Search

Feedforward neural networks with a single hidden layer of neurons and a linear output layer are a convenient way to model a nonlinear input-output mapping. If the activation weights, i.e. the weights between input and hidden-layer neurons, are known, an estimation problem remains that is linear in the parameters. This can easily be solved by standard least-squares methods. The problem thus reduces to finding appropriate activation weights. This paper describes a method to obtain the activation weights, based on local linear approximations, which also can be solved with standard least-squares techniques. The local linear models can be obtained by fuzzy clustering methods. The method is demonstrated on a simple example. With the proposed method the weights are obtained very fast, and the results are good. The method is also flexible with respect to the incorporation of a priori process knowledge.

In this paper a grey-box modelling approach for the identification of nonlinear, time-varying dissolved oxygen dynamics in an activated sludge wastewater treatment process (ASP) is presented. Herein, singular value decomposition of the locally available Jacobian matrix, or equivalently eigenvalue decomposition of the parameter covariance matrix, as well as parameter transformation are essential tools to arrive at a model suitable for adaptive, long-range prediction. The use of on-line respiration rate measurements circumvents the need to model substrate dynamics, and thus greatly simplifies the modelling procedure.

Optimal control of greenhouse climate entails the operation of the control variables such that the economic return to the grower is maximised. The basis of this approach is a model that describes the crop behaviour under the influence of the indoor climate conditions which in turn are determined by the outside weather conditions and the exerted control actions. Several simulation experiments have been reported in the literature (e.g. Gutman et al. (1993) and van Henten (1994), but there is no record of any results of experiments using optimal control at an operational level in a real greenhouse. From January until June 1994, experiments have been performed in a real greenhouse with tomatoes, comparing two compartments of the same greenhouse, one controlled by a current commercial greenhouse climate control computer and one controlled by a receding horizon optimal control (RHOC) algorithm. This paper presents the optimal control problem and some preliminary results of the experiment.

Thus far, optimal control has primarily been investigated for seasonal crop growth optimization. On-line aspects have received much less attention. The decomposition between long term strategies and on-line control, however, is not trivial. Appreciable losses occur when set-points generated by seasonal optimization using nominal weather and the static physics assumption are put to the fully dynamic system. It is argued that the main product of seasonal optimization are not the state trajectories of the fast variables, but rather the state trajectories, and the co-state trajectories of the slow variables. The co-states represent marginal values of the slow variables. The actual on-line control requires the solution of a new dynamic optimization problem over a far shorter horizon, with a modified goal function incorporating the marginal value of all slow crop variables (not just the harvested parts). Feed-back and feed-forward to account for modelling errors and weather deviations can be accounted for by receding horizon optimal control, using the actual measured weather and fast states. An feasible economically optimal greenhouse crop control and operation system thus may consist of three major elements: (i) specification of constraints that pertain to unmodelled management aspects, possibly supplied by expert systems or other AI decision support, (ii) a solution to the seasonal slow crop response optimization using known long term nominal weather, and (iii) a receding horizon optimal control to generate short term controls responding to the actual weather and state, using crop state and co-state information from the slow optimization.

To evaluate microalgae production in large scale open ponds under different climatologic conditions, a model-based framework is used to study the effect of light conditions, water temperature and reactor design on trends in algae productivity. Scenario analyses have been done for two algae species using measured weather data of the Netherlands and Algeria. The effects of temperature control, photo-inhibition and using monthly or yearly fixed biomass concentrations are estimated by a sensitivity analysis. The calculation-based results show that climate conditions such as solar irradiation and temperature dynamics play an important role in open raceway ponds. In moderate climate zones low and high temperatures over a season suppress growth. At high latitudes this effect is important as light levels vary much during the day and between seasons. Optimal biomass concentrations in ponds depend on location, pond depth and algae species. Pond design, location and algae species interact and productivity cannot be based solely on general or assumed efficiencies. It is essential to select algae species that have a suitable growth rate, light absorption coefficient and the ability to grow over a broad temperature range. The presented approach gives a framework to validate specific cultivation systems.

This work establishes a relationship between dryer energy performance and controllability using energy balances and process resiliency analysis. It is shown that using the process gain matrix, the dryer energy efficiency can be reliably calculated with conditions for simultaneous controllability improvement established. By incorporating a drying rate modifying system such as a desiccant dehumidifier as an add-on, these conditions are shown to be achievable due to the extra dehumidification which can be manipulated using the additional degrees of freedom introduced by the sorption system. Due to the adsorbent regulation properties which are enhanced by high-temperature regeneration, the resilience of energy performance to disturbances is significantly improved compared to conventional dryers. Also, a desiccant system performance indicator, the "adsorber-regenerator net energy efficiency ARNEE" is introduced and it is shown that energy efficiency improvement is possible only if the ARNEE is greater than the energy efficiency of the stand-alone dryer. © 2013 Elsevier Ltd.

This work analyzes the synergy between two complementary unit operations - adsorbent dehumidification and drying - and presents a mixed integer nonlinear programming approach to optimize energy performance in a two-stage system. Combined with active constraint analysis, the adsorbent properties that promote energy performance are derived. Microporous adsorbents with higher sorption capacities at low vapor pressures and requiring higher regeneration temperatures are preferred for ambient air dehumidification in the first stage. For exhaust air dehumidification, mesoporous adsorbents with lower regeneration temperatures are preferred such that the exhaust air from the first regeneration stage can sufficiently regenerate them. For drying below 50 C, energy consumption reductions of about 70% are achieved compared to conventional dryers without adsorbent dehumidification depending on adsorbent properties. The results demonstrate the usefulness of superstructure optimization in matching the drying process with the capabilities of the adsorbents to enhance process synergy for improved energy efficiency. © 2013 American Chemical Society.

This work presents an approach to compute dryer energy efficiency using air flowrate step responses and establish a link between drying energy efficiency and process controllability. The approach is based on the temperature drop between the dryer inlet and outlet air under adiabatic conditions and so decouples water evaporation from heat loss and product heating effects on dryer temperature drop. As such, the computation is accurate even for dryers with significant heat losses where the traditional use of actual temperature drop measurements is grossly inaccurate. The approach is tested and verified on two experimental case studies involving significant heat losses: the first, a continuous fluidized-bed dryer (from literature); the second, a conventional and zeolite wheel-assisted batch dryer designed in the current study. © 2013 Copyright Taylor and Francis Group, LLC.

An integrated neural-fuzzy approach named NUFZY has been developed in order to model non-linear systems. A set of membership functions at the input layer is linked with a layer of rules, using pre-set parameters. The consequent part of the rules are combined to form the output value by some weighted sum. Very fast training of the weights is achieved by using the orthogonal least squares method, which also provides a method to efficiently remove redundant rules from the prototype rule base of NUFZY. A practical example involving tomato plant growth data is used to demonstrate the capability of NUFZY.

This work shows from energy balances and process resiliency analysis, the relationship between dryer controllability and energy performance. It is shown that using the process gain matrix, the dryer temperature drop and hence, energy efficiency can be reliably calculated. Conditions necessary for simultaneous energy efficiency and controllability improvement are established. By pre-conditioning the drying air using a desiccant adsorption system, such conditions are shown to be achievable for the same input choice as a conventional dryer. Also, extra degrees of freedom introduced by the desiccant system promote controllability with the advantage best exploited using full multivariable control. Energy efficiency disturbance resilience is also improved. © 2012 Elsevier B.V.

This work presents a mixed integer nonlinear programming (MINLP) formulation for the design of energy-efficient multistage adsorption dryers within constraints on product temperature and moisture content. Apart from optimizing temperatures and flows, the aim is to select the most efficient adsorbent per stage and product to air flow configuration. Superstructure models consisting of commonly used adsorbents such as zeolite, alumina, and silica-gel are developed and optimized for a two-stage, low-temperature, adsorption drying system. Results show that the optimal configuration is a hybrid system with zeolite as the first-stage adsorbent and silica-gel as the second-stage adsorbent in counter-current flow between drying air and product. A specific energy consumption of 2,275 kJ/kg is achieved, which reduces to 1,730 kJ/kg with heat recovery by a heat exchanger. Compared to a conventional two-stage dryer at the same drying temperature, this represents a 59% reduction in energy consumption. The optimal system ensures the exhaust air temperature of the first-stage regenerator is high enough to regenerate the second-stage adsorbent so no utility energy is spent in the second stage. A higher second-stage adsorbent wheel speed favors energy performance as it becomes optimized for energy recovery while the first is optimized for dehumidification. Although this work considers three candidate adsorbents in a two-stage system, the same reasoning can be applied to systems with more stages and adsorbents. The developed superstructure optimization methodology can, by extension, be applied to optimize multistage hybrid drying systems in general for any objective. © 2012 Copyright Taylor and Francis Group, LLC.

This work presents a mixed integer nonlinear programming method for the design of energy efficient multistage adsorption dryers within product temperature and moisture constraints. The aim is to find the adsorbent type and properties that minimize specific energy consumption. The results show that the adsorbents chosen in each stage have highest sorption capacities for corresponding drying air conditions. Heat requirements are matched so, zero utility energy is spent regenerating the second stage adsorbent. The adsorbent flow speeds are such that the first stage is optimized for dehumidification and the second for heat recovery. Overall, the optimal system reduces specific energy consumption by about 64% compared to a conventional system at the same drying temperature. Also, drying capacity is improved which permits the use of smaller dryers. © 2012 Elsevier B.V.

Low-temperature drying is important for heat-sensitive products, but at these temperatures conventional convective dryers have low energy efficiencies. To overcome this challenge, an energy efficiency optimization procedure is applied to a zeolite adsorption dryer subject to product quality. The procedure finds a trade-off between the improved drying capacity due to dehumidification and energy expenditure due to regeneration while incorporating product drying properties. By optimizing the regeneration air inlet temperature, drying air, adsorbent, and regeneration air flow rates as well as sensible and latent heat recovery from the regenerator exhausts, the energy efficiency is improved by up to 45% compared to the state-of-the-art. The high mass transfer effect of high temperatures is utilized in the regenerator to boost dehumidification while isolating the heat-sensitive dried product from the quality-degrading effect. A process optimization procedure for maximizing energy efficiency is applied to a zeolite adsorption dryer subject to product quality. By optimizing the regeneration air inlet temperature, drying air, adsorbent, and regeneration air flow rates as well as sensible and latent heat recovery from the regenerator exhausts, the energy efficiency is improved by up to 45% compared to the state-of-the-art. © 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

The cultivation of algae in photo-bioreactors shows similarities to crop cultivation in greenhouses, especially when the reactors are driven by sun light. Advanced methodologies for dynamic optimization and optimal control for greenhouses are known from earlier research. The aim here is to extend these methodologies to microalgae cultivated in a flat plate photo-bioreactor. A one-state space model for the algal biomass in the reactor is presented. The growth rate vs. light curve is parameterized on the basis of experimental evidence. Spatial distribution of light and growth rate between the plates is also considered. The control variable is the dilution rate. Dynamic optimal control trajectories are presented for various choices of goal function and external solar irradiation trajectories over a horizon of 3 days. It was found that the algae present in the reactor at final time represent a value for the future. Numerical and theoretical results suggest that the control is bang-(singular-)bang, with a strong dependence on the weather. The optimal biomass also depends on the available light, and achieving it to reach a new optimal steady cycle after a prolonged change in weather may take several days. A preliminary theoretical analysis suggests a control law that maximizes the effective growth rate. The analysis shows that like in the greenhouse case, the co-state of the algal biomass plays a pivot role in developing on-line controllers.

The systematic design of an autonomous platform for robotic weeding research in arable farming is described. The long term objective of the project is the replacement of hand weeding in organic farming by a device working autonomously at field level. The distinguishing feature of the described design procedure is the use of a structured design approach, which forces the designer to systematically review and compare alternative solution options, thus preventing the selection of solutions based on prejudice or belief. The result of the design is a versatile research vehicle with a diesel engine, hydraulic transmission, four-wheel drive and four-wheel steering. The robustness of the vehicle and the open software architecture permit the investigation of a wide spectrum of research options for intra-row weed detection and weeding actuators.

The shear stresses in printed scaffold systems for tissue engineering depend on the flow properties and void volume in the scaffold. In this work, computational fluid dynamics (CFD) is used to simulate flow fields within porous scaffolds used for cell growth. From these models the shear stresses acting on the scaffold fibres are calculated. The results led to the conclusion that the Darcian (k 1) permeability constant is a good predictor for the shear stresses in scaffold systems for tissue engineering. This permeability constant is easy to calculate from the distance between and thickness of the fibres used in a 3D printed scaffold. As a consequence computational effort and specialists for CFD can be circumvented by using this permeability constant to predict the shear stresses. If the permeability constant is below a critical value, cell growth within the specific scaffold design may cause a significant increase in shear stress. Such a design should therefore be avoided when the shear stress experienced by the cells should remain in the same order of magnitude.

Application of advanced controllers in horticultural practice requires detailed models. Even highly sophisticated models require regular attention from the user due to changing circumstances like plant growth, changing material properties and modifications in greenhouse design and layout. Moreover, their calibration is data demanding and laborious. This study explores the suitability of the extended Kalman filter (EKF) for automatic, on-line estimation and adaptation of parameters in a physics-based greenhouse model. The method was tested with measured data recorded over a period of 1 year, and with a model that describes the air temperature and moisture content in the Watergy greenhouse. In order to keep the parameters estimation problem tractable, and to improve the local accuracy of the parameters, separate EKFs are applied to sub-systems, using observation data at the sub-system boundaries. The filter adequately adjusts parameter values, thus significantly improving the model fit as compared to simulations with no-varying parameters. It appears that the filter is robust with respect to sudden changes in the system; when a disturbance occurs, such as pruning of plants or emergency opening of the windows, the EKF changes the parameter values accordingly. The result suggests that the extended Kalman filter is, indeed, a suitable method to provide the required automatic adaptation to time-varying phenomena when modeling is impractical.

This work discusses the evaluation of multi-stage adsorption dryers with air dehumidification by zeolite and alumina pillared clay. In a multi-stage dryer, product is dried in succeeding stages while air leaving a stage is fed to the next stage after dehumidification by an adsorbent. Energy efficiency of the drying system is evaluated for low-temperature drying (10-50°C) and compared with conventional condenser drying. Results showed that the efficiency of the multi-stage adsorption dryers increases with the number of stages. For low drying temperatures, zeolite is most favorable; for drying temperatures 40-50°C, alumina pillared clay needs less cooling and deserves preference.

Many papers describe applications of advanced controllers in greenhouses. As control literature focusses on control algorithms, the layout of the measurement and control system is usually underexposed. Unfortunately, commercially available greenhouse climate control systems do not have the necessary flexibility to accommodate these types of controllers. This paper describes a systematic approach to the design of a flexible measurement and control system, that can be adjusted to suit most control research in horticulture. Individual functions within a measurement and control system are identified and alternative solutions are given. The approach is applied to the design of the measurement and control system for the Watergy greenhouse. The controllers (in software) are versatile and setpoints can be generated by the easy-to-use user interface as well as by external software (like Matlab). Data is centrally stored and the system is easy to expand. It has been shown that the system has functioned robustly for the last 3 years. It has the capability to serve as basis for research and testing of various advanced control strategies in the Watergy system with adaptive model-based control as a final goal.

Measurement of the key process variables is essential during biopharmaceutical production. These measurements are often not available online. This work combines frequent online measurements (oxygen uptake rate) with infrequent offline measurements (biomass) to estimate the specific growth rate, biomass, and the oxygen transfer coefficient (kLa) online. The system consists of an extended Kalman filter and parameter adaptation for the time-varying kLa. Tuning is based on minimization of the error between the simulation and the estimation. Although the process itself is not stable, stability of the observer is evaluated heuristically by application of the Routh criterion. Performance and convergence of the observer are shown in both simulations and experiments in continuous and fed-batch cultivations of Bordetella pertussis.

The available time span for food product and process innovation is steadily decreasing, and to increase the efficacy of the development cycles, systematic design procedures can be used to develop new and to redesign existing processes. The Conceptual Process Design (CPD) methodologies used in chemical industry might also be applied in the food industry to rethink their systems and to break down the complexity of problems into several hierarchical levels. The Delft design matrix, a combination of the iterative design procedure, hierarchical decomposition and task driven methods, is a promising tool which can cover the different objectives and criteria of food design. This article analyzes the potential applicability of Delft Design Matrix as a CPD methodology for food process design and illustrates its functionality with the design of a bakery production system.

Quality driven design uses specified product qualities as a starting point for process design. By backward reasoning the required process conditions and processing system were found. In this work dynamic optimization was used as a tool to generate processing solutions for baking processes by calculating optimal operation strategies which give a basis for process and unit operation design. Two different approaches for dynamic optimization had been applied: calculation of continuous trajectories based on the calculus of variations (1) and calculation of switching trajectories by using control vector parameterization (2). Optimization of bakery processes was performed for different product specifications and by using different heating sources: convective, microwave and radiation heating. Moreover, effects of variations in dough properties on the optimal processing system were also evaluated. The results showed that dynamic optimization procedures were versatile tools to achieve a better and a more flexible design by generating a number of solutions from the specified final product qualities. Furthermore, the results underpinned the well known empirical fact that different final product specifications require different baking strategies. It was also shown that the initial dough properties have significant effect on baking procedures, and that combining several heating inputs improved the flexibility of the process operation. In addition, optimization for continuous trajectories gave overall a better result than using the switching trajectories.

Two-dimensional computational fluid dynamics calculations for multistage zeolite drying are performed for two dryer configurations (1) a continuous moving bed zeolite dryer and (2) a discrete bed zeolite dryer. The calculations concern drying of tarragon (Artemisia dracunculus L.) as an herbal product. The results reveal the profiles of water, vapor, and temperature in dryer, adsorber, and regenerator in the flow directions. The thermal efficiency ranges between 80 and 90% and is close to overall model calculations. The performance of continuous moving bed zeolite dryer is the best. Residence time of air, product, and zeolite are in accordance to other drying systems.

Model reduction, linearisation and discretisation are standard techniques in systems theory to simplify nonlinear models. In this paper, these techniques are used on a nonlinear model of a storage facility with forced ventilation with ambient air. Consequently, weather forecasts are needed to predict the systems’ trajectory. Because weather forecasts inherit uncertainty, the storage model is amenable to uncertainty. Based on a linearised version of the storage model, standard error propagation rules have been used to predict the system uncertainty analytically. Optimal control calculates controls in such a way that a prespecified cost criterion is minimised. As the uncertainty of the storage system is subject to ventilation with outside air, uncertainty of the model state increases with increased ventilation. The model uncertainty is integrated into the cost criterion to allow a trade-off between an optimal nominal solution and a minimum variance control solution. In this respect, the predicted 95% confidence limits are kept as close as possible to the reference trajectory.

This work presents the evaluation of multi-stage adsorption dryers with air dehumidification by zeolite and alumina pillared clay. In a multi-stage dryer product is dried in a number of succeeding stages and for each stage the air is dehumidified by the adsorbent. The exhaust air from each stage is reused for product drying in a following stage after passing a bed with adsorbent. The heat efficiency of the drying system is evaluated and compared with a conventional drying system using air dehumidified by condensation. The evaluation concerns the effect of stage number, inlet dryer temperatures 10-50°C and inlet ambient air 10-25°C. Results show that the performance of the multi-stage adsorption dryers increases with the number of stages and a three stage system is economically most promising. For varying ambient air conditions, zeolite as adsorbent is the most favourable.

In bakery production, product quality attributes as crispness, brownness, crumb and water content are developed by the transformations that occur during baking and which are initiated by heating. A quality driven procedure requires process optimization to improve bakery production and to find operational procedures for new products. Control vector parameterization (CVP) is an effective method for the optimization procedure. However, for accurate optimization with a large number of parameters CVP optimization takes a long computation time. In this work, an improved method for direct dynamic optimization using CVP is presented. The method uses a sensitivity based step size refinement for the selection of control input parameters. The optimization starts with a coarse discretization level for the control input in time. In successive iterations the step size was refined for the parameters for which the performance index has a sensitivity value above a threshold value. With this selection, optimization is continued for a selected group of input parameters while the other nonsensitive parameters (below threshold) are kept constant. Increasing the threshold value lowers the computation time, however the obtained performance index becomes less. A threshold value in the range of 10–20% of the mean sensitivity satisfies well. The method gives a better solution for a lower computation effort than single run optimization with a large number of parameters or refinement procedures without selection.

The development of food production processes is facilitated by tools which explore the interaction between process design, operation conditions and product characteristics. In this work an approach how to set-up a simulation model is presented for the phenomena and transformations which occur during baking and which fix the product quality. The simulation model has three consecutive parts: mass and heat transport in the product, transformations concerning starch state transition and color, and the formation of quality attributes (color, softness, crispness and staling). The model for mass and heat transfer is based on laws of conservation and expressed in partial differential equations for spatial products. The starch state transition and color formation are a mixture of qualitative and quantitative information, while the product quality model is mainly based on qualitative information. The model is applied to three bakery products: bread, biscuit and a cake-type. The results show that the model estimates the product quality and its transformations as a function of dough composition, baking and storage condition. The results fit well to observed changes of properties and product quality during baking. Industrial relevance Food industries require tools to evaluate processing options in the feasibility phase of process design. Therefore, simulation of process models is important for this purpose. However, knowledge of different aspects is subject to area of expertise (for example heat and mass transfer versus product quality formation) and often these areas are hardly connected. This work presents a systematic modeling approach for the dominant processes during baking and their interconnection. The main functions of the model are to explore the consequences of choices in design, to rank design options and to find in what direction properties will change when operational conditions change. Moreover, the model can be used for sensitivity analysis to explore on what items further information must be gathered.

This work discusses the potential of three multistage zeolite drying systems (counter-, co-, and cross-current) with a varying number of stages. The evaluation showed that for 2¿4 stages with heat recovery the efficiency of the systems ranges between 80 and 90%. Additionally, by introducing a compressor, the latent heat in the exhaust air from the regenerator is recovered and used to heat the inlet air for an additional drying stage. As a result, for the counter-current drying system and compressor pressure 1.5¿2 bar, a maximum energy efficiency of 120% is achieved, which results in halving the energy consumption compared to conventional drying systems.

A wide range of various classical and modern control methods is classified from the point of view of required information and put in the context of optimization.

Monitoring and control of production processes for biopharmaceuticals have become standard requirements to support consistency and quality. In this paper, a constant specific growth rate in fed-batch cultivation of Bordetella pertussis is achieved by a newly designed specific growth rate controller. The performance of standard control methods is limited because of the time-varying characteristics due to the exponentially increasing biomass and volume. To cope with the changing dynamics, a stable model reference adaptive controller is designed which adapts the controller settings as volume and biomass increase. An important asset of the design is that dissolved oxygen is the only required online measurement. An original design without considering the dissolved oxygen dynamics resulted experimentally in oscillatory behaviour. Hence, in contrast to common believes, it is essential to include dissolved oxygen dynamics. The robustness of this novel design was tested in simulation. The validity of the design was confirmed by laboratory experiments for small-scale production of B. pertussis. The controller was able to regulate the specific growth rate at the desired set point, even during a long fed-batch cultivation time with exponentially increasing demands for substrates and oxygen.

We will argue in this paper that the nature and magnitude of model uncertainty dictate the appropriateness of the control system design methodology. To obtain these arguments, we will pursue optimality of the control system design. Note this is a philosophical paper , in which we present our arguments in qualitative terms . We will identify the circumstances under which several control system design methodologies are appropriate. The design methodologies that will be considered are (1) optimal control and optimal feedback design, (2) model predictive control (receding horizon control), (3) active and passive adaptive control and (4) robust control. Robust control strategies are conservative and limit the possible improvements that can be obtained over say classical PID control. However, it will turn out that for certain types of control problems, improvements in performance are indeed limited. Roughly speaking, the area where we may expect (significant) improvement from advanced (optimal) control system design is very much limited by the uncertainty of the systems model. Miracles are not to be expected from advanced control. What can be expected is reflected in an active adaptive optimal control scheme introduced and discussed in this paper.

An optimisation procedure is presented for direct control of product quality of agro-material and minimisation of its quality variation. The procedure builds on a previously presented model structure, which is briefly reviewed, together forming a methodological framework for direct product quality control in climate controlled operations. The economic-based optimisation procedure is part of a hierarchical control structure and utilizes the model to achieve the required control. The result is a trade-off between operational cost and product quality. The characteristic variation in natural products is treated by a lumped approach, which uses discretised quality classes to describe not only product quality but also a measure of its variation. Although direct measurement of product properties is often difficult, the approach is validated using both simulated and measured data. The procedure has been implemented and tested in a full-scale industrial operation of storing large quantities of potatoes, using recorded real data to illustrate the benefits of the procedure.

A novel solar humid-air-collector system for combined water treatment, space-cooling and ¿heating has been designed in an EU framework-5 financed project called Watergy. The design consists of a construction of two prototypes for applications in architecture and greenhouse horticulture: a South European variant (for arid climate and emphasis on agricultural use) and a Central-/North European variant (for temperate climate and emphasis on integral building design). The core is the development of a collector system, following the principle of a closed, two phase thermo siphon. It achieves combined evaporation and ¿condensation, efficient heat transfer to a central heat exchanger as well as increased heat conduction from (humid) air to water. Main improvements are cost reduction in space heating and -cooling of buildings and greenhouses. Furthermore, viability increases by additional integration of greenhouse irrigation water recycling, desalination and building grey-water recycling. Sensors and actuators, connected to low-level controllers, activate a model-based control system to manage these processes. The paper describes the different appendages, sensor systems, network connections, databases, alarming systems, user interfaces and remote management from the Netherlands to Spain through the Internet.

The NICOLET model has been developed to predict the growth and nitrate content of greenhouse lettuce. Four single-organ versions have been developed: [1] abundant supply of nitrogen (1998), [2] mild N-stress (1999), [3] severe N-stress (2003), and [4] ontogenetic changes of organic-N and water content (in preparation). The `abundant-N¿ model [1] and the `mild-stress¿ model [2] have two compartments: `structure¿ and `vacuole¿, while the `severe-stress¿ model [3] requires a third compartment: `excess-carbon¿, and the `ontogenetic¿ model [4] has separate `metabolic¿ and `support¿ structural sub-compartments. The main special features of the NICOLET model are [1] the osmotica balance of the `vacuole¿, where nitrate and hydrocarbons play a complementary role in maintaining a constant osmotic potential, [2] the excess-carbon compartment, where `dry¿ carbon compounds are stored, and [3] the sub-division of the `structure¿ into sub-compartments of different compositions. Loosely speaking, the first feature controls the nitrate concentration, the second controls the organic-N and water contents, and the third controls the ontogenetic changes. The NICOLET model has been able to mimic `normal¿ seasonal variations of nitrate content, as well as the effects of drastic N-stress treatments. These results are illustrated by comparing measured data with model-simulations. Accurate prediction of nitrate concentration is difficult, due to its sensitivity to changes in the environment. Exact control of nitrate under commercial conditions may require transient corrective measures, such as N-interruption, in conjunction with a good plant-nitrate monitoring system.

A solar greenhouse has been designed that maximizes solar energy use and minimizes fossil energy consumption. It is based on a conventional greenhouse extended with a heat pump, a heat exchanger, an aquifer and ventilation with heat recovery. The aim is to minimize fossil energy consumption, while maximizing crop dry weight and keeping temperature and humidity within certain limits. These requirements are defined in a goal function, which is minimized by optimal control. A greenhouse with crop model is used to simulate the process behaviour. It is found that open loop optimal control trajectories can be determined. The boiler use is reduced to a minimum, thus reducing fossil energy use. Compared to a conventional greenhouse it is found that the energy costs are decreased and the crop dry weight is increased.

Many methods for solving optimal control problems, whether direct or indirect, rely upon gradient information and therefore may converge to a local optimum. Global optimisation methods like Evolutionary algorithms, overcome this problem. In this work it is investigated how well novel and easy to understand Evolutionary algorithms, referred to as Differential Evolution (DE) algorithms, and claimed to be very efficient when they are applied to solve static optimisation problems, perform on solving multimodal optimal control problems. The results show that within the class of evolutionary methods, Differential Evolution algorithms are very robust, effective and highly efficient in solving the studied class of optimal control problems. Thus, they are able of mitigating the drawback of long computation times commonly associated with Evolutionary algorithms. Furthermore, in locating the global optimum these Evolutionary algorithms present some advantages over the Iterative Dynamic Programming (IDP) algorithm, which is an alternative global optimisation approach for solving optimal control problems. At present little knowledge is available to the selection of the algorithm parameters in the DE algorithm when they are applied to solve optimal control problems. Our study provides guidelines for this selection. In contrast to the IDP algorithm the DE algorithms have only a few algorithm parameters that are easily determined such that multimodal optimal control problems are solved effectively and efficiently.

In the field of water technology, forward uncertainty propagation is frequently used, whereas backward uncertainty propagation is rarely used. In forward uncertainty analysis, one moves from a given (or assumed) parameter subspace towards the corresponding distribution of the output or objective function. However, in the backward uncertainty propagation, one moves in the reverse direction, from the distribution function towards the parameter subspace. Backward uncertainty propagation, which is a generalisation of parameter estimation error analysis, gives information essential for designing experimental or monitoring programmes, and for tighter bounding of parameter uncertainty intervals. The procedure of carrying out backward uncertainty propagation is illustrated in this technical note by working example for an oxidation ditch wastewater treatment plant. Results obtained have demonstrated that essential information can be achieved by carrying out backward uncertainty propagation analysis.

The Takagi-Sugeno (T-S) type of fuzzy model combines linguistic information and quantitative training procedures, and is therefore a most suitable candidate for accommodating additional a-priori knowledge in input-output modeling. The idea is to regard the a-priori knowledge as constraints that penalize a performance criterion used for identifying the unknown parameters in the T-S fuzzy model. The relative importance of various sources of a-priori knowledge, as expressed in the penalty weighting parameters, can be assessed by optimizing a generalized cross-validation criterion. A synthetic example is presented, showing that significant effects can be achieved by adding penalties to the optimization performance criterion. The identified T-S fuzzy model has comparative performance in the interpolation (training) range and a convincing improvement in the extrapolation (validation) range as compared to the T-S fuzzy model without adding any a-priori knowledge.

This paper presents validation and implementation results of a benchmark developed for a specific full-scale oxidation ditch wastewater treatment plant. A benchmark is a standard simulation procedure that can be used as a tool in evaluating various control strategies proposed for wastewater treatment plants. It is based on model and performance criteria development. Testing of this benchmark, by comparing benchmark predictions to real measurements of the electrical energy consumptions and amounts of disposed sludge for a specific oxidation ditch WWTP, has shown that it can (reasonably) be used for evaluating the performance of this WWTP. Subsequently, the validated benchmark was then used in evaluating some basic and advanced control strategies. Some of the interesting results obtained are the following: (i) influent flow splitting ratio, between the first and the fourth aerated compartments of the ditch, has no significant effect on the TN concentrations in the effluent, and (ii) for evaluation of long-term control strategies, future benchmarks need to be able to assess settlers' performance.

The recent market introduction of ion-selective sensors in horticulture removes one of the barriers towards accurate control of individual ions supply in closed water systems in greenhouse cultivation. In previous work, controllers have been designed that are able to compensate for evapotranspiration and nutrient uptake by tracing a set-point for pulse-averaged drain flow and individual ion concentrations in the drain, based on ion-selective sensor information. The objective of the current work is to investigate the desired operation mode of the nutrient controller under the assumption of successful constant drain flow control. Two cases are distinguished: demand satisfying control and supply regulating control. Using fundamental mass balances and transport equations, conditions are derived to which set-points of the constant drain concentration controllers should obey in order to ensure non-inhibiting nutrient supply. It is concluded that uptake regulation below the demand is most likely difficult to achieve with a drain concentration controller, whereas it is very suitable for demand satisfying control.

To enhance development and acceptance of new control strategies, a standard simulation benchmarking methodology to evaluate the performance of wastewater treatment plants has recently been proposed. The proposed methodology is, however, for a typical plant and typical loading and environmental conditions. Thus, benchmarking a full-scale plant working under different situations is still a problem that needs to be solved. This paper proposes a realistic approach to benchmark specific full-scale activated sludge plants used for carbon and nitrogen removal, based on real design, operational and performance data. An illustrative example is also presented in this paper.

In this paper a modelling and control structure for product quality control is presented for a class of operations that processes agro-material. This class can be characterised as climate-controlled operations, such as storage, transport and drying. The basic model consists of three parts. These are the quality and behaviour of the product, and its direct and indirect environment. This decomposition is reflected in the proposed control structure. The significance of the explicit inclusion of product behaviour is that the control is much more geared to the demands of the product. The applicability and potential advantages of the proposed structure are illustrated in different industrial cases, such as the storage of potatoes and the transport of apples in reefer-containers.

When applied to large simulation models, the process of parameter estimation is also called calibration. Calibration of complex non-linear systems, such as activated sludge plants, is often not an easy task. On the one hand, manual calibration of such complex systems is usually time-consuming, and its results are often not reproducible. On the other hand, conventional automatic calibration methods are not always straightforward and often hampered by local minima problems. In this paper a new straightforward and automatic procedure, which is based on the response surface method (RSM) for selecting the best identifiable parameters, is proposed. In RSM, the process response (output) is related to the levels of the input variables in terms of a first- or second-order regression model. Usually, RSM is used to relate measured process output quantities to process conditions. However, in this paper RSM is used for selecting the dominant parameters, by evaluating parameters sensitivity in a predefined region. Good results obtained in calibration of ASM No. 1 for N-removal in a full-scale oxidation ditch proved that the proposed procedure is successful and reliable.

Standard methods for the determination of oxygen transfer rate are based on assumptions that are not valid for oxidation ditches. This paper presents a realistic and simple new method to be used in the estimation of oxygen transfer rate in oxidation ditches from clean water measurements. The new method uses a loop-of-CSTRs model, which can be easily incorporated within control algorithms, for modelling oxidation ditches. Further, this method assumes zero oxygen transfer rates ( K L a ) in the unaerated CSTRs. Application of a formal estimation procedure to real data revealed that the aeration constant ( k = K L aV A , where V A is the volume of the aerated CSTR) can be determined significantly more accurately than K L a and V A . Therefore, the new method estimates k instead of K L a . From application to real data, this method proved to be more accurate than the commonly used Dutch standard method (STORA, 1980).

Since high concentration levels of nitrate in lettuce are undesirable, its control is currently an important problem in the context of European Union regulations. Using a dynamic model that predicts the amount of nitrate at harvest time, an optimal control problem is formulated and solved through an enhanced classical gradient method and a Differential Evolution algorithm. This work shows that in order to avoid local minima an efficient evolutionary algorithm may be applied to solve optimal control problems or to provide a good initial guess for a classical method, which solves smooth continuous-time optimal control problems more accurately and efficiently.

When applied to large simulation models, the process of parameter estimation is also called calibration. Calibration of complex non-linear systems, such as activated sludge plants, is often not an easy task. On the one hand, manual calibration of such complex systems is usually time-consuming, and its results are often not reproducible. On the other hand, conventional automatic calibration methods are not always straightforward and often hampered by local minima problems. In this paper a new straightforward and automatic procedure, which is based on the response surface method (RSM) for selecting the best identifiable parameters, is proposed. In RSM, the process response (output) is related to the levels of the input variables in terms of a first- or second-order regression model. Usually, RSM is used to relate measured process output quantities to process conditions. However, in this paper RSM is used for selecting the dominant parameters, by evaluating parameters sensitivity in a predefined region. Good results obtained in calibration of ASM No.1 for N-removal in a full-scale oxidation ditch proved that the proposed procedure is successful and reliable.

This paper demonstrates the application of the factorial sensitivity analysis methodology in studying the influence of variations in stoichiometric, kinetic and operating parameters on the performance indices of an oxidation ditch simulation model (benchmark). Factorial sensitivity analysis investigates the sensitivities in a region rather than in a point. Hence, it has the advantage of giving more information about parameter interactions (non-linearity). Short-term results obtained have shown the following. The index AE is not significantly affected by variations in the value of parameters of the activated sludge model (ASM) No 1. The index TSP is greatly influence by heterotrophic yield (YH), heterotrophic decay (bH) and specific hydrolysis (kh) and the index EQ is dominated by YH, Monod coefficient (KS), bH, kh, anoxic condition correction factors (g, h), hydrolysis half-saturation coefficient (KX), autotrophs maximum specific growth rate (?A) and ammonia half-saturation coefficient (KNH). Furthermore, the index EQ has been shown to be very sensitive to parameter interactions, at certain regions.

From the analysis of a mass balance of nutrients and water, a strategy for nutrient application in closed growing systems is suggested that could be of use in the design of control algorithms. This strategy is to keep the solute drain flow - time averaged over one supply cycle - at a constant level, and to adjust the incoming concentration of nutrients in proportion to the difference between incoming ion concentration and ion concentration in the drain.

Theoretically, using information about crop growth would allow the extension of present greenhouse control strategies towards a truly economic optimal control strategy. A brief survey is given of developments in the scientific literature. A full solution would require to consider the long-term crop development as well as all relevant short-term dynamics of the crop, the greenhouse and the external weather. Obstacles for the acceptance of such solutions are briefly discussed. One of the key factors is the lack of reliable crop development models for the wide variety of crops encountered in practice, and the need to leave part of the decision freedom in the hands of the grower. An analysis is given of simplified approaches resulting from integrating the crop equations over a day or more. The temperature integral concept, a specific example of such approach, is gaining popularity, despite the fact that it lacks exploitation of knowledge about the fast crop responses. The discussion leads to the concept of separation of responsibilities, where the short-term effects, including photosynthesis and evapo-transpiration, are handled by an automated model-predictive optimal controller, while the long-term effects are left to the grower, with support from a flexible decision support system based on crop models whenever they become available.