Search

By using small computing devices deployed at user premises, Autonomous Demand Response (ADR) adapts users electricity consumption to given time-dependent electricity tariffs. This allows end-users to save on their electricity bill and Distribution System Operators to optimise (through suitable time-dependent tariffs) management of the electric grid by avoiding demand peaks. Unfortunately, even with ADR, users power consumption may deviate from the expected (minimum cost) one, e.g., because ADR devices fail to correctly forecast energy needs at user premises. As a result, the aggregated power demand may present undesirable peaks. In this paper we address such a problem by presenting methods and a software tool (APD-Analyser) implementing them, enabling Distribution System Operators to effectively verify that a given time-dependent electricity tariff achieves the desired goals even when end-users deviate from their expected behaviour. We show feasibility of the proposed approach through a realistic scenario from a medium voltage Danish distribution network., 6 pages, 1 figure. In SmartGridComm 2018. IEEE, 2018

Recently, the topic of grid integration of distributed energy resources (DERs) has been closely linked with developments in power electronics research. The experimental evaluation of control strategies for power interfaces usually requires a significant allocation of both time and resources, and this is why the power hardware-in-the-loop (PHIL) techniques have been introduced to help accelerate the design and evaluation process. This article presents a PHIL-based microgrid test bed, Smart Energy Integration Lab (SEIL), that has been specifically designed for evaluation studies on the grid integration of DERs, with an emphasis on the system dynamic performance and power dispatch scenarios. The SEIL is equipped with a remotely controlled network configuration, power electronics converters, passive load banks, a battery storage system, and real-time (RT) control and acquisition systems that allow the emulation of several microgrids and connection of several DERs at the same time. To demonstrate the functionality of the SEIL, experimental results for two typical scenarios will be presented: 1) a power dispatch scenario in a microgrid and 2) a wind turbine (WT) operation in grid connection.

The continuous increase in energy demand and the integration of a large number of renewable energy sources with intermittent production require an advanced energy management to guarantee an uninterrupted supply. Deployment of smart meters in wide areas of power networks and availability of affordable storage systems encourage development and implementation of local energy management systems. Local energy management is considered a feasible and inexpensive approach to reduce the impact of load variations and intermittent generation. Demand response and load shaping techniques provide customers with the facility to contribute to system balancing and improve power quality. This paper presents a building energy management which determines the optimal scheduling of the components of the local energy system. The developed two-stage optimization is based on a receding horizon strategy and minimizes the energy costs under consideration of the physical system constraints. The proposed building energy management has been implemented and tested with a medium-size hotel emulated in a microgrid testbed using power-hardware-in-the-loop techniques. The obtained results underline that energy management strategies can be validated under realistic conditions using flexible power scenarios.

In modern societies buildings largely contribute to the energy demand and account in some countries for up to 45% of the primary energy consumption. Detailed consumption data can be used to offer advanced services and provide new business opportunities to all participants in the energy supply chain. In recent years, the intensified research effort by the scientific community and companies from the energy sector has led to an improved estimation of the building energy consumption. This paper proposes a tool that combines on one hand architectural characteristics and, on the other, user interactions with a building to precisely estimate its energy consumption and supply costs. The tool also determines the sensitivity of the energy consumption and the related costs for any variations applied in the building configuration or in the energy tariffs. The identification of critical parameters allows building managers to adopt appropriate measures to improve energy efficiency of the considered building. For the demonstration and validation of the proposed approach the developed tool is then applied to a case of an office building located in Madrid (Spain). It is explained in detail how the estimates of the building energy consumption and the costs were calculated and the sensitivity analysis performed.

This paper deals with the short-term forecasting of electrical energy demands at the local level, incorporating advanced metering infrastructure (AMI), or ‘smart meter’ data. It provides a study of the effects of aggregation on electrical energy demand modelling and multi-nodal demand forecasting. This paper then presents a detailed assessment of the variables which affect electrical energy demand, and how these effects vary at different levels of demand aggregation. Finally, this study outlines an approach for incorporating AMI data in short-term forecasting at the local level, in order to improve forecasting accuracy for applications in distributed energy systems, microgrids and transactive energy. The analysis presented in this study is carried out using large AMI data sets comprised of recorded demand and local weather data from test sites in two European countries. The authors kindly acknowledge the support of the European Commission provided through the Marie Curie researcher mobility action (FP7-PEOPLE-2013-COFUND), and the SmartHG research project (FP7-ICT-2011-8). peer-reviewed

This paper discusses the design and simulation of an integrated load forecasting and state estimation tool for distribution system operations. A predictive database is created and applied to forecast future network states in order to allow short-term (e.g., hours/days ahead) planning to be carried out. The predictive database is based on adaptive nonlinear auto-regressive exogenous (NARX) load estimation and forecasting models, which are continuously updated using feedback from the state estimator. This creates a closed-loop information flow designed to continuously monitor and improve the system state estimation performance by updating and retraining models where appropriate. The aim of this methodology is to improve situational awareness and help to provide network operators with early warning of potential issues, in medium voltage (MV) networks where the number of on-line measurements is limited, and state estimation relies heavily on estimates of power injections. The applicability of the approach is demonstrated through simulation using supervisory control and data acquisition (SCADA) and smart meter measurements recorded from an actual MV distribution network.

The increasing trends of energy demand and renewable integration call for new and advanced approaches to energy management and energy balancing in power networks. Utilities and network system operators require more assistance and flexibility shown from consumers in order to manage their power plants and network resources. Demand response techniques allow customers to participate and contribute to the system balancing and improve power quality. Traditionally, only energy-intensive industrial users and large customers actively participated in demand response programs by intentionally modifying their consumption patterns. In contrast, small consumers were not considered in these programs due to their low individual impact on power networks, grid infrastructure and energy balancing. This paper studies the flexibility of aggregated demands of buildings with different characteristics such as shopping malls, offices, hotels and dwellings. By using the aggregated demand profile and the market price predictions, an aggregator participates directly in the day-ahead market to determine the load scheduling that maximizes its economic benefits. The optimization problem takes into account constraints on the demand imposed by the individual customers related to the building occupant comfort. A case study representing a small geographic area was used to assess the performance of the proposed method. The obtained results emphasize the potential of demand aggregation of different customers in order to increase flexibility and, consequently, aggregator profits in the day-ahead market. The authors kindly acknowledge the support of the Spanish Ministry of Economy and Competitiveness project RESmart (ENE2013-48690-C2-2-R). peer-reviewed 2019-04-26

Summary In model predictive control (MPC), the input sequence is computed, minimizing a usually quadratic cost function based on the predicted evolution of the system output. In the case of nonlinear MPC (NMPC), the use of nonlinear prediction models frequently leads to non-convex optimization problems with several minimums. This paper proposes a new NMPC strategy based on second order Volterra series models where the original performance index is approximated by quadratic functions, which represent a lower bound of the original performance index. Convexity of the approximating quadratic cost functions can be achieved easily by a suitable choice of the weighting of the control increments in the performance index. The approximating cost functions can be globally minimized by convex optimization techniques in order to compute the input sequence. The minimization of the performance index is carried out by an iterative optimization procedure, which guarantees convergence to the solution. Furthermore, for a nominal prediction model, asymptotic stability for the proposed NMPC strategy can be shown. In the case of considering an estimation error in the prediction model, input-to-state practical stability is assured. The control performance of the NMPC strategy is illustrated by experimental results. Copyright © 2014 John Wiley & Sons, Ltd.

Detailed residential energy consumption data can be used to offer advanced services and provide new business opportunities to all participants in the energy supply chain, including utilities, distributors and customers. The increasing interest in the residential consumption data is behind the roll-out of smart meters in large areas and led to intensified research efforts in new data acquisition technologies for the energy sector. This paper introduces a novel model for generation of residential energy consumption profiles based on the energy demand contribution of each household appliance and calculated by using a probabilistic approach. The model takes into consideration a wide range of household appliances and its modular structure provides a high degree of flexibility. Residential consumption data generated by the proposed model are suitable for development of new services and applications such as residential real-time pricing schemes or tools for energy demand prediction. To demonstrate the main features of the model, an individual household consumption was created and the effects of a possible change in the user behaviour and the appliance configuration presented. In order to show the flexibility offered in creation of the aggregated demand, the detailed simulation results of an energy demand management algorithm applied to an aggregated user group are used.

The mathematical model used in Min–Max MPC (MMMPC) to predict the future trajectory of the system explicitly considers disturbances and uncertainties. Based on the future trajectory, the control sequence is computed minimizing the worst case cost with respect to all possible trajectories of the disturbances and uncertainties. This approach leads to a more robust control performance but also complicates the practical implementation of MMMPC due to the high computational burden required to solve the optimization problem. This computational burden is even worse if a nonlinear prediction model is used. In fact, to the best of the authors’ knowledge, there have not yet been reported any applications of nonlinear MMMPC to real processes. In this paper a nonlinear MMMPC strategy based on a second order Volterra series model is presented. The particular structure of the used prediction model allows to obtain an explicit formulation of the worst case cost and its computation in polynomial time. Real time applications with typical prediction and control horizons are possible because of the reduced complexity of the proposed control strategy. Furthermore, input-to-state practical stability for the proposed control strategy is guaranteed under certain conditions. The MMMPC strategy is implemented and validated in experiments with a continuous stirred tank reactor whose temperature dynamics are approximated by a second order Volterra series model. The control performance of the proposed MMMPC strategy is illustrated by the obtained experimental results.

The particular industrial interest in fuel cells is based on the possibility to generate clean energy both for stationary and automotive applications. Performance and safety issues of fuel cells are closely related to the control strategy used for the fuel and oxidant supply. In PEM (polymer electrolyte membrane or proton exchange membrane) fuel cells the oxygen excess ratio expresses the proportion between oxygen reacting in the cells and oxygen entering the stack and represents a decisive variable for the mentioned issues. This work is focused on the design of a nonlinear model predictive control (NMPC) strategy manipulating the air flow rate in order to maintain the oxygen excess ratio in a desired value, both for safety and performance reasons. The designed NMPC, based on a second order Volterra series model, was implemented on a commercial fuel cell. The proposed NMPC strategy is validated in experiments and compared to a linear model predictive controller and to the original built-in controller.

Suitable environmental conditions are a fundamental issue in greenhouse crop growth and can be achieved by advanced climate control strategies. In different climatic zones, natural ventilation is used to regulate both the greenhouse temperature and humidity. In mild climates, the greatest problem faced by far in greenhouse climate control is cooling, which, for dynamical reasons, leads to natural ventilation as a standard tool. This work addresses the design of a nonlinear model predictive control (NMPC) strategy for greenhouse temperature control using natural ventilation. The NMPC strategy is based on a second-order Volterra series model identified from experimental input/output data of a greenhouse. These models, representing the simple and logical extension of convolution models, can be used to approximate the nonlinear dynamic effect of the ventilation and other environmental conditions on the greenhouse temperature. The developed NMPC is applied to a greenhouse and the control performance of the proposed strategy will be illustrated by means of experimental results.

Min–Max MPC (MMMPC) offers the possibility to consider disturbances and uncertainties in the mathematical model used to predict the future trajectory of the system. The explicit consideration of disturbances and uncertainties in order to obtain a more robust control performance complicates the practical implementation of MMMPC due to the high computational burden required to compute the control law. The computational complexity of the optimization problem can be reduced by using approximate solutions or upper bounds of the worst case cost of the objective function. A computationally efficient MMMPC strategy based on such an upper bound was presented in a previous work also published in this journal (see Section 1 ). One of the main drawbacks of that strategy is the lack of a stability guarantee. In this paper it is shown that input-to-state practical stability of the MMMPC strategy can be guaranteed if a certain initial condition and a semi-feedback approach are used. Furthermore, the MMMPC strategy is validated in experiments with a continuous stirred tank reactor in which the temperature of the reactor is controlled. The behavior of the system and the controller are illustrated by means of experimental results.

There is a large demand to apply nonlinear algorithms to control nonlinear systems. With algorithms considering the process nonlinearities, better control performance is expected in the whole operating range than with linear control algorithms. Three predictive control algorithms based on a Volterra model are considered. The iterative predictive control algorithm to solve the complete nonlinear problem uses the non-autoregressive Volterra model calculated from the identified autoregressive Volterra model. Two algorithms for a reduced nonlinear optimization problem are considered for the unconstrained case, where an analytic control expression can be given. The performance of the three algorithms is analyzed and compared for reference signal tracking and disturbance rejection. The algorithms are applied and compared in simulation to control a Wiener model, and are used for real-time control of a chemical pilot plant. Copyright © 2009 John Wiley & Sons, Ltd.

The practical implementation of min–max MPC (MMMPC) is limited by the computational burden required to compute the control law. This problem can be circumvented by using approximate solutions or upper bounds of the worst possible case of the performance index. In a previous work, the authors presented a computationally efficient MMMPC control strategy in which a close approximation of the solution of the min–max problem is computed using a quadratic programming problem. In this paper, this approach is validated through its application to a pilot plant in which the temperature of a reactor is controlled. The behavior of the system and the controller are illustrated by means of experimental results.

Fuel cells represent an area of great industrial interest due to the possibility to generate clean energy, especially PEM (Polymer Electrolyte Membrane or Proton Exchange Membrane) fuel cells for their use in stationary and automotive applications. The oxygen excess ratio expressing the proportion between oxygen reacting in the cells and oxygen entering the stack is widely used to guarantee safety and to reach a high performance. To control the oxygen excess ratio a model predictive control (MPC) has been developed using the compressor motor voltage to manipulate the air flow rate entering the fuel cell. The design of the MPC is carried out with a commercial fuel cell. The proposed MPC is implemented and validated in experiments and compared with the original built-in controller. The behaviour of the system and the controller is illustrated by means of experimental results.

Model-based control techniques are commonly applied to control the greenhouse climate. As well-known, these techniques require accurate models for adequate results. Several first-principle models have been developed for the greenhouse climate control problem considering all the physical and physiological processes. However, these models are too complex to be used for control purposes. On the other hand, empirical models based on input/output real data allow to obtain better results and less complex model structures. In the Mediterranean areas the main problem is cooling the greenhouse and this leads to natural ventilation as a standard tool. This paper presents the development and the results of a Volterra model for the greenhouse temperature including the crop effect and using natural ventilation.

[EN] The present publication demonstrates the application of a nonlinear predictive control (NMPC) strategy based on Volterra models to a pilot plant in which the temperature of the reactor's content is controlled. The control is based on a second-order diagonal Volterra model in order to consider nonlinear effects. To calculate the control action, an iterative method with low computational cost, is used. The behavior of the process and the controller will be presented by means of experimental results. Finally, the experimental results of the NMPC will be compared to the results of a linear model based predictive controller (MPC)., [ES] El presente artículo muestra la aplicación de una estrategia de control predictivo no lineal basado en modelos (Nonlinear Model Predictive Control, NMPC) a una planta piloto en la que se controla la temperatura de un reactor. El controlador está basado en un modelo diagonal de Volterra de segundo orden para considerar efectos no lineales. Para calcular la acción de control se utiliza un método iterativo que requiere pocos recursos computacionales. El comportamiento del proceso y del controlador se presenta mediante resultados experimentales. Finalmente se comparan los resultados experimentales del NMPC con los resultados de un controlador predictivo lineal (MPC)., Este trabajo ha sido financiado parcialmente por el Ministerio de Educación y Ciencia a través del proyecto de investigación DPI-2004-07444-C04-01

Recent developments in active distribution networks, and the availability of smart meter data has led to much interest in Short-Term Load Forecasting (STLF) of electrical demand at the local level, e.g. estimation of loads at substations, feeders, and individual users. Local demand profiles are volatile and noisy, making STLF difficult as we move towards lower levels of load aggregation. This paper examines in detail the correlations between demand and the variables which influence it, at various levels of load disaggregation. The analysis investigates the forecasting capability of both linear and non-linear STLF approaches for forecasting local demands, and quantifies the forecast uncertainty for each level of load aggregation. The results demonstrate the limitations of several of the most commonly-used STLF approaches in this context. It is shown that, at the local level, standard STLF models may not be effective, and that simple load models created from historical smart meter data can give similar prediction accuracies. The analysis in the paper is carried out using two large smart meter data sets recorded at distribution networks in Denmark and in Ireland.

In order to optimally manage a modern electricity distribution network, peaks in residential users demand should be avoided, as this can reduce energy and network asset management costs. Furthermore, this must be done without compressing residential users demand. To this aim, in a demand response setting, residential users are given a price policy, which economically motivates them to shift their loads in order to achieve this goal. However, if the price policy for all users is similar, this demand response may result in simply shifting the demand peaks (peak rebound), leaving the problem unsolved. In this paper we propose a novel methodology which i) for each network substation s, automatically computes the desired power profile to be kept in order to optimally manage the network itself, ii) for each network substation s, automatically synthesizes individualized price policies for residential users connected to s, so that s is kept at the desired profile. Note that price policies individualization avoids the peak rebound problem, as different users have different low tariff areas. Furthermore, our methodology measures the flexibility of a residential user as the capacity needed by a home energy storage system (e.g., a battery) to always follow the given price policy, thus mitigating residential users discomfort. We show the feasibility of our approach on a realistic scenario taken from an existing medium voltage Danish distribution network.

The SmartHG project goal is to develop a suite of integrated software services (the SmartHG Platform) aiming at steering residential users energy demand in order to: keep operating conditions of the electrical grid within given healthy bounds, minimize energy costs, and minimize CO2 emissions. This is achieved by exploiting knowledge (demand awareness) of electrical energy prosumption of residential users as gained from SmartHG sensing and communication infrastructure. This paper describes such an infrastructure along with user demand patterns emerging from the data gathered from ~600 sensors installed in ~40 homes participating in SmartHG test-beds.

Grid integration of distributed energy resources (DERs) has been a hot topic in the field of power electronics in recent years. Time and resources required in experimental verification of control algorithms and equipment can be significantly reduced by using power-hardware-in-the-loop (PHIL) techniques. The paper presents the Smart Energy Integration Lab (SEIL), a PHIL based microgrid testbed specifically built for analyzing grid integration aspects of DERs with a focus on system dynamic performance and power dispatch scenarios. The SEIL consists of a remotely-controlled AC network configuration, power electronics converters, passive load banks, a battery storage system and real-time control and data acquisition systems. By using these resources, the laboratory allows configuration and simultaneous operation of several microgrids. To demonstrate the functionality of the SEIL, a power dispatch scenario applied to a microgrid was used.

[ES] Este artículo trata el problema del diseño de un controlador predictivo para seguimiento de referencias cambiantes, en el caso de sistemas no lineales. Los controladores predictivos proveen leyes de control adecuadas para regular sistemas lineales o no lineales a un punto de equilibrio dado garantizando la satisfacción de restricciones y la estabilidad asintótica. Pero si este punto de equilibrio cambia, el controlador podría perder la estabilidad o incluso la factibilidad y por lo tanto sería incapaz de seguir la referencia deseada. En (Ferramosca et al., 2009a) se ha propuesto un controlador predictivo para seguimiento de referencias capaz de garantizar factibilidad y convergencia al punto de equilibrio a pesar de los cambios que este pueda sufrir. En este artículo, este controlador se utiliza para controlar en tiempo real una planta piloto de procesos. Los resultados obtenidos demuestran que el controlador predictivo para seguimiento es capaz de controlar plantas con dinámicas no lineales y restricciones. El experimento demuestra cómo el controlador garantiza estabilidad, factibilidad y convergencia también en caso de referencias no alcanzables., [EN] This paper deals with the tracking problem for constrained nonlinear systems using a model predictive control (MPC) law. MPC provides a control law suitable for regulating constrained linear and nonlinear systems to a given target steady state. However, when the target operating point changes, the feasibility of the controller may be lost and the controller fails to track the reference. Recently, a novel MPC for tracking constrained nonlinear systems has been presented (Ferramosca et al., 2009a). This is capable to steer the system to any reference, even in the case of changing references. In this paper, this controller is used for the real-time control of a chemical pilot plant. The obtained experimental results demonstrate that the MPC for tracking is suitable for the control of plants with nonlinear dynamics since it ensures stability and offset-free convergence in case of large changes in the reference even using short prediction horizons. Besides, in case of unreachable set points, the controller steers the system to the closest reachable equilibrium point., Este trabajo ha sido realizado gracias al apoyo del proyecto Europeo HD-MPC (FP7-ICT- 223854) y del Ministerio de Educación y Ciencia Español (contracto DPI2008-05818).

Energy demand management schemes, applied to individual and aggregated residential demands, have advanced significantly in recent years. In order to improve the demand models used for the management schemes there have been several attempts to create a representative model of a household's consumption. This paper presents a novel approach to generation of energy consumption profiles for one or several households based on the energy demand contribution of each household appliance and calculated by using a probabilistic approach. The proposed energy demand model takes into consideration a wide range of household appliances and provides a high degree of flexibility by benefiting from its modular structure. Artificially generated profiles of detailed energy consumption can be readily used for testing and verification of demand control techniques designed for both individual and aggregated household demands. The features of the proposed energy demand model along with its configuration are demonstrated by generating and illustrating the profiles of consumer energy demand.

This paper presents a novel approach to use second order Volterra series models in nonlinear model predictive control. A common technique in model predictive control is the minimization of a quadratic cost function with respect to the future input sequence. In the case of nonlinear models, the resulting cost function is a possibly non-convex function. The proposed strategy uses quadratic cost functions to approximate the original cost function. For the quadratic cost functions, convexity can be achieved easily by adding a weighting function of the control increments. The approximated convex cost functions are minimized globally by means of an iterative approach with guaranteed convergence. The proposed control strategy is applied to a continuous stirred tank reactor and the control performance is illustrated by experimental results.

This paper shows the application of a Min-Max Model Predictive Control (MMMPC) strategy to a pilot plant in which the temperature of a reactor is controlled. An approximation of the worst case cost is used to obtain the control action. This approximation can be easily computed yielding a solution of the min-max problem very close to the exact one. The complexity of the algorithm allows the real time implementation for typical prediction and control horizons. The behavior of the system and the controller will be illustrated by means of experimental results.

This document presents the design of a hierarchical control to regulate the oxygen excess ratio of a fuel cell. The master controller calculates the necessary air flow to stabilize the oxygen excess ratio at a fixed set point. A nonlinear model based predictive controller (NMPC) using a Volterra series model is used as a master controller. The slave controller, a nonlinear PI, uses the reference of the air flow calculated by the master controller to stabilize the air flow in the compressor and allows reference tracking. The proposed control strategy is applied to full nonlinear model of a fuel cell in which simulations are carried out.

The practical implementation of min-max MPC (MMMPC) controllers is limited by the computational burden required to compute the control law. This problem can be circumvented by using approximate solutions or upper bounds of the worst possible case of the performance index. In a previous work, the authors presented a computationally efficient MMMPC control strategy in which a close approximation of the solution of the min-max problem is computed using a quadratic programming problem. In this paper, this approach is validated through its application to a pilot plant in which the temperature of a reactor is controlled. The behavior of the system and the controller are illustrated by means of experimental results.

[ES] Este trabajo muestra la aplicación de una estrategia de control predictivo mín-máx basado en modelos (Min-Max Model Predictive Control, MMMPC) a una planta piloto en la que se controla la temperatura de un reactor. Para calcular la acción de control se utiliza una aproximación del peor caso, que permite calcular una solución muy cercana a la del problema mín-máx con un coste computacional mucho menor. Por tanto, es un algoritmo cuya complejidad hace factible una implementación en tiempo real para valores típicos de los horizontes de predicción y control. El comportamiento del proceso y del controlador se ilustran mediante resultados experimentales., [EN] This work shows the application of a min-max model predictive control (MMMPC) strategy to a pilot plant in order to control the temperature of a reactor. An approximation of the worst case cost is used to compute the control action. This approximation allows us to compute a solution very close to that of the original min-max problem, but with a much lower computational burden. Due to its reduced complexity, it is easy to implement this strategy in real time for typical values of the prediction and control horizons. The behaviour of the process and the controller is illustrated by means of experimental results., Este trabajo ha sido financiado parcialmente por el Ministerio de Educación y Ciencia a través de los proyectos de investigación DPI2004-07444 y DPI2005-04568.

In recent years, the energy sector has undergone an important transformation as a result of technological progress and socioeconomic development. The continuous integration of renewable technologies drives the gradual transition from the traditional business model based on a reduced number of large power plants to a more decentralized energy production. The increasing energy demand and intermittent generation of renewable energy sources require modern control strategies to provide an uninterrupted service and guarantee high energy efficiency. Utilities and network operators permanently supervise production facilities and grids to compensate any mismatch between production and consumption. The enormous potential of local energy management contributes to grid stability and can be used to reduce the adverse effects of load variations and production fluctuations. This paper presents a building energy management which determines the optimal scheduling of all components of the local energy system. The two-stage optimization is based on a receding horizon strategy and minimizes two economic functions subject to the physical system constraints. The performance of the proposed building energy management is validated in simulations and the results are compared to the ones obtained with other energy management approaches.

In mild climates, the greatest problem faced by far in greenhouse climate control is cooling, which, for economical reasons, leads to natural ventilation as a standard tool. The nonlinear relationship between ventilation and temperature can be captured by Volterra models. These models represent the simple and logical extension of convolution models and can be successfully applied in nonlinear model-based predictive control. This paper presents the development of a nonlinear model predictive controller (NMPC) based on the identification of a Volterra model from input/output data considering the natural ventilation and the most relevant disturbances acting on the system. Finally, the NMPC is applied to a detailed simulation model of the greenhouse and the control behavior will be illustrated by means of simulation results.