scispace - formally typeset
Search or ask a question
Journal Article•DOI•

Use of model predictive control and weather forecasts for energy efficient building climate control

Frauke Oldewurtel1, Alessandra Parisio2, Colin N. Jones3, Dimitrios Gyalistras1, Markus Gwerder4, Vanessa J. Stauch5, Beat Lehmann6, Manfred Morari1 •
ETH Zurich1, University of Sannio2, École Polytechnique Fédérale de Lausanne3, Siemens Building Technologies4, MeteoSwiss5, Swiss Federal Laboratories for Materials Science and Technology6
01 Feb 2012-Energy and Buildings (Elsevier)-Vol. 45, Iss: 45, pp 15-27
TL;DR: In this paper, the authors investigated how ModelPredictive control and weatherpredictions can increase the energy efficiency in Integrated Room Automation (IRA) while respecting occupant comfort.
About: This article is published in Energy and Buildings.The article was published on 2012-02-01 and is currently open access. It has received 1070 citations till now. The article focuses on the topics: Model predictive control & HVAC.

Summary (6 min read)

Jump to: [1.1. Literature review] – [1.2. Main idea and outline of the paper] – [2. Introduction to model predictive control for building climate control] – [Cost function type Mathematical description] – [2.1. Cost function] – [2.2. Constraints] – [2.3. Current state] – [2.4. Dynamics] – [3. Control strategies and benchmarks for building climate control] – [3.1.1. Rule-Based Control (RBC)] – [3.1.2. Deterministic MPC (DMPC)] – [3.2.1. Performance Bound (PB)] – [4. Implementation of MPC] – [. Weather forecast] – [4.1.3. Local measured weather] – [4.2.1. Building modeling] – [4.2.3. Building model data] – [4.3.1. MPC formulation] – [• Deterministic MPC (DMPC)] – [• Stochastic MPC with chance constraints] – [4.3.4. Soften constraints] – [4.4. Step 4: solve optimization problem] – [5.1. Controller assessment concept] – [Q2 Prediction horizon length] – [Q3 Performance of DMPC] – [Q4 Performance of SMPC] – [Q5 Importance of Weather Predictions] – [Q1 Theoretical savings potential] – [Q5 Importance of weather predictions] and [7. Conclusion]

1.1. Literature review

  • In conclusion, from the literature it can be learned that predictive control strategies as well as including automated blinds and lighting in the control action provide potential benefits for energy efficient building climate control.
  • The reasons for MPC in buildings being rarely used until now are primarily the difficulties/costs of obtaining a model of an building that can be used in the MPC controller and the fact that energy costs played a minor role in the past.
  • The use of simulation tools in building planning are becoming standard and can help to obtain models for the MPC controller.
  • The quality of weather predictions is increasing and hence its usefulness for building climate control.
  • Energy costs are rising, and finally, there is a desire to handle time-varying electricity prices and the possibility of MPC to do so.

1.2. Main idea and outline of the paper

  • The paper aims at introducing the possibilities and new developments in Stochastic MPC to the building research and development community; it is organized as follows.
  • Section 2 explains the basic concepts of MPC with emphasis on the use for building climate control.
  • Section 3 describes the control strategies as well as a benchmark for building climate control that are used in the presented study.
  • Also the newly developed Stochastic MPC approach is introduced here.
  • Section 5 reports on the performance of the proposed controllers and, in Section 6, simulation results are presented and discussed.

2. Introduction to model predictive control for building climate control

  • In terms of building climate control, this means that at the current point in time, a heating/cooling, etc. plan is formulated for the next several hours to days, based on predictions of the upcoming weather conditions, see Fig. 1 .
  • Predictions of any other disturbances (e.g., internal gains), as well as time-dependencies of the control costs (e.g., dynamic electricity prices) or of the constraints (e.g., thermal comfort range) can be readily included in the optimization.
  • The first step of the control plan is applied to the building, determining the setting of all the heating, cooling and ventilation.

Cost function type Mathematical description

  • By this receding horizon approach feedback is introduced into the system, since the new optimal control problem solved at the beginning of the next time interval is a function of the new measured state at that point in time and hence of any disturbances that have meanwhile acted on the building.
  • The cost function and the constraints are the main pieces of the MPC design, the current state is used as the initial state for control predictions, and the dynamics of the system have to be modeled to a reasonable precision such that a good control performance is achieved.

2.1. Cost function

  • Be equivalent to the cost of the Linear Quadratic Regulator/Linear Quadratic Gaussian controller (classic optimal control problem).
  • If one wishes to minimize 'amounts', outliers, or economically motivated signals, then the linear cost function is more suitable than the quadratic one.
  • This cost function would also be a common choice for minimizing energy consumption of buildings.

2.2. Constraints

  • This is the most common type of constraint and is used to put upper and/or lower bounds on variables.
  • Since an optimization problem can only be solved if all variables are deterministic, chance constraints need to be reformulated into deterministic constraints.
  • This type of constraint is called conic, since the feasible region of the constraint has the form of a cone.
  • Second order cone constraints can -under special circumstances -result from reformulations of chance constraints.
  • This is a common type of constraint in hybrid systems, i.e. systems that exhibit both continuous and discrete time behavior.

2.3. Current state

  • The system model is initialized to the measured current state of the building and all predictions begin from the system in this initial state.
  • If some states cannot be measured but are observable, a Kalman filter would commonly be used for state estimation.

2.4. Dynamics

  • The system model is a critical piece of the MPC controller.
  • In the presented investigation a bilinear model was used, which will be derived in Section 4.2.

3. Control strategies and benchmarks for building climate control

  • Three different control strategies are presented that are compared in the investigation: Rule-Based Control (RBC), Deterministic MPC (DMPC), and Stochastic MPC (SMPC).
  • The control strategies under investigation are DMPC, which is a standard MPC formulation and SMPC, which is a newly developed MPC strategy for the purpose of building climate control that can handle the uncertainties resulting from the use of weather forecasts.
  • How the costs and constraints are defined for the MPC strategies is detailed in Section 4.
  • A second benchmark is given by the so-called Performance Bound (PB), which is defined as optimal control with perfect information, in particular with a perfect weather prediction, i.e. the realization is equal to the prediction.

3.1.1. Rule-Based Control (RBC)

  • The current control practice in Integrated Room Automation is RBC.
  • RBC determines all control inputs based on a series of rules of the form "if condition, then action".
  • The conditions and actions are usually associated with numerical parameters (e.g., threshold values) that need to be chosen.

3.1.2. Deterministic MPC (DMPC)

  • DMPC is the standard MPC approach that is used in virtually all commercial MPC applications.
  • It uses the imperfect/uncertain weather forecast and takes its control decision under the assumption that the predictions are correct (i.e. equal to certain).
  • Therefore, it is also often called Certainty Equivalence Control.

3.2.1. Performance Bound (PB)

  • For analysis purposes, it is possible to formulate such a problem for a given building and a given year after the weather has been recorded and hence is known.
  • PB is itself not a controller but rather a concept, that can serve as a benchmark for the investigated control strategies.
  • In order to compute PB, the same MPC algorithm was used in the investigation as for DMPC, but with perfect weather predictions.
  • To compute PB, a prediction horizon of seven days and a control horizon (i.e. the number of time steps that control inputs are applied in open-loop) of three days were used.

4. Implementation of MPC

  • It is explained how the various inputs to the controller (weather predictions, local weather and building measurements, building model data, etc.) are translated to a mathematical structure, that can be processed by standard optimization software.
  • This picture also gives an outline of this section.

. Weather forecast

  • The weather predictions were given by archived forecasts of the numerical weather prediction model COSMO-7 operated by MeteoSwiss [22] .
  • The forecast data comprised the outside air temperature, the wetbulb temperature and the incoming solar radiation.
  • For this study, four meteorological measurement sites in different European countries were chosen in order to represent different climatic zones in Europe.

4.1.3. Local measured weather

  • For the local measured weather, archived measurements of the Swiss Meteorological Network at the same sites as given in Table 3 were used.
  • Having the sites for the weather forecasts and the local measured weather coincide means that there is no spatial error between the building and the weather station.

4.2.1. Building modeling

  • Since the dynamic behavior of the building is bilinear between inputs, states and weather parameters, the dynamic equations of the MPC problem result in a non-convex optimization, which can be difficult to solve.
  • The approach that is taken is a form of Sequential Linear Programming (SLP) for solving non-linear problems in which the system is iteratively linearized around the current solution, the optimization problem is solved and it is repeated until a convergence condition is met [29] .

4.2.3. Building model data

  • These parameters can be either determined from the construction plan according to the materials used and their tabular values or, alternatively, the parameters can be determined via estimation methods.
  • For this investigation, only simulation models needed to be constructed, therefore, tabular values for the materials were used.

4.3.1. MPC formulation

  • One needs to choose between the two principle MPC formulations: deterministic MPC and stochastic MPC.
  • This choice determines the formulation of the constraints and the cost function.
  • Hence, in the following these formulations are detailed for both choices, deterministic MPC and stochastic MPC.

• Deterministic MPC (DMPC)

  • The result is a simple linear constraint of the form EQUATION where x min t,1 and x max t,1 are the desired upper and lower comfort bounds, which can be time-varying, e.g., due to night-setbacks for the room temperature.
  • The primary limitation of DMPC is the fact that, if the weather does not actually equal the expected value, then the bound may be violated.
  • This is most often dealt with by artificially tightening the upper and lower bounds, which provides a buffer zone and can be effective for small variances.
  • The cost to be paid is the additional energy required to hold the room temperature further away from the bounds.

• Stochastic MPC with chance constraints

  • One method of automatically determining an appropriate amount to tighten the constraints (shown for the upper bound here) is to formulate them as chance constraints as discussed in Section 3.1.3.2.
  • This is why the procedure discussed in the next section was introduced.
  • This mitigation is, however, bounded by the amount of input energy available to the controller, since the approach requires that some input energy is allocated for compensating for the disturbance (M k ), and some for steering the system to a desired state (h k ).
  • The resulting optimization problem has a second order cone constraint, which is convex and therefore tractable.
  • For this reason, the matrices M k were chosen during a pre-processing step and fixed for the entire year, which turns the constraints into linear constraints.

4.3.4. Soften constraints

  • It is not always possible to satisfy all constraints of the building and so a standard relaxation-procedure, the so-called softconstraints [14, 15] , is required, that chooses automatically which constraints are to be violated first.
  • This is achieved by adding variables to the optimization routine which allows every constraint to be violated.
  • These variables are however heavily penalized, which forces them to zero, i.e. to the satisfaction of all constraints, if at all possible.
  • If this is not possible, then these additional variables give the optimizer sufficient flexibility to always find a solution that can be applied to the building.
  • One can define the relative importance of each constraint by tuning the relevant weighting matrices and thereby have the system violate the least important first [31] .

4.4. Step 4: solve optimization problem

  • Once the constraints and cost have been formulated, the resulting problem can be passed to a standard optimization routine.
  • For this work, the commercial package CPLEX [32] was used, since it is effective for the large-scale and sparse problems that result.

5.1. Controller assessment concept

  • The Non-Renewable Primary Energy (NRPE) usage was assessed as well as the amount and number of constraint violations of the room temperature.
  • In the investigation of the theoretical potential, the variants listed in Table 5 were considered.
  • The combination of all possible variants makes in total 1280 cases.
  • For PB and RBC, no Kalman filter was necessary, since PB has a perfect prediction available and RBC does not have any prediction available.
  • In the following, it is detailed how the six questions Q1 to Q6 were addressed: Q1 Theoretical savings potential For the corresponding analysis PB and RBC were compared for 1280 cases.

Q2 Prediction horizon length

  • For the corresponding analysis PB with a prediction horizon of seven days was compared with PB simulations with shorter prediction horizons for the 1280 cases.
  • The analysis is separated for different HVAC systems and the two building standards.
  • One can see that for most systems (on average) a prediction horizon of 24 h is sufficient to deviate not more than 5% from the PB performance.
  • Lengths that was found for HVAC system S5 can be attributed to the fact that Passive house buildings have much lower heating and cooling energy demand as compared to Swiss average buildings.
  • Accordingly, in the latter case the TABS's thermal storage capacity becomes more important, and because TABS are thermally inert systems their predictive control requires in general much longer prediction horizons.

Q3 Performance of DMPC

  • For the corresponding analysis DMPC was employed using weather forecasts from the COSMO-7 numerical weather prediction model.
  • The performance of DMPC in terms of NRPE use and violations for all cases in Table 6 was compared to PB and RBC and is plotted in Fig. 7 . DMPC and RBC had a larger NRPE use than PB and had both many violations.
  • When comparing RBC and DMPC, both RBC as well as DMPC exceeded 70 Kh/a, but DMPC violated this threshold more clearly.
  • The number of violations were also typically larger for DMPC.
  • Note, for both RBC and DMPC the data points for Cases 4 and 8 are not shown because they were far beyond the axes ranges of the plots.

Q4 Performance of SMPC

  • For the corresponding analysis SMPC was employed using weather forecasts from the COSMO-7 numerical weather prediction model.
  • In Fig. 8 , SMPC is directly compared for the same six cases with the performance of RBC.
  • Figs. 9 and 10 show the resulting room temperature profiles throughout the whole year for Case 3 in Table 6 when using RBC and SMPC respectively.
  • Furthermore, the diurnal temperature variations are much smaller with SMPC, which is a more favorable behavior in terms of comfort.
  • This behavior is observable for the other cases in a similar fashion.

Q5 Importance of Weather Predictions

  • This question was treated by comparing SMPC performance using COSMO-7 weather predictions, i.e. provided by a weather service, versus using 24 h persistence predictions, i.e. continuous recycling of the data from the last 24 h.
  • Again, the same six example cases from Table 6 were analyzed.

Q1 Theoretical savings potential

  • In the investigated 1280 cases, the reasonable amount of violations of 70 Kh/a was exceeded for many cases when using the RBC controller.
  • Fig. 5 shows the joint cumulative distribution of the theoretical energy savings potential (as additional NRPE use in % of PB) and the amount of comfort violations in Kh/a.
  • It can be seen that more than half of the considered cases show an additional NRPE use of more than 40 % (rightmost brown column).
  • Thus, for many cases there is a significant savings potential, which can potentially be exploited by SMPC.

Q5 Importance of weather predictions

  • Fig. 11 depicts the performance of SMPC with persistence predictions (SMPC pers ) versus COSMO-7 weather predictions (SMPC C7 ), (which was used for all investigations above).
  • SMPC pers shows in all cases a clearly higher NRPE use.
  • In two cases each, SMPC pers shows slightly less violations, equal amounts of violations, and clearly larger violations than SMPC C7 .

7. Conclusion

  • In summary, SMPC is a promising approach for building climate control.
  • Its performance in real applications can be expected to vary with the quality of the model and the available input data (model parameters, system states, weather predictions, etc.) to an extent that remains to be investigated.

Did you find this useful? Give us your feedback

Figures (10)
Citations
More filters
Journal Article•DOI•
Theory and applications of HVAC control systems – A review of model predictive control (MPC)

[...]

Abdul Afram1, Farrokh Janabi-Sharifi1•
Ryerson University1
01 Feb 2014-Building and Environment
TL;DR: In this paper, the authors present a literature review of model predictive control (MPC) for HVAC systems, with an emphasis on the theory and applications of MPC for heating, ventilation and air conditioning (HVAC) systems.

899 citations

Cites background from "Use of model predictive control and..."

  • ...The cost function is based on the desired behavior of the system and serves to stabilize the system if the optimal cost can be described by a Lyapunov function [91]....

    [...]

Journal Article•DOI•
Occupant behavior modeling for building performance simulation: Current state and future challenges:

[...]

Da Yan1, William O'Brien2, Tianzhen Hong3, Xiaohang Feng1, H. Burak Gunay2, Farhang Tahmasebi4, Ardeshir Mahdavi4 •
Tsinghua University1, Carleton University2, Lawrence Berkeley National Laboratory3, Vienna University of Technology4
15 Nov 2015-Energy and Buildings
TL;DR: In this paper, the state-of-the-art research, current obstacles and future needs and directions for the following four-step iterative process: (1) occupant monitoring and data collection, (2) model development, (3) model evaluation, and (4) model implementation into building simulation tools.

629 citations

Cites background from "Use of model predictive control and..."

  • ...9 weather should be obtained from an on-site weather station due to spatial differences from local weather stations [79]....

    [...]

Journal Article•DOI•
Benefits and challenges of electrical demand response: A critical review

[...]

Niamh O׳Connell1, Pierre Pinson1, Henrik Madsen1, Mark O׳Malley2•
Technical University of Denmark1, University College Dublin2
01 Nov 2014-Renewable & Sustainable Energy Reviews
TL;DR: A critical review of the literature in the field of demand response can be found in this article, where the authors provide an overview of the benefits and challenges of the demand response resource, including the ability to balance fluctuations in renewable generation and facilitate higher penetrations of renewable resources on the power system.
Abstract: Advances in IT, control and forecasting capabilities have made demand response a viable, and potentially attractive, option to increase power system flexibility. This paper presents a critical review of the literature in the field of demand response, providing an overview of the benefits and challenges of demand response. These benefits include the ability to balance fluctuations in renewable generation and consequently facilitate higher penetrations of renewable resources on the power system, an increase in economic efficiency through the implementation of real-time pricing, and a reduction in generation capacity requirements. Nevertheless, demand response is not without its challenges. The key challenges for demand response centre around establishing reliable control strategies and market frameworks so that the demand response resource can be used optimally. One of the greatest challenges for demand response is the lack of experience, and the consequent need to employ extensive assumptions when modelling and evaluating this resource. This paper concludes with an examination of these assumptions, which range from assuming a fixed linear price–demand relationship for price responsive demand, to modelling the highly diverse, distributed and uncertain demand response resource as a single, centralised negative generator, adopting fixed characteristics and constraints.

475 citations

Journal Article•DOI•
Review of building energy modeling for control and operation

[...]

Xiwang Li1, Jin Wen1•
Drexel University1
01 Sep 2014-Renewable & Sustainable Energy Reviews
TL;DR: In this paper, an up-to-date overview of research on application of building energy modeling methods in optimal control for single building and multiple buildings is also summarized in this paper, and different model-based and model-free optimization methods for building energy system operation are reviewed and compared.
Abstract: Buildings consume about 41.1% of primary energy and 74% of the electricity in the U.S. Better or even optimal building energy control and operation strategies provide great opportunities to reduce building energy consumption. Moreover, it is estimated by the National Energy Technology Laboratory that more than one-fourth of the 713 GW of U.S. electricity demand in 2010 could be dispatchable if only buildings could respond to that dispatch through advanced building energy control and operation strategies and smart grid infrastructure. Energy forecasting models for building energy systems are essential to building energy control and operation. Three general categories of building energy forecasting models have been reported in the literature which include white-box (physics-based), black-box (data-driven), and gray-box (combination of physics based and data-driven) modeling approaches. This paper summarizes the existing efforts in this area as well as other critical areas related to building energy modeling, such as short-term weather forecasting. An up-to-date overview of research on application of building energy modeling methods in optimal control for single building and multiple buildings is also summarized in this paper. Different model-based and model-free optimization methods for building energy system operation are reviewed and compared in this paper. Agent based modeling, as a new modeling strategy, has made a remarkable progress in distributed energy systems control and optimization in the past years. The research literature on application of agent based model in building energy system control and operation is also identified and discussed in this paper.

470 citations

Cites background from "Use of model predictive control and..."

  • ...[38] developed a new MPC strategy for building climate control....

    [...]

Journal Article•DOI•
Importance of occupancy information for building climate control

[...]

Frauke Oldewurtel1, David Sturzenegger1, Manfred Morari1•
ETH Zurich1
01 Jan 2013-Applied Energy
TL;DR: In this paper, the authors investigated the potential of using occupancy information to realize a more energy efficient building climate control, focusing on Swiss office buildings equipped with Integrated Room Automation (IRA), i.e. the integrated control of Heating, Ventilation, Air Conditioning (HVAC) as well as lighting and blind positioning of a building zone or room.

338 citations

References
More filters
Book•
Predictive Control With Constraints

[...]

Jan M. Maciejowski
01 Dec 2001
TL;DR: A standard formulation of Predictive Control is presented, with examples of step response and transfer function formulations, and a case study of robust predictive control in the context of MATLAB.
Abstract: 1. Introduction to Predictive Control. 2. A Standard Formulation of Predictive Control. 3. Solving Predictive Control Problems. 4. Step Response and Transfer Function Formulations. 5. Tuning. 6. Stability. 7. Robust Predictive Control. 8. Perspectives. 9. Case Studies. 10. The Model Predictive Control Toolbox. References Appendices A. Some Commercial MPC Products B. MATLAB Program basicmpc C. The MPC Toolbox D. Solutions to Problems

5,468 citations

Journal Article•DOI•
Model predictive control: theory and practice—a survey

[...]

Carlos E. García1, D. M. Prett2, Manfred Morari1•
Royal Dutch Shell1, California Institute of Technology2
03 May 1989-Automatica
TL;DR: The flexible constraint handling capabilities of MPC are shown to be a significant advantage in the context of the overall operating objectives of the process industries and the 1-, 2-, and ∞-norm formulations of the performance objective are discussed.

5,188 citations

Journal Article•
Optimal Filtering

[...]

Brian D. O. Anderson, John B. Moore, Mansour Eslami1•
State University of New York System1
01 Mar 1982-IEEE Transactions on Systems, Man, and Cybernetics
TL;DR: This book helps to fill the void in the market and does that in a superb manner by covering the standard topics such as Kalman filtering, innovations processes, smoothing, and adaptive and nonlinear estimation.
Abstract: Estimation theory has had a tremendous impact on many problem areas over the past two decades. Beginning with its original use in the aerospace industry, its applications can now be found in many different areas such as control and communjcations, power systems, transportation systems, bioengineering, image processing, etc. Along with linear system theory and optimal control, a course in estimation theorycan be found in the graduate system and control curriculum,of most schools in the country. In fact, it is probably one of the most,salable courses as far as employment is concerned. However, despite its economic value and the amount of activities in the field, very few books on estimation theory have appeared recently. This book helps to fill the void in the market and does that in a superb manner. Although the book is called OptimalFiltering, the coverage is restricted to discrete time filtering. A more appropriate title would thus be Optimal Discrete Time ,Filtering. The authors’ decision to concentrate on discrete time f lters is due to “recent technological developments as well as the easier path offered students and instructors.” This is probably a wise move since a thorough treatment of continuous time filtering will require a better knowledge o f stochastic processes than most graduate students or engineers will have. As it stands now, the text requires little background beyond that of linear system theory and probability theory. Written by active researchers, in the area, the book covers the standard topics such as Kalman filtering, innovations processes, smoothing, and adaptive and nonlinear estimation. Much of the material in the book has been around for a long time and has been widely used, by practitioners in the area: Some results are more recent. However,-it .has been difficult to locate all of them presented in a n organized manner within a single text. This is especially true of the chapters dealing with the computation aspects and nonlinear and adaptive estimation. After a short introductory chapter, Chapter 2 introduces the mathematical model to be used throughout most of the book. The discrete time Kalman filter is 1 hen presented in Chapter 3, along with some applications. Chapter 4 contains a treatment

4,836 citations

Book•
Stochastic programming

[...]

András Prékopa
01 Jan 1995

1,224 citations

Journal Article•DOI•
Meso-gamma scale forecasts using the nonhydrostatic model LM

[...]

J. Steppeler1, G. Doms1, U. Schättler1, H. W. Bitzer1, Almut Gassmann1, U. Damrath1, G. Gregoric2 •
Deutscher Wetterdienst1, University of Ljubljana2
01 Jan 2003-Meteorology and Atmospheric Physics
TL;DR: It is concluded, that orographic filtering is necessary and the effect of such filtering on precipitation forecasts is investigated, and examples of model predictions at the meso-γ scale are given.
Abstract: ¶The nonhydrostatic model LM was developed for small scale operational predictions. Advances in computer development will give the possibility of operational models of a rather fine scale, which will cover the meso-gamma scale. The LM is currently applied at a scale of 7 km and an increase of the operational resolution to 2.5 km is planned for the next few years. Predictions of such high resolution require to abandon the hydrostatic assumption, which is used with most current operational weather prediction models. The LM was designed to cover all resolutions from 50 m to 50 km with an efficiency making it suitable for operational use. It is a fully elastic model, using second order centred finite differences. The time integration is done using the Klemp–Wilhelmson method, treating the slow modes by a larger time step than the fast modes. The vertical propagation of the fast waves is done implicitly. After describing the design of the LM, this paper gives examples of model predictions at the meso-γ scale. Some results of the current operational application at the resolution 7 km are presented. Deficiencies in the localisation of model generated precipitation are investigated using an idealised bell shaped mountain and applying different resolutions. In this way the convergence to the correct solution can be investigated. From these results it is concluded, that orographic filtering is necessary and the effect of such filtering on precipitation forecasts is investigated. Finally, the prediction of a squall line over northern Germany is shown in order to demonstrate the potential of the model in forecasting the meso-γ scale.

829 citations

Related Papers (5)
Experimental analysis of model predictive control for an energy efficient building heating system

[...]

01 Sep 2011-Applied Energy
Jan Široký, Frauke Oldewurtel, Jiří Cigler, Samuel Prívara 
Model Predictive Control for the Operation of Building Cooling Systems

[...]

01 May 2012-IEEE Transactions on Control Systems and Technology
Yudong Ma, Francesco Borrelli, Brandon Hencey, Brian Coffey, Sorin Bengea, Philip Haves 
Theory and applications of HVAC control systems – A review of model predictive control (MPC)

[...]

01 Feb 2014-Building and Environment
Abdul Afram, Farrokh Janabi-Sharifi
A review on buildings energy consumption information

[...]

01 Jan 2008-Energy and Buildings
Luis Pérez-Lombard, José Ortiz, Christine Pout
Energy efficient building climate control using Stochastic Model Predictive Control and weather predictions

[...]

29 Jul 2010
Frauke Oldewurtel, Alessandra Parisio, Colin N. Jones, Manfred Morari, Dimitrios Gyalistras, Markus Gwerder, Vanessa J. Stauch, Beat Lehmann, Katharina Wirth 
Frequently Asked Questions (1)
Q1. What have the authors contributed in "Use of model predictive control and weather forecasts for energy efficient building climate control" ?

This paper presents an investigation of how Model Predictive Control ( MPC ) and weather predictions can increase the energy efficiency in Integrated Room Automation ( IRA ) while respecting occupant comfort. In this paper it is reported on the development and analysis of a Stochastic Model Predictive Control ( SMPC ) strategy for building climate control that takes into account hance-constrained control the uncertainty due to the use of weather predictions. As first step the potential of MPC was assessed by means of a large-scale factorial simulation study that considered different types of buildings and HVAC systems at four representative European sites. The findings suggest that SMPC outperforms current control practice.