetylneuraminic acid with continuously operated enzyme membrane reactors on a milliliter scale

Multi-step enzymatic syntheses often show very complex regulation mechanisms. Hence, extensive studies are necessary to identify the optimal mode of operation and suitable reaction conditions. Besides the application of mechanistic process models, the miniaturization and parallelization of laboratory equipment can accelerate bioprocess optimization. For the first time, parallel enzyme membrane reactors (EMRs) on a milliliter scale were established, which allow the long-term operation of enzymatic processes in continuously operated EMRs and cascades thereof with an excellent parallel reproducibility. Exemplarily, the miniaturized EMRs were used for the validation of an established model for the two-step synthesis of the high-value pharmaceutical starting material N-acetylneuraminic acid (Neu5Ac) with an N-acyl-d-glucosamine 2-epimerase (AGE) and an N-acetylneuraminate lyase (NAL). Validation experiments for different reactor configurations, i.e. continuous EMRs and cascades of EMRs, revealed the high prediction accuracy of the model. The subsequent model analysis enabled an in silico optimization of the process with regard to reaction conditions and mode of operation. Thus, a combined approach of using mechanistic models and parallel EMRs on a milliliter scale is applicable for the first time and will contribute to the acceleration of the implementation and optimization of enzymatic production processes.

Application of a mechanistic model as a tool for on-line monitoring of pilot scale filamentous fungal fermentation processes-The importance of evaporation effects.

A mechanistic model-based soft sensor is developed and validated for 550L filamentous fungus fermentations operated at Novozymes A/S. The soft sensor is comprised of a parameter estimation block based on a stoichiometric balance, coupled to a dynamic process model. The on-line parameter estimation block models the changing rates of formation of product, biomass, and water, and the rate of consumption of feed using standard, available on-line measurements. This parameter estimation block, is coupled to a mechanistic process model, which solves the current states of biomass, product, substrate, dissolved oxygen and mass, as well as other process parameters including kL a, viscosity and partial pressure of CO2 . State estimation at this scale requires a robust mass model including evaporation, which is a factor not often considered at smaller scales of operation. The model is developed using a historical data set of 11 batches from the fermentation pilot plant (550L) at Novozymes A/S. The model is then implemented on-line in 550L fermentation processes operated at Novozymes A/S in order to validate the state estimator model on 14 new batches utilizing a new strain. The product concentration in the validation batches was predicted with an average root mean sum of squared error (RMSSE) of 16.6%. In addition, calculation of the Janus coefficient for the validation batches shows a suitably calibrated model. The robustness of the model prediction is assessed with respect to the accuracy of the input data. Parameter estimation uncertainty is also carried out. The application of this on-line state estimator allows for on-line monitoring of pilot scale batches, including real-time estimates of multiple parameters which are not able to be monitored on-line. With successful application of a soft sensor at this scale, this allows for improved process monitoring, as well as opening up further possibilities for on-line control algorithms, utilizing these on-line model outputs. Biotechnol. Bioeng. 2017;114: 589-599. © 2016 Wiley Periodicals, Inc.

Resource Efficiency Studies using a New Operator Training Simulator for a Bioethanol Plant

Biofuel production plants are expected to significantly contribute to an environmentally sustainable future. However, a prerequisite for this goal is the resource conserving and energy efficient operation of these plants. Complex process dynamics are supposed to have an important influence on resource and energy efficiency of biofuel plants. This leads to strong demands on plant operators and automation systems. In this contribution a newly developed operator training simulator (OTS) is used to demonstrate the influence of process control and automation strategies on the sustainable operation of a bioethanol plant. The OTS presented here is based on mechanistic and dynamic process models describing the unit operations and some additional equipment such as valves, pumps, etc.. The models were verified using pilot plant data and have been implemented in the control system and simulation software WinErs. Automation sequences for the full plant have been developed using GRAFCET. Simulation experiments with the operator training simulator of the bioethanol production plant clearly indicate the strong influence of operational strategies and conditions on the resource efficiency and bioethanol yield of the process. The operational strategies and conditions for the bioethanol fermentation unit, the cross flow unit as well as the rectification unit and the overall process may lead to a variation of the energy demands for ethanol production of 48 %.

A Realistic Process Example for MIMO MPC based on Autoregressive Models

Advanced controllers such as model predictive control are in use for a wide range of application in the process industry. The potential utilization of such advanced predictive controllers is far from exhausted. One barrier for more widespread implementation is the lack of simple methodologies for advanced control design development which may be used by non experts in control theory. This paper presents and illustrates the use of a simple methodology to design an offset-free MPC based on ARX models. Hence a mechanistic process model is not required. The forced circulation evaporator by Newell and Lee is used to illustrate the offset-free MPC based on ARX models for a nonlinear multivariate process.

Alstom's Regenerative Calcium Cycle - Norcem Derisking Study: Risk Mitigation in the Development of a 2nd Generation CCS Technology.

Alstom is a pioneer and industrial leader in the development of post combustion CCS technologies. Alstom's Regenerative Calcium Cycle (RCC) is a 2nd generation post combustion CCS technology utilizing a calcination/ carbonation loop to capture CO2 from flue gas at high temperatures. The CO2 capture cycle is driven by introducing heat required to regenerate sorbent at 900 °C (4.0 GJ/t CO2, chemically bound CO2) the regenerated sorbent is then used for CO2 capture from flue gas where the exothermic heat of the reverse reaction (-4.0 GJ/t CO2 captured) is recovered in a power cycle. Due to the option of using natural sorbent materials (limestone), RCC also provides attractive integration opportunities with industrial processes such as cement production. Recently indirectly fired RCC calcination concepts also open the door for game changing performance gains using natural or even synthetic materials. However, the wide range of performance variability characteristic of natural sorbents contributes to uncertainty in defining process performance. Reliable sorbent deactivation models, which describe both cyclic and chemical deactivation, in combination with mechanistic reactor models which predict gas and solids phase conversions are required to lower the uncertainty associated with new reactor concepts and larger geometries required for commercialization for retrofit power-plant flue gas or for integrated solutions for the cement industry. Parallel to RCC process development, Alstom is currently executing a “Derisking” or risk-mitigation study for Norcem, a Norwegian cement producer, who seeks to take a leading position in the development of technologies for CO2-capture from cement production facilities. Norcem is currently heading up a project on behalf of the European cement industry (partners are Norcem, HeidelbergCement and the European Cement Research Academy) investigating various technological options at the pilot scale. In relation to this project, Alstom's RCC technology was selected with the objective to define process performance and the associated uncertainty. Pilot testing was conducted at the Institute of Combustion and Power Plant Technology at the University of Stuttgart (IFK) planned by Alstom focusing on the validation of process models and less the targeting of pilot plant performance, which is often limited by reactor geometries. This paper discusses the characterization of sorbent performance and the gap between the performance expected from thermo- gravimetric analysis (TGA) and that realized during steady state pilot testing under representative operating conditions. Operational stability, material balance closure and the confirmation of sorbent attrition characteristics from pilot testing are also discussed and steady-state operational data, gained from pilot testing are applied to initialize mechanistic process models which allow prediction of reactor performance for larger more efficient designs.

Integrative modelling of animal movement: incorporatingin situhabitat and behavioural information for a migratory marine predator

A fundamental goal in animal ecology is to quantify how environmental (and other) factors influence individual movement, as this is key to understanding responsiveness of populations to future change. However, quantitative interpretation of individual-based telemetry data is hampered by the complexity of, and error within, these multi-dimensional data. Here, we present an integrative hierarchical Bayesian state-space modelling approach where, for the first time, the mechanistic process model for the movement state of animals directly incorporates both environmental and other behavioural information, and observation and process model parameters are estimated within a single model. When applied to a migratory marine predator, the southern elephant seal (Mirounga leonina), we find the switch from directed to resident movement state was associated with colder water temperatures, relatively short dive bottom time and rapid descent rates. The approach presented here can have widespread utility for quantifying movement-behaviour (diving or other)-environment relationships across species and systems.

Comparative Study of Intelligent Soft-Sensors for Bioprocess State Estimation

 Abstract—In this article, application of soft-sensors for indirect determination of biomass and product concentration in a complex fed-batch biotechnological process is discussed. Three advanced techniques for softsensor design were investigated: feed-forward artificial neural networks, support vector regression model, and relevance vector regression model. Glucose /lactose feed rates and oxygen uptake rate along with its integrated quantity were used as direct reference measurements for estimation. Estimation quality of analyzed soft-sensors was tested using data generated by mechanistic process model. All three analyzed estimation techniques provided very similar estimation results from statistical point of view; nevertheless employment of regression models has some advantage because of its simplicity. Based on that, recommendations for application of the elaborated softsensors are given.