Energy retrofit of educational buildings: Transient energy simulations, model calibration and multi-objective optimization towards nearly zero-energy performance

Energy and Buildings 144 (2017) 303–319

Contents lists available at ScienceDirect

Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild

Energy retrofit of educational buildings: Transient energy simulations, model calibration and multi-objective optimization towards nearly zero-energy performance Fabrizio Ascione a,∗ , Nicola Bianco a , Rosa Francesca De Masi b , Gerardo Maria Mauro a , Giuseppe Peter Vanoli c a

Università degli Studi di Napoli Federico II, DII – Department of Industrial Engineering, Piazzale Tecchio, 80, 80125 Napoli, Italy Università degli Studi del Sannio, DING – Department of Engineering, Piazza Roma, 21, 82100 Benevento, Italy c Università degli Studi del Molise, Dipartimento di Medicina e Scienze della Salute “Vincenzo Tiberio”, Via Francesco De Sanctis, 1, 86100 Campobasso, Italy b

a r t i c l e

i n f o

Article history: Received 9 November 2016 Received in revised form 16 February 2017 Accepted 24 March 2017 Available online 27 March 2017 Keywords: Energy modelling and model calibration Educational buildings Building energy retrofit Multi-objective optimization Cost-optimal analysis Genetic algorithm

a b s t r a c t Building activity is the sector that affects for the most part the anthropogenic climate change. Indeed, even if differences can be found among countries, according to the level of development and climates, buildings require about 30–40% of the overall energy demand, with similar share concerning the greenhouse emissions. According to the more recent EU Directives in matter of energy efficiency in energy use, a great attention has to be paid to energy refurbishments of existing buildings. Indeed, the turnover rate of the EU building stock is generally low in fully-developed countries, so that the energy retrofits are also more important compared to new nearly zero-energy buildings. The proposed investigation concerns the demonstrative role of the public hand, whose necessity is underlined by the EU Directives 2010/31/EU and 2012/27/EU, through a multi-step and multi-objective optimization of an educational building of an Italian University. All preliminary investigations aimed at a reliable modelling, the iterative method that combines genetic algorithms and transient energy simulations tailored on calibrated numerical models make the investigation complete and repeatable. All levers of energy efficiency have been pressed, and thus the refurbishment of building envelope, HVAC systems, integration of energy supply by renewables. The multi objective optimisation concerns costs, incentives, indoor comfort, energy demands for heating and cooling and a novel approach is proposed for choosing the best configuration of retrofit. It is resulted that the most profitable energy efficiency measures involve the modernization of energy systems, even if also the retrofit of the building envelope can be profitable under favourable financial conditions. The cost-effective refurbishment reduces the primary energy demand up to a value of 12 kWh/m2 a, so that the building can be surely considered as nZEB. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Space heating and cooling are the prevalent energy usages of European Union (EU) [1], with 50% (546 Mtoe) of final energy consumption in 2012, and it is expected that these will remain quite high. Thus, energy-efficient and low/zero-carbon energy technologies for heating and cooling in buildings will play a key role to achieve a significant reduction of greenhouse gas emissions at European level, as well as to meet the targets of the COP21 climate

∗ Corresponding author. E-mail addresses: [email protected] (F. Ascione), [email protected] (N. Bianco), [email protected] (R.F. De Masi), [email protected] (G.M. Mauro), [email protected] (G.P. Vanoli). http://dx.doi.org/10.1016/j.enbuild.2017.03.056 0378-7788/© 2017 Elsevier B.V. All rights reserved.

conference 2015 of Paris. In this frame, the Energy Performance of Buildings Directive (EPBD) Recast (2010/31/EU) [2] showed a framework for improving energy performance of European buildings and for increasing the exploitation of renewable sources. A key element of the EPBD Recast, in order to achieve long-term efficiency objectives, is the introduction of the standard of ‘nearly zero-energy building (nZEB)’. According to the mentioned Directive, a nZEB is “a building that has a very high energy performance, for which the nearly zero or very low amount of energy required should be covered to a very significant extent by energy from renewable sources, including energy from renewable sources produced on-site or nearby”. More in detail, the European Directive 2010/31/EU [2], established that the definition of nZEB is a task that should be defined

304

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

Nomenclature Symbols a cov dGC dPEC e gmax r s t x A COP DH EER F GC IC PEC P SHGC Tvis TED U ␩

Absorptance to solar radiation [–] Percentage of roof area covered by PV panels [–] Difference in GC compare to BB [D Difference in PEC compare to BB [Whp /m2 a] [–] Thermal (infrared) emissivity [–] Maximum number of GA generations [–] discount rate [%] Population size of the GA [–] Thickness of thermal insulation layer [m] Vector of design variables [–] Surface of envelope compenents [m2 ] Coefficient of performance of heat pumps [Wt /Wel ] Percentage of discomfort hours [%] Energy efficiency ratio of chillers [Wt /Wel ] Objective functions [–] Global cost [D ] Investment cost [D ] Primary energy consumption per unit of conditioned area and year [Whp /m2 a] Power [W] Solar heat gain coefficient [–] Transmittance to visible radiation [–] Thermal energy demand for space conditioning [Wh/m2 a] Thermal transmittance [W/m2 K] Efficiency of boilers [Wt /Wp ]

Subscripts or Superscripts Referred to the base building configuration BB C Referred to space cooling el Referred to electrical energy or power Referred to the external floor f H Referred to space heating p Referred to primary energy or power Referred to the roof r t Referred to thermal energy or power v Referred to the external vertical walls w Referred to the windows Acronyms BB Base building configuration Domestic hot water DHW ERM Energy retrofit measure Genetic algorithm GA HVAC Heating, ventilating and air conditioning PV Photovoltaic RES Renewable energy source

at national level. Today, where a numerical indicator is set, the requirements – in terms of primary energy – range rather widely from 0 to 270 kWh/m2 a. For residential buildings, the higher energy demand ranges between 33 kWh/m2 a in Croatia and 95 kWh/m2 a in Latvia, with a majority of countries aiming at 45–50 kWh/m2 a. Few Member States mentioned objectives that go beyond nZEB requirements, and thus the targets of net zero-energy buildings (ZEB) in the Netherlands, positive-energy buildings in Denmark and France, climate neutral new buildings in Germany and the zero-carbon standard in the UK [3]. In Italy, the definition of nZEB standard has been completed by the Ministerial Decree 06/26/2015 [4]. The Italian procedure consists of determining the building energy class, assuming conventional boundary conditions (i.e.,

“asset rating”). The energy class is derived from the value of the overall energy performance index EPgl,NREN which represents the annual amount of non-renewable primary energy needed to meet the different needs associated with a standardized use of the building, divided by the useful area of building [kW/m2 ], and thus heating, cooling, sanitary hot water, illumination, ventilation, and electrical consumption for people moving inside the building. In detail, building under examination must achieve energy class IV, the best one, to be classified as NZEB. This means that EPgl,NREN should be lower than 40% (at least) compared with EPgl,NREN index calculated for reference building. This is defined as the one characterized by same geometry, orientation, functionality, climatic conditions and geographic location of the building under investigation, and it respects for the building envelope and the systems, the minimum requirements defined by [4] from 1 January 2019 for public buildings, and from 1 January 2021 for all others. However, new buildings are few in all EU developed countries (a reasonable range is 0.4–1.2% per year) and thus the major way for achieving de-carbonization targets is the renovation of the existing building stock. About it, two thirds of the EU’s buildings were built when mandatory energy efficiency requirements were poor or absent. For instance, almost half of buildings have individual boilers installed before 1992, with efficiency equal to 0.60 or less, and, moreover, 22% of individual gas boilers, 34% of direct electric heaters, 47% of oil boilers and 58% of coal boilers are older than their technical lifetime. Thus, large savings can be achieved through simple retrofits of building envelope and HVAC systems. For what concerns the Italian building stock, Corrado et al. [5] elaborated data of the national pilot action for the IEE-EPISCOPE project, concerning the residential buildings of Piedmont region (North Italy). The analyses underlined that the current trend of the average yearly refurbishment rate, included in the range 0.06–1.05%, is not sufficient to achieve the CO2-eq emissions’ reduction target. In Europe, offices and educational buildings account for about 40% of the entire non-residential floor space, such as specified by the first version of the EPBD. In the non-residential sector, since 1990, electricity consumption has increased by a remarkable 74%. Nonresidential average specific energy consumption was estimated around 280 kWhp/m2 a (with reference to all end-uses) [6]. More in detail, educational buildings and sport facilities account for a further 18% of the energy use [6]. In China, surveys on energy consumption of colleges and universities in Guangdong province [7] showed that campuses are among the major energy consumers. Chung and Rhee [8] established that potential energy savings, in the range 6–29%, can be reached in university buildings of Seoul, South Korea. As underlined by Santamouris [9], high-energy consumption of the building sector, local climate change and energy poverty are strongly interrelated. According to the author, the innovation of the built environment of Europe can contribute to the minimization of consumption of the building sector, eradication of energy poverty and mitigation of urban heat island effect and local climate change. The study here presented concerns the energy retrofit of an educational (i.e., university) building located in South Italy, by showing that the application of a rigorous optimization methodology provides robust cost-optimal retrofit solutions that ensure very high levels of energy performance, near to those typical of nZEBs. Thus, this study aims to demonstrate that the cost-optimal building retrofit towards nZEBs is possible, thereby coupling the drastic reduction of energy consumption, and thus of environmental impact, with cost-effectiveness and economic feasibility. The methodology is applied after the calibration of the building energy model by means of energy audit and measured data. Before the description of methodology and outcomes, the following lines propose a brief review of research in matter of building energy retrofitting aimed at the achievement of the nZEB goal. Since the proper selection of the best packages of energy retrofit measures

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

305

Scheme 1. Sequence of the optimization procedure by means of the applied Genetic Algorithm, taken from [36].

(ERMs) requires to solve multi-objective optimization problems (with several potential objective functions, design variables and constraints), the main methodological approaches for the conversion of existing buildings into nZEBs are discussed.

1.1. Energy retrofit for the goal of nearly zero-energy buildings (nZEBs) The zero-energy concept for existing buildings involves two main design strategies: I) minimizing building energy needs by means of the application of efficient and effective ERMs, II) exploiting renewable energy sources (RESs). Li et al. [10] reviewed the most worthwhile scientific studies related to these two strategies, while Kapsalaki and Leal [11] reviewed definitions, methodologies, technologies and cases of residential and non-residential nZEBs. In the same vein, several papers investigated the potential conversion of buildings with poor energy performance into nearly zero systems, by discussing the technical and economic feasibility of different case studies. The results of Serghides et al. [12] indicated that the refurbishment of an old single-family house, in Cyprus, into a nZEB is financially viable, with a payback period of 8 years. Ferreira et al. [13] concluded that the renovation of the Portuguese existing residential building stock can be made with packages of measures that lead to the cost-optimal point with very low energy consumption (0–15 kWh/m2 a) and low carbon emissions (0–2.16 kgCO2-eq /m2 a). Ciulla et al. [14] investigated the most common retrofit solutions used in Italy, by focusing mainly on the energy performance of historical building envelopes, and showed potential primary energy savings in the range 44.6–56.7%. For ex-Soviet Union countries, Kuusk et al. [15] showed that, through a deep renovation, nearly zero-energy apartments can be achieved, since space heating and electricity demands can be reduced of around 70%, with payback periods, generally, around 30 years or 8 years in the best financial scenario. Also, Passer et al. [16] investigated a case study in Austria and the results indicated that a high-quality refurbishment of the thermal envelope with prefabricated fac¸ade elements, inclusive of solar thermal collectors and PV panels, is the optimal energy retrofit. Patterson et al. [17] explored the potential of curtainwall technologies for achieving net zero-energy in existing buildings. Dabaieh et al. [18] demonstrated that vernacular structures in the Western Desert of Egypt can become nZEBs when retrofitted and equipped with renewable energy solutions. Moreover, Alirezaei et al. [19] investigated the role of the Vehicle to Home technology in satisfying the energy requirements for a nZEB. The results indicated that grid electricity consumption can still be effectively reduced by up to 68% compared to a conventional building design. About the non-residential sector, Aksamija [20] discussed feasibility of achieving net zero-energy goals with appropriate design manipulations and use of multiple renewable energy sources in retrofitting commercial buildings, by considering a case study in Massachusetts (U.S.A). Concerning hot

climates, AlAjmi et al. [21] investigated three scenarios to convert a public building from inefficient energy consumer into a nZEB. The best scenario, designed to operate in synergy with the local energy grid, provides annual avoided CO2-eq emissions of about 748 tons and payback period of 11.8 years. For what concerns the educational building sector, Niemala et al. [22] evaluated the impact of different ERMs for typical university buildings built in 1960s and 1970s in Finland. The results indicated that the proposed national nZEB target can be cost-effectively achieved in renovations. For the historical building of the School of Engineering and Architecture in Bologna (North Italy), Semprini et al. [23] showed that heating energy consumption can be reduced from 15%, through the optimization of building management, to 26%, through the energy retrofit of heating system and windows. Tahsildoost et al. [24] presented the retrofit of two typical schools in Iran, in which airtightness, window replacement and roof insulation were designed. After the refurbishment, energy consumption and thermal environment of the schools were monitored, by showing 29.87% and 38.29% primary energy savings, respectively, and the improvement of indoor environmental also testified by the occupants. Concerning Italian buildings, Ascione et al. [25] investigated an educational edifice in the ancient centre of Benevento (South Italy, the same area here analysed), by showing that the proper management of occupant behaviour, indoor controls and comfort ranges provide significant energy savings with zero or low-capital investments. Moreover, Dalla Mora et al. [26] presented a radical refurbishment of an historical building, Ca’ S. Orsola, which is certified by the seismic point of view and each apartment of the block is certified in energy class A. Really, properly-designed ERMs decreased global energy consumption up to 90%. About it, Becchio et al. [27] underlined that the success of a nZEB depends not only on how the building is designed and built, but, definitely, also on the occupant behaviour. Indeed, the interaction between occupants and building systems has an impact on actual energy consumption up to 36%.

1.2. Methodological approach for the design of high energy performance buildings and nZEBs Several studies concern the application of the cost-optimal methodology, established by the EPBD Recast [2,28], especially to design nearly zero-energy buildings. Becchio et al. [29], for a single-family house in North Italy, compared different solutions of HVAC systems according to the macro-economic approach of the EPBD guidelines. They concluded that nZEBs are technically feasible but these are still quite far to be the best solution in terms of cost-optimality. Kurnitski et al. [30] gave the same conclusion, with reference to the Estonian climate. According to these studies, nearly zero-energy performance level is not yet the cost-optimal one, by taking into account current prices of technologies and thus it is suggested the target of technically-reasonably achievable levels of energy efficiency, with current best-practices

306

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

and on site energy conversion from renewables. Ferrara et al. [31] applied the cost-optimal methodology to a low-consumption French single-family house, in order to study how the primary energy system influences the envelope design of a cost- optimal nZEB. Congedo et al. [32] presented the application of cost-optimal methodology for an office building in a warm climate and they obtained that the best configuration has a global cost lower of 62.4 D /m2 compared to the reference building. Moreover, Lizana et al. [33] presented a multi-criteria assessment method within the ‘(Re)Programa’ research project for the South-West of Europe. During the energy retrofitting, this methodology allows to consider the requirements of each intervening agent, through combined economic, environmental and social evaluations. Briefly, the design of high-performance buildings, up to zero-energy buildings, is a multivariable problem, and it includes essential requirements such as energy performance, visual and thermal comfort. However, as discussed by Sesana and Salvalai [34], the nZEB design would certainly benefit from the adoption of life cycle methodologies, as done in this study. Carlucci et al. [35] provided the design of a detached net zero-energy house in South Italy, by using a non-dominated sorting genetic algorithm, implemented in the GenOpt optimization engine, through the Java genetic algorithms’ package, to instruct the EnergyPlus [36] simulation code. In the same vein, a multi-objective optimization was carried out using a genetic algorithm (NSGA-II) coupled with a dynamic simulation tool by Penna et al. [37]. They investigated the relationship between the initial characteristics of residential buildings and the definition of optimal retrofit solutions in terms of either maximum economic performance or energy consumption minimization towards nZEB behaviour for the lowest achievable thermal discomfort. Hamdy et al. [38] compared the performances of seven commonly-used multi-objective optimization algorithms in solving the design problem of a nearly zero-energy building, by finding that 1400–1800 were the minimum required number of evaluations to stabilize optimization results of the building energy model. Ascione et al. [39] proposed a new methodology ® by means of the coupling between MATLAB [40] and EnergyPlus, by implementing a genetic algorithm, and this allows the evaluation of profitable and feasible packages of energy efficiency measures applied to buildings. Analogously, Hamdy et al. [41] proposed a multi-stage optimization method to find the cost-optimal and nZEB energy performance levels for a single-family house in Finland. 2. Aim of paper and methodology The proper selection of energy retrofit measures (ERMs) for existing buildings, even more than the energy design of new constructions, requires the handling of complex optimization problems with many objective functions and several contrasting targets. More in detail, given the target of net and nearly zero-energy buildings (nZEBs), according to the new EU indications stated by the European guidelines [2,28], several objectives have to be taken into account simultaneously, among which: • • • •

reduction of energy demand for the space heating and cooling, savings for what concerns the indoor artificial lighting, thermal, visual and environmental comfort, reduction of operating costs, investment costs and thus global costs.

Sometimes, these targets are in mutual contradiction. For instance, a reduction of power provided for artificial lighting is, obviously, beneficial in terms of electric energy demand, but it affects negatively the heating need for the microclimatic control

(because of the lack of a thermal gain). Moreover, a smart control (e.g., automatic model predictive regulation) for the heating system can affect the indoor comfort and, moreover, it rises up the investment cost. Many other examples could be provided, such as the indoor summer overheating caused by high thicknesses of thermal insulation of the opaque envelope or due to the adoption of lowemissive glasses. Finally, with reference to both thermodynamics (i.e., thermal comfort, energy demands for heating and cooling) and costs (investments and operating costs), the building energy design requires deep and careful investigations, so that suitable and tailored energy simulations, by taking into account several objectives, are necessary. As discussed in the previous section, optimization approaches, more or less advanced, are suitable for finding the best trade-off among all possible adoptions (single ERMs) and combinations (packages) of solutions. In this study, a multi-objective and multi-stage optimization approach is applied in order to find costoptimal retrofit solutions towards nearly zero-energy performance for the investigated educational building. In detail, once fixed the main boundary conditions, concerning geometry, climatic location, profiles of occupancy and patterns of building use, as well as minimum satisfactory conditions of thermal comfort, the most suitable trade-off among several ERMs is investigated. The considered energy retrofit measures concern all levers of energy efficiency in buildings, such as evidenced in [42], and thus: • the building envelope (i.e., particular building plasters and coatings, addition of thermal insulation, thermal mass, new kind of low-emissive of selective windows); • the primary energy systems (i.e., adoption of efficient air-source heat pump, ground source heat pump, water-cooled chiller with cooling tower); • electric on-site renewables (i.e., photovoltaic generators). All energy conservation measures, with detailed description of the investigated options (for instance, thicknesses of thermal insulation layers or overall heat transfer coefficients of windows, energy efficiency ratios of heat pumps) are discussed in the following subsections. In order to address building retrofit towards the nZEB target, the proposed methodology integrates reliable simulations to predict building energy performance with robust optimization algorithms. In detail, the previously cited EnergyPlus [36] is used for energy simulations because it ensures high accuracy and levels ® of detail in building modelling. On the other hand, MATLAB [40] is used to run the optimization algorithms and to perform dataprocessing because of its large opportunities and capabilities of programming. Definitely, the selection of these tools is not binding, since the paper is not focused on integrating software but proposes integrated methods, which can be replicated also in other software in order to support the application of the methodology. It is noticed that the same authors applied a similar procedure to optimize the energy retrofit of residential [39,43] and hospital buildings [44] by finding cost-optimal solutions also in presence of different constraints concerning the available economic budget. More in detail, the employed methodology is similar to the one developed in [44] but, compared to this latter, it takes account of indoor thermal comfort and allows to pursue the nZEB target. Thus, compared to current scientific literature about building performance optimization (see the review of Section 1), the paper proposes an original multi-stage and multi-objective framework that provides a comprehensive optimization of building energy retrofit. In this regard, the methodology allows to achieve nZEB performance by also considering the other objective functions to which building stakeholders are generally interested, namely thermal comfort, investment costs, global costs. Notably, a significant weight is given to the minimization of global costs (i.e., costoptimality), because these take account of both operation costs,

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

linked to energy consumption and thus polluting emissions, and investment costs, thereby providing a good indicator to represent the interests of all building stakeholders, in particular private and public perspectives. The multi-stage framework permits to investigate a huge scenario of retrofit scenarios in order to find a constrained cost-optimal solution, which ensures a substantial reduction of energy needs in order to reach the nZEB target, and complies with occupants’ requirements about thermal comfort. It is highlighted that the current body of knowledge includes several methods to optimize building energy retrofit (see Section 1), but a comprehensive framework, as the one here proposed, that allows to achieve the nZEB target in existing buildings, by considering all main objectives of building stakeholders and a huge domain of retrofit solutions, is a novelty. In particular, the methodology performs a multi-stage and multi-objective optimization by implementing a Pareto genetic algorithm, i.e., GA, (1st stage) and a smart exhaustive sampling (2nd stage). The GA, born as a modification of NSGA II [39,45], operates by continuatively and iteratively improving the models of the building in order to identify the non-dominated solutions (i.e., the front of Pareto) for what concerns the design of building energy retrofit by minimizing two or more objective functions. Different conflicting functions can be chosen, as shown in the example of Fig. 1. The method is powerful but, just for these reasons, a deep knowledge and professional expertise in matter of selection of decision variables and their ranges of variability are needed. For instance, by citing the same example mentioned in the previous lines, regarding the building envelope, merely commercial (i.e., discrete) thicknesses of thermal insulation can be chosen. All told, the application of the method is quite complex, being necessary mathematical knowledge and skills, without compromising the reliability of the application in the building sector (professional approach) and building energy efficiency more specifically. As aforementioned, the proposed methodology is articulated in two stages, which consider different objective functions, in order to ensure the reliable investigation of a wide solutions domain, and thus a robust assessment of cost-optimal retrofit solutions, in a reasonable computational time, as thoroughly argued in [44]. Furthermore, the multi-stage development allows to find ‘constrained’ cost-optimal retrofit packages that imply a drastic reduction of building energy consumption, since energy demand is minimized in the 1 st stage in order to make possible the nZEB goal. The following subsections provide the two methodology stages. 2.1. 1st methodology stage: implementation of the three-objective genetic algorithm (GA) In this stage, the mentioned GA is implemented, by using a three-objective approach, to select optimal packages of ERMs by minimizing thermal energy demand for space heating (TEDH ), space cooling (TEDC ) and the annual percentage of discomfort hours (DH). DH is assessed according to the procedure described in [39]. Thus, for what concerns the aforementioned macro-categories of ERMs (i.e., thermal envelope, HVAC systems and equipment, renewable energy sources), the measures, explored in this stage, are addressed to the energy optimization of building envelope and operation of space conditioning systems. Each selected ERM, characterized by a range of values, defines a design decision. The GA carries out a smart research within the solution domain by investigating a limited number of solutions, i.e., retrofit scenarios, properly selected by the optimization logic. Hence, a large amount of computational time is saved, compared to exhaustive researches. The algorithm provides one three-dimensional (3-D) and three bidimensional (2D) Pareto fronts (one for each couple of objective functions), collecting non-dominated solutions, which represent optimal packages of the investigated ERMs. In order to optimize

307

building energy performance towards the nZEB goal, the thermal energy needs for space heating (TEDH ) and cooling (TEDC ) have been set as the first two objective functions. These two components of energy consumption are chosen, instead of others (e.g., energy demand for direct electric uses or domestic hot water) because most measures commonly used to address building energy retrofit affect these thermal demands, often with divergent effects. For instance, building envelope thermal insulation has, definitely, a positive effect on TEDH , but it can have a negative effect on TEDC because of the risk of indoor overheating. Furthermore, the reliable assessment of these two energy indicators requires transient and, generally, time-consuming simulations through proper software, such as EnergyPlus. Thus, the use of the GA is particular effective when these objective functions are chosen, since it produces substantial savings of computational times compared to exhaustive researches. Besides the optimization of energy performance, a proper retrofit strategy should consider also the occupants’ thermal comfort [28], and thus DH is chosen as third objective function. It is clear that the minimization of energy requests for space conditioning and of discomfort hours are mutually conflicting. Indeed, the reduction of TEDH and TEDC can lead to a worsening of indoor conditions (for instance, because a ‘reduced-in time’ use of systems and equipment for microclimatic control is preferred, or the identified set-points and set-backs of air conditioning are the ones that minimize the energy demands). The TED minimization can be the target preferred by the building owner. Conversely, the minimization of low-comfort hours can be the one preferred by the occupants. Indeed, their expectation after retrofit is, commonly, the improvement, as much as possible, of the thermal comfort, in terms of temperatures, air changes, air quality, and also of low time required for the achievement of suitable indoor conditions. About this last point, high winter set-backs of temperature can allow, during the intermittent use of building, that the desired air temperature is achieved in a short time (this implies high set-backs, near to set-points). Clearly, the ‘utopic’ (because not obtainable) goal is the minimization of all targets at the same time. The GA provides the trade-off, i.e., non-dominated, solutions collected in the aforemen® tioned Pareto fronts. This algorithm has been written in MATLAB environment according to the scheme reported in the next lines (Scheme 1), already used by the same authors in [39,43,44], where the vector (F = [TEDH , TEDC , DH]) collects the objective functions, whereas the vector x is composed of bits, which encode the design variables that represent ERMs. Thus, each design variable can assume a limit number of discrete options. This allows to cut the solution domain and is closer to reality and commercial world [39]. However, as aforementioned, the possible discrete options must be carefully set according to experience, expertise and bestpractices. The GA performs, iteratively, an evolution of a population of s (population size) individuals, the so-called chromosomes, each one characterized by a set of values of the aforementioned vector x, whose components are called genes, thus corresponding to a building retrofit scenario. The process is performed through a series of iterations, called generations. It is required to improve the population by means of the selection of the best chromosomes as well as through the mutation and crossover (i.e., combination) of their genes (e.g., the bits encoding the thicknesses of thermal insulation layers) in order to have new individuals that improve building energy and thermal performance. The best chromosomes, the so-called parents, are chosen based on the values of objective functions (i.e., the rank value) and on the average crowding distance among population’s individuals. The best parents form the so-called elite and survive to the generation. After the random creation of the initial population, the described ‘Darwinian evolution’

308

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

Fig. 1. Examples of Pareto multi-objective optimization of building energy retrofit: two (A-graph) and three (B-graph) conflicting objectives.

occurs during each generation, and the procedure ends when one of the below-reported conditions is satisfied: • a threshold number of generations (gmax ) is reached, • the Pareto Front does not change significantly between two following generations. Of course, a not significance variation of the Pareto Front means that a fixed tolerance (tol), for what concerns the spread of change, has to be defined previously. For a deeper description of the GA, the reader is invited to refer to [44], in which the authors applied the same algorithm to a similar case study. Clearly, for each retrofit scenario, encoded by certain values of ® the vector x, MATLAB interrogates EnergyPlus in order to run a transient energy simulation. The simulation’s outcomes are, then, prost-processed to assess the values of TEDH , TEDC and DH for each explored scenario. The coupling scheme between the two programs is shown in Fig. 2, where the flowchart evidences the logical sequence of the procedure. The so-called ‘coupling func® tion EnergyPlus ↔ MATLAB ’ has been written for making possible the communication between the programs. More in detail, that function operates by converting x in a new building model to be simulated (the .idf files) and by handling, suitably, the output files of EnergyPlus (the .csv files) to evaluate the objectives F. The selection of a genetic algorithm was particularly proper, because it is heuristic and iterative, so that suitable for the communication between ® the software (i.e., according to MATLAB , EnergyPlus is a mere generator of hidden functions). It should be noted that the ERMs are implemented and parameterized directly into the.idf of EnergyPlus. Finally, all boundary conditions and simulation criteria have been defined as accurately as possible, according to previous experiences [39,43,44]. Please note that also a constraint is defined, since all solutions that cause an increase of DH compared to the base building configuration are excluded. Indeed, a proper and effective retrofit strategy must not produce a worsening of indoor thermal comfort conditions. As aforementioned, the three-objective GA provides a set of nondominated packages of ERMs for TED and/or DH reduction (Pareto fronts). Thus, the algorithm implementation must be followed by the decision-making process (see Fig. 3), which aims at selecting one optimal solution from the Pareto fronts. This process is performed in the second methodology stage. 2.2. 2st methodology stage: smart exhaustive sampling and Cost-optimal analysis In this stage, the decision-making process is carried out, in order to select one package of ERMs among the non-dominated building configurations. It is a delicate and active task for stakeholders, since

it can be conducted according to different criteria. Indeed, no one of the Pareto solutions can be defined, ‘a priori’, better than another, so that a criterion of selection is required. Commonly, one can select the solution with the shortest geometrical distance from the ideal point that minimizes both objectives (i.e., the ‘utopia point’) or the one that minimizes the energy demand by fixing a value of thermal discomfort. In previous studies ([39,43]), some of the same authors applied both the mentioned approaches, called respectively the ‘utopia point’ and the ‘minimum comfort’ criteria. Conversely, in this study, a third criterion of choice is applied. Indeed, the costoptimal analysis is applied for conducting the decision-making after a smart exhaustive sampling, which allows to consider further ERMs addressed to energy systems and renewable energy sources’ (RESs) exploitation. Finally, the selected retrofit package, among all investigated ERMs (in both methodology stages), is the cost-optimal solution, which is the one that minimizes global costs over building lifecycle. Global costs are assessed according to the guidelines of the EPBD Recast [2,28] and take account of initial expenditure and running costs with reference to a certain number of years, properly actualized at the starting time. In particular, in this stage, a smart exhaustive sampling is implemented by investigating the energy performance of different technical solutions for primary energy systems, in presence of the non-dominated ERMs’ packages for TED and/or DH reduction, selected in the first stage, and also in absence of ERMs for building envelope and operation (BB: base building configuration). For each retrofit scenario, investment cost (IC), primary energy consumption (PEC) and global cost (GC) are assessed, and, finally, the cost-optimal solution is identified in correspondence of different values of the discount rate (i.e., r), thereby performing a sensitivity analysis at the variation of r, as recommended by the guidelines of ® the EPBD Recast [28]. This stage is carried out in MATLAB environment by employing the procedure described in [44], without needing further EnergyPlus simulations, and thus it needs a negligible computational time compared to the first stage. The exhaustive sampling is defined ‘smart’ because: • it explores, besides the BB, only the packages of ERMs, for the reduction of TED and/or DH, that are properly selected through the GA implementation; • it is performed in MATLAB® environment, without needing further EnergyPlus simulations, and thus it needs a very low computational time (i.e., order of magnitude of few minutes). More properly, the chosen retrofit package represents a ‘constrained’ cost-optimal solution because only proper packages of ERMs are selected for the cost-optimal analysis based on the outcomes of the 1st methodology stage. This allows to achieve a cost-optimal solution that produces, at the same time, a substantial

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

309

Fig. 2. Framework of the proposed methodology.

®

Fig. 3. Multi-objective optimization. Iterative coupling of MATLAB and EnergyPlus and decision-making process.

improvement of building energy performance thereby making possible the cost-effective design towards nZEBs. Clearly, the achieved solution could not ensure the achievement of the nZEB goal. If this latter represents a must, a constraint can be set on PEC, i.e., PEC ≤ PECnZEB, in order to obtain a “more constrained” cost-optimal solution, which definitely provides nZEB performance. 3. Case study: energy audit and calibration of the building numerical model The investigated building, briefly named ‘SEA’, hosts offices and classrooms of the Department of Law, Economics, Management and Quantitative Methods of University of Sannio (see Fig. 4). Built in the early ‘90 s and previously used as post office, it is located just outside the ancient centre of the Italian city of Benevento (South Italy, Italian climatic zone C). All information necessary for the building performance assessment have been obtained through in-situ surveys, interviews with managers and occupants, in-field measurements. In order to simulate the real energy performance, hourly energy simulations have been carried out according to the procedure of ‘tailored ratings’ as defined by the international standard EN 15603 [46]. The simulations have been performed through the use

of EnergyPlus software. In this section, the audit procedure, aimed at building numerical modelling as well as at model calibration, is described. 3.1. Building energy audit The building has rectangular shapes, being placed on a sloped area of around 7000 m2 , and the articulation of volumes provides a large amount of roof, sun- and wind- exposed. The global elevation is about 17 m, with 4 usable floors above the ground. The ‘surface to volume ratio’ (S/V) is equal to 0.41 m−1 . The main geometrical characteristics are summarized in Table 1. 3.1.1. Characterization of building uses and definition of thermal zones Direct surveys have permitted the census of all major energy uses. At the basement, the main rooms are used as classrooms (Fig. 5a) and there is a bar/lunch area (Fig. 5c); the ground and second floors mainly host classrooms, libraries and offices of professors. At the first floor, there are mostly offices, as well as a conference room (Fig. 5b) and some laboratories. By combining information about occupancy schedules and installed equipment,

310

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

Fig. 4. Investigated building: A) overall building layout; B) main fac¸ade.

Table 1 Building information and geometrical characteristics. Building net floor Area

6459 m2

Gross Volume

21216 m3

Height

≈17 m

Surface to volume ratio

0.41 m−1

Geometry

Total

North(315◦ –45◦ )

East(45◦ –135◦ )

South(135◦ –225◦ )

West(225◦ –315◦ )

Gross Wall Area [m2 ] Window Opening Area [m2 ] Window-Wall Ratio [%]

3584.08 982.87 27.42

1135.70 322.37 28.39

656.44 133.42 20.32

1135.52 335.84 29.58

656.41 191.24 29.13

Fig. 5. Thermal zones: A) classroom; B) bar; C) conference room; D) geometrical model of 1st floor.

the numerical model has been defined. Fig. 5d shows a typological floor modelled with DesignBuilder [47], one of the most authoritative graphical interface of EnergyPlus. The average air change rate has been defined around 1.5 ACH, for allowing the required comfort conditions [48], even it was detailed according to the specific use of rooms. Moreover, an additional air change rate equal to 0.5 ACH has been considered too, because of the lack of airtightness. A fluorescent lighting system is installed in the whole building. 3.1.2. Building envelope audit Several surveys concerned the study of the building envelope, mainly supported by the coupling of infrared thermography (through of an infrared camera FLIR p660 with nominal accuracy of ±2%) with measurements of heat flux through walls. With reference to the proposed study, the infrared thermography study was carried out in the morning, during the running of the heating system (Fig. 6a), for having optimal monitoring conditions. Starting from the outdoor thermography, the windows have shown – obviously – the highest surface temperature compared to the other structural components. The building has double-glazing windows, with air filling, aluminium frame and external shadings. The estimated overall transmittance UW is 2.70 W/m2 K−1 . The infrared inspection around the window’s frame evidenced also air leakages and infiltrations.

By means of the envelope audit, it was known that the building has a reinforced concrete structure with insulated external walls. Thermographic images have allowed to support in-situ measurements of thermal transmittances by means of heat flux meters. More in detail, a wireless heat flux meter has been used with accuracy of ±5%. The monitoring has been carried out during 17–21 March 2016 (Fig. 6b). The selected sampling time-step was 1800s and the Uvalue resulted from the average calculation of 189 measures, according to the methodology of the “Average Method ISO 9869” [49]. During the in-field tests, the indoor spaces were conditioned also during the night, and an average temperature difference of 6 ◦ C has been registered between indoor and external environment. After a deviation characterizing the first day, assuming conventional internal and external heat transfer coefficients for convection and radiation, a convergence around the value of 0.55 W/m2 K−1 was obtained. The measured Uvalue has been compared to those derived from the analytical calculations, according to the standard ISO 6946 [50]. The average thickness is 0.25 m and the calculation provided a thermal transmittance value equal to 0.53 W/m2 K−1 . This means that good convergence was found. Considering the technical literature experiences, the values of the measured thermal transmittance can significantly diverge from those derived by the analytical calculation, also around ±20%, because of execution not

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

311

Fig. 6. A) outdoor thermographs of the southern fac¸ade; B) in situ measurement of thermal conductance.

fully in compliance with the design, material degradation and environmental conditions different compared to the designed ones. With reference to the floor on the ground and the building roof, experiments have been performed by the Laboratory of Building Construction of the Department of Engineering of the University of Benevento. More in detail, ceiling, slab on the ground and roofs have mixed structures, given by the parallel presence of concrete beams, joists and interposed hollow bricks, with insulating layers (i.e., 3.0 cm of expanded polystyrene). Notably, the Uvaue is equal 0.73 W/m2 K−1 for the external floor (a part of the floor is external) and 0.76 W/m2 K−1 for the roof. The attention was focused on these two latter envelope components and on external vertical walls, because these separate the building from the outdoor environment, and thus they will subjected to the application of a further thermal insulation layer as potential ERM. 3.1.3. Audit of technical systems and equipment Hydronic air-conditioning systems are installed in classrooms and offices, for both space heating and cooling. The hot water is produced by a condensing gas boiler (nominal power of 388 kWt and efficiency at the low calorific value, ␩, equal to 0.98) and the chilled water is supplied by an electric air-cooled chiller (nominal power of 406 kWt , energy efficiency ratio, EER, at rated conditions, equal to 3.2 Wt /Wel ), significantly oversized. Furthermore, at the basement floor, all rooms for educational activities have a mixed air/water HVAC (heating, ventilating and air conditioning) system, given by the combination of fan coils and air handling units, not equipped with heat recovery systems. An air-cooled electric heat pump with nominal heating (cooling) power of 91.2 kW (85.5 kW) is used for the heat (cool) generation. Finally, bar and dining hall have autonomous direct expansion systems. Domestic hot water (DHW) is supplied by a quite new natural gas boiler. Benevento is inside the Italian climatic zone “C”, characterized by 1315 Heating Degree-Days, with a heating period length established from November 15 to March 31, according to the Italian law. Conversely, the space cooling is generally admitted in the interval June 1–September 30. In wintertime, the building is heated at 20 ◦ C from 8:00 am to 6:00 pm from Monday to Friday. The cooling setpoint temperature is 26 ◦ C, between 10:00 am and 6:00 pm also in this case from Monday to Friday. 3.2. Calibration of the building numerical model Geometrical and thermal–physical characteristics of the building have been defined in the simulation model as described in the previous section. The heating and cooling systems have been modelled according to the surveyed characteristics. As aforementioned, EnergyPlus has been used for simulating the energy performance

Fig. 7. A) heating energy request; B) electricity energy request.

Table 2 Calibration indexes evaluation. ERRyear EH 1.5%

MBE Eel 1.5%

EH 2.1%

CV(RMSEmonth ) Eel 2.6%

EH 3.7%

Eel 10%

of the building, by adopting transient heat transfer algorithms. The outcomes of the energy simulation, compared to the consumptions derived by the energy billings, are reported in Fig. 7a and b, with reference to the primary energy need for the space heating and the electric energy requests, respectively. The simulation outputs have been compared to the billing data, according to the “Whole Building Level Calibration with Monthly Data” approach proposed by FEMP M&V Guideline [51]. The adopted statistical indices are the annual energy consumption (ERRyear ), the coefficient of variation of the root mean squared error CV(RMSEmonth ) and the mean bias errors (MBE). Typically, models are declared to be calibrated if these produce an ERRyear within ±5%, CV(RMSEmonth ) within +15% and MBE within ±5%. Table 2 shows the evaluated indexes and the tolerance values (in the last line) separately for the space heating (EH ) and the electric energy demand for all building uses (Eel ).

312

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

A satisfactory convergence is quite evident according to the literature indications, so that the energy model can be considered as well-calibrated. 4. Optimization of building energy retrofit: results and discussion The optimization methodology, proposed in section 2, is applied to the described case study in order to identify cost-optimal retrofit solutions that imply, at the same time, a significant improvement of building energy performance towards the goal of nZEB. The following sub-sections show the outcomes provided by the two stages in which the methodology is articulated. 4.1. 1st methodology stage: three-objective optimization of ERMs to minimize TEDH , TEDC and DH This first stage aims at finding optimal packages of ERMs for the reduction of thermal energy demand for space heating (TEDH ), space cooling (TEDC ) and percentage of discomfort hours (DH). It is noted that the base building configuration (BB) is characterized by the following values of the mentioned objective functions: • TEDH,BB = 28.80 kWh/m2 a; • TEDC,BB = 7.62 kWh/m2 a; • DHBB = 43.02%. Thus, the ERMs investigated in this stage are addressed to the energy optimization of the building envelope and operation of space conditioning systems, in terms of set point temperatures. The proposed measures are listed in Table 3 in which also some constraints established by the Italian rules [52] are specified. They have been chosen based on building peculiarities, bestpractices, authors’ expertise and previous retrofit studies addressed to similar case studies [39,44,53]. Notably, Table 3 characterizes such ERMs, which are 10, by proposing a brief description, the considered options, the related investment costs as well as public incentives that are taken into account in the calculation of global costs. It should be noticed that investment costs are deduced from direct quotations of suppliers. Since the implementations of ERMs 3), 4) and 5) are connected, the cumulative investment cost is considered. The same occurs for ERMs 6), 7) and 8). Each ERM, delineated in Table 3, introduces a design variable of the optimization problem that is solved by running the GA in the 1st methodology stage. The resulting 10 design variables are characterized in Table 4, which reports the number of considered discrete options and the bits required to encode each variable. Also in this case, the variability ranges of each design variable have been chosen based on the peculiarities of the case study, best-practice, authors’ expertise and previous similar optimization studies [39,44,53]. In particular, the number of investigated options for each variable is always a power of 2 because variables are encoded by strings of bits. The variability ranges (i.e., [0.1, 0.25, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]) of infrared emissivity (e) and solar absorptance (a) of external plasters are slightly moved towards 1 (and not towards 0) because, generally, paints used for building applications present values of ‘e’ and ‘a’ closer to 1 than to 0. Concerning, the thermal insulation of external walls (ERM 5), roof (ERM 8) and external floor (ERM 9), the minimum values of the insulation layer thickness (t) are set in order to comply with the limit values of thermal-transmittance (Uvalue ) established by the Italian law [4] in case of building envelope retrofit, as shown in Table 3. The variation step of ‘t’ increases for higher values of this variable (e.g., tv varies within the range [0,0.02, 0.03, 0.04, 0.05, 0.06,

Fig. 8. Optimization of the ERMs for the reduction of TED and/or DH: Threedimensional Pareto front by minimizing TEDH , TEDC and DH in presence of the constraint DH < DHBB (43.02%).

0.08, 0.10 m]), because higher the values of t, lower the reduction of the Uvalues produced by an increment of insulant thickness. Finally, as regards the replacement of the windows, three highly efficient solutions, according to the best-practice for the considered case study, are explored, as done in a similar application proposed by the authors in [44]. Table 4 shows that the total number of bits, required to encode the 10 design variables, is 27, and thus the investigated domain includes 227 = 134217728 solutions. The assessment of energy performance of each solution needs to run an EnergyPlus simulation, which takes around 10 min for the investigated case study by ® using a processor Intel CoreTM i7 at 2.00 GHz. Therefore, the computational time required by an exhaustive sampling would be prohibitive, around ‘hundreds of years’. The GA implementation allows to tackle this issue by performing a smart research within the solution domain. Definitely, the parameters most affecting computational burden and reliability of a GA are the population size (s) and the maximum number of generation (gmax ), since the product between s and gmax provides the limit number of solutions to be explored, and thus, in the proposed application, of required EnergyPlus simulations. As shown in [44] for a similar case study – in terms of number of design variables and peculiarities of the case study as concerns use destination, climatic location and building type – a reliable value of s is 4 times the number of design variables, whereas a dependable choice for gmax is 20. These values are employed in the proposed study, and thus the maximum number of EnergyPlus simulations to be run is (s × gmax ) = (40 × 20) = 800, thereby requiring an acceptable computational time, around 5.6 days. Also the other parameters of the GA are set equal to the values used in [44], namely ce (elite count) = 2, fc (crossover fraction) = 0.6, fm (mutation probability) = 0.1, tol (tolerance in the variation of Pareto front spread as stop criterion) = 0.001. The implementation of the three-objective GA provides nondominated solutions, which represent optimal packages of ERMs and form the so-called Pareto fronts. A constraint is established, since all solutions that cause an increase of DH compared to DHBB are excluded. Thus, the acceptable solutions (DH < DHBB ) are 360. All told, the outcomes are depicted in Figs. 8–11, and in particular: • Fig. 8 shows the three-dimensional Pareto front, which collects all non-dominated solutions that lead to the minimization of TEDH , TEDC and DH; • Fig. 9 shows the bi-dimensional Pareto front of the nondominated solutions that lead to the minimization of TEDH and TEDC ;

Table 3 Characterization of the proposed ERMs for the reduction of TED and/or DH. Energy Retrofit Measures (ERMs) for the reduction of Thermal Energy Demand (TED) and/or Discomfort Hours (DH) Description and considered Options

Investment Cost

Incentives

1) Variation of space heating set point temperature (TH )

The value of TH highly affects TED and DH during the heating season. Thus, different values are considered. The established upper limit is 22 ◦ C, because higher heating set point temperatures are prohibited by the Italian law [52]. The value of TC highly affects TED and DH during the cooling season. Thus, different values are considered. The established lower limit is 24 ◦ C, as done in [38,52]. Plastering of the external vertical walls by using paints with a certain value of thermal infrared emissivity, denoted with ev . Different options are explored. The base building configuration (BB) has ev = 0.9. Plastering of the external vertical walls by using paints with a certain value of absorptance to solar radiation, denoted with av . Different options are explored. The BB has av = 0.6. Installation of an external layer of polyurethane** on the external vertical walls. Different values of the layer’s thickness (tv ) are investigated. All options provide an Uvalue of the walls, denoted with Uv , lower than the upper limit established by the Italian law in case of building envelope retrofit for climatic zone C, which is equal to Uv lim = 0.40 W/m2 K [4]. The BB is characterized by Uv BB = 0.55 W/m2 K. Plastering of the roof by using paints with a certain value of thermal infrared emissivity, denoted with er . Different options are explored. The BB has er = 0.9. Plastering of the roof by using paints with a certain value of absorptance to solar radiation, denoted with ar . Different options are explored. The BB has ar = 0.7. Installation of an external layer of polyurethane** on the roof. Different values of the layer’s thickness (tr ) are investigated. All options provide an Uvalue of the roof, denoted with Ur , lower than the upper limit established by the Italian law in case of building envelope retrofit the climatic zone C, which is equal to Ur lim = 0.34 W/m2 K [4]. The BB is characterized by Ur BB = 0.760 W/m2 K. Installation of an external layer of polyurethane on the exterior side of the external floor. Different values of the layer’s thickness (tf ) are investigated. All options provide an Uvalue of the floor, denoted with Uf , lower than the upper limit established by the Italian law in case of building envelope retrofit for climatic zone C, which is equal to Uv lim = 0.42 W/m2 K [4]. The BB is characterized by Uf BB = 0.730 W/m2 K. I) Solar Control Replacement of the existing double-glazed Uw = 2.40 W/m2 K air-filled external windows (Uw = 2.70 W/m2 K, SHGC = 0.499 SHGC = 0.760, Tvis = 0.812) with argon-filled Tvis = 0.664 double-glazed ones characterized by PVC II) Low-e frames. The transmittance to visible radiation Uw = 1.71 W/m2 K (Tvis ) is always higher than 0.65 in order to SHGC = 0.691 prevent a significant electricity demand Tvis = 0.744 increase (for artificial lighting), which is III) Solar therefore neglected. Different options are Control + Low-e explored (see on the right), characterized by Uw = 1.64 W/m2 K solar control and/or low-emissive (low-e) coatings. The same options were considered in SHGC = 0.433 Tvis = 0.691 [44] for a similar application.

Absent

Absent

Absent

Absent

[(500–2000 × tv ) × tv + 20] × Av tv is expressed in m and Av denotes the walls’ surface (=3584.1 m2 )

* 45% (55% if these ERMs are combined with the installation of efficient heating systems, such as the investigated heat pumps, see Table V) of IC up to a limit of 400000 D , accorded in 5 years. Concerning climatic zone C, for each envelope component, the incentive is provided only if thermal insulation is implemented and this induces:

2) Variation of space cooling set point temperature (TC ) 3) Variation of the walls’ thermal emissivity (ev ) 4) Variation of the walls’ solar absorptance (av )

5) Walls’ thermal insulation of thickness tv

6) Variation of the roof’s thermal emissivity (er ) 7) Variation of the roof’s solar absorptance (ar )

8) Roof’s thermal insulation of thickness tr

9) External floor’s thermal insulation of thickness tf

10) Energy efficient windows characterized by lower Uvalue (Uw ) and solar heat gain coefficient (SHGC) compared to the BB

[(500–2000 × tr ) × tr + 15] × Ar tr is expressed in m and Ar denotes the roof’s suurface (=2698.9 m2 )

• Uv ≤ 0.30 W/m2 K • Ur ≤ 0.27 W/m2 K • Uf ≤ 0.30 W/m2 K. Thus, for the considered case study, in order to have access to these incentives the following conditions must be satisfied, respectively:

• tv ≥ 0.043 m • tr ≥ 0.067 m • tf ≥ 0.055 m [(500–2000 × tf ) × tf +10] × Af tf is expressed in m and Af denotes the floor’s surface (= 404.5 m2 )

240 D per m2 of windows’ surface

250 D per m2 of windows’ surface

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

ERM

* 45% (55% in the case mentioned above) of IC, accorded in 5 years. Concerning climatic zone C, the incentive is provided only if Uw ≤ 1.75 W/m2 K (thus, only for options II and III) and the upper limit is 75 000 D .

280 D per m2 of windows’ surface

* The financial incentives is the capital grants provided by the Italian Government for the implementation of some ERMs according to [52]. ** k (thermal conductivity) = 0.028 W/m K; ␳ (density) = 25 kg/m3 , c (specific heat) = 1340 J/kg K. 313

314

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

Table 4 Characterization of the GA design variables, which are related to the implementation of ERMs for the reduction of TED and/or DH. Design Variable

Investigated options

Number of discrete Options

Bits needed for variable encoding

1) TH : space heating set point temperature 2) TC : space cooling set point temperature 3) ev : external walls’ thermal emissivity 4) av : external walls’ solar absorptance 5) tv : external walls’ insulation thickness

19, 20 (BB*), 21, 22 ◦ C 24, 25, 26 (BB), 27 ◦ C 0.1, 0.25, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 (BB) 0.1, 0.25, 0.4, 0.5, 0.6 (BB), 0.7, 0.8, 0.9 0 (BB), 0.02, 0.03, 0.04, 0.05, 0.06, 0.08, 0.10 m 0.1, 0.25, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 (BB) 0.1, 0.25, 0.4, 0.5, 0.6, 0.7 (BB), 0.8, 0.9 0 (BB), 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.12 m 0 (BB), 0.03, 0.04, 0.05, 0.06, 0.08, 0.10, 0.12 m - air-filled double glazed with aluminium frames (BB) - argon-filled double glazed with solar control coatings and PVC frame - argon-filled double glazed with low-emissive coatings and PVC frame - argon- filled double glazed with solar control and low-e coatings and PVC frame

4 4 8 8 8

2 2 3 3 3

8 8 8

3 3 3

8

3

4

2

6) er : roof’s thermal emissivity 7) ar : roof’s solar absorptance 8) tr : roof’s insulation thickness 9) tf : external floor’s insulation thickness 10) Windows’ type

* BB: base building configuration.

Fig. 9. Optimization of the ERMs for the reduction of TED and/or DH: Bi-dimensional Pareto front by minimizing TEDH and TEDC in presence of the constraint DH < DHBB (43.02%).

Fig. 11. Optimization of the ERMs for the reduction of TED and/or DH: Bidimensional Pareto front by minimizing TEDC and DH in presence of the constraint DH < DHBB (43.02%).

• Fig. 11 shows the bi-dimensional Pareto front of the nondominated solutions that lead to the minimization of TEDC and DH. Clearly, all non-dominated solutions included in any of the three bi-dimensional Pareto fronts also belong to the three-dimensional Pareto front, which collects a total of 77 solutions. These latter represent the optimal packages of ERMs for the reduction of TED and/or DH, and are thoroughly investigated in the second methodology stage. Thus, the three-objective GA allows to select recommended packages of retrofit measures for the energy optimization of building thermal envelope and space conditioning operation, which are then subjected to a deeper energy end economic analysis.

Fig. 10. Optimization of the ERMs for the reduction of TED and/or DH: Bidimensional Pareto front by minimizing TEDH and DH in presence of the constraint DH < DHBB (43.02%).

• Fig. 10 shows the bi-dimensional Pareto front of the nondominated solutions that lead to the minimization of TEDH and DH;

4.2. 2st methodology stage: assessment of cost-optimal whole retrofit solutions A smart exhaustive sampling is performed in order to investigate the performance of different technical solutions for primary energy systems in presence of the 77 optimal ERMs’ packages (included in the 3D Pareto front of Fig. 7) and also in absence of ERMs for building envelope and operation, i.e., the base building (BB) enve-

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

lope and operation. For each retrofit scenario, IC, PEC and GC are evaluated and the cost-optimal solution is identified for different values of the discount rate (i.e., r). Concerning PEC assessment, the primary energy conversion factor is set equal to 1.05 for natural gas and 1.95 for electricity, as established by the Italian law [4]. The global cost is assessed over a calculation period of 20 years, as stated in [28] for non-residential buildings, by considering the specific prices of natural gas and electricity equal to 0.91 D /Nm3 and 0.24 D /kWhel , respectively, as deduced by the building energy bills. Produced electric energy that is sold to the grid (when photovoltaic panels are implemented) is assumed remunerated at the price of 0.08 D /kWhel , as done in [44], and this value is quite reasonable according to the Italian present situation. It is noted that the BB is characterized by the following values of the cited performance indicators (clearly ICBB = 0 D ):

315

Fig. 12. Assessment of IC, dPEC and dGC for the investigated 8 190 retrofit scenarios. The considered economic discount rate (r) for GC assessment is equal to 3%.

• PECBB = 137.96 kWh/m2 a; • GCBB = 287.74 kWh/m2 a for r = 3% (as recommended in [28]). Table 5 characterizes all investigated technical solutions for primary energy systems, by delineating also the BB’s systems. Concerning the proposed system, the peak thermal power is set by considering the peak thermal load of the BB, overestimated of 5%. Only photovoltaics (PV) are investigated as RES (renewable energy source) system because, definitely, they are the most energy-efficient and cost-effective renewable technology for the considered building, given the climatic location and the huge electricity demand, as argued in [56] for a similar case study. No retrofit solutions are proposed for the domestic hot water (DHW) primary system, because the existing equipment (i.e., a quite new natural gas boiler) is characterized by a satisfying energy efficiency (i.e., ␩ = 0.92) and, furthermore, the DHW demand is quite low for the investigated use destination (i.e., an educational building). The same occurs for the existing system for artificial lighting, which is already characterized by a satisfactory efficiency. All told, the possible combinations of the considered primary energy systems are 105, and thus the total number of whole retrofit scenarios is (105 × 78) = 8190, where 78 means 77 non-dominated ERMs’ packages (detected by the GA) plus the BB configuration as regards building envelope and operation of space conditioning systems (in terms of set point temperatures). The retrofit scenarios are defined ‘whole’ because they address all levers affecting building energy performance, namely the thermal characteristics of building envelope, the operation and efficiency of energy systems and the exploitation of RESs. The outcomes are shown in the following figures, where the differences of primary energy consumption and global cost for each retrofit scenario compared to the BB – denoted as dPEC and dGC, respectively – are reported instead of the absolute values of PEC and GC in order to highlight the energy and economic savings produced by the achieved solutions. In particular: • Fig. 12 shows the values of IC, dPEC and dGC for each of the 8 190 retrofit scenarios by considering a discount rate (r) equal to 3% as recommended in [28]. The values of IC allow to appreciate the different amounts of investment required by the solutions, thereby showing that the IC of retrofit packages with ERMs for the building envelope (clear points in the figure) is, generally, more than double compared to the IC of packages without this kind of ERMs (dark points); • Fig. 13 shows the cost-optimal curve, i.e., dGC in function of dPEC, for the investigated retrofit scenarios by considering r = 3%. It is obtained from Fig. 13 by projecting all points onto the plane dPEC – dGC; • Fig. 14 shows the cost-optimal curve by considering r = 1%; • Fig. 15 shows the cost-optimal curve by considering r = 5%.

Fig. 13. Cost-optimal curve (dGC vs dPEC) for the investigated 8 190 retrofit scenarios. The considered economic discount rate (r) for GC assessment is equal to 3%.

Fig. 14. Cost-optimal curve (dGC vs dPEC) for the investigated 8 190 retrofit scenarios. The considered economic discount rate (r) for GC assessment is equal to 1%.

In all mentioned figures, the ‘star’ markers represent the costoptimal retrofit packages, which are characterized in Table 6, in function of the interest rate r, by reporting the composition in terms of ERMs and the values of IC, dPEC, dGC and DH. Also DH has been considered, in order to assess the impact of the detected costoptimal solutions on occupants’ thermal comfort as recommended in [2,28]. The outcomes are consistent with energy and economic considerations. First, as expected, when r decreases, the potential economic savings (in terms of reduction of discounted operating costs during building lifecycle), produced by the ERMs, become

316

Table 5 Characterization of the investigated primary energy systems. Considered primary energy systems Description and considered Options Heating System

Condensing gas boiler (BB*)

Heating & Cooling

Ground-source reversible heat pump

Cooling System

Air-cooled Chiller (BB) Water-cooled chiller

DHW System RESs

Efficient gas boiler (BB) Solar photovoltaic (PV) panels

Solar photovoltaic (PV) panels to be installed on the roof, south-oriented with 34◦ tilt angle. The size is expressed by the parameter cov, which provides the percentage of the available roof surface (Ar = 2698.9 m2 ) covered by PV panels. The panels are properly positioned by avoiding mutual shading. Cov varies in the range 0% (BB) – 100% with a step equal to 10%. Two popular PV panel kinds are considered (see on the right).

Investment Cost

Incentives

120000 D

16 413 D per year for a total duration of 5 years**

190000 D

21 887 D per year for a total duration of 5 years**

50000 D

poly-crystalline silicon

250 D per m2 of panels’ surface

mono-crystalline silicon

430 D per m2 of panels’ surface

* BB: base building configuration. *The financial incentives is the capital grants provided by the Italian Government for the implementation of some ERMs according to [54]. *** The financial incentives is the capital grants provided by the Italian Government for the implementation of some ERMs according to [55].

50% of IC up to a limit of 96000 D , accorded in 10 years ***

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

Air-source heat pump

Condensing natural gas boiler with peak thermal power (Pth ) of 388 kWt and LCV (lower calorific value) efficiency (␩) equal to 0.98 at rated conditions (water inlet/outlet temperatures equal to 35/55 ◦ C). Air-source electric heat pump with Pth equal to 430 kWt and COP (coefficient of performance) equal to 3.8 at rated conditions (water inlet/outlet temperatures equal to 40/45 ◦ C and outdoor temperature of 7 ◦ C). Reversible ground-source electric heat pump with geothermal vertical probes: − Heating operation: the Pth is 430 kWt and the COP is 5.1 at rated conditions (water inlet/outlet temperatures = 40/45 ◦ C); − Cooling operation: the Pth is 150 kWt and the EER (energy efficiency ratio) is 6.5 at rated conditions (water inlet/outlet temperatures = 12/7 ◦ C). Air-cooled electric chiller with Pth equal to 406 kWt and EER equal to 3.2 at rated conditions (water inlet/outlet temperatures equal to 12/7 ◦ C and outdoor temperature of 35 ◦ C). Water-cooled electric chiller equipped with Pth equal to 150 kWt and EER equal to 5.5 at rated conditions (water inlet/outlet temperatures = 12/7 ◦ C). Energy efficient natural gas boiler, nominal ␩ equal to 0.92.

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

317

Table 6 Cost-optimal whole energy retrofit packages in correspondence of difference available economic budgets: characterization of the ERMs and values assumed by the objective functions. Interest Rate (r)

See Figs. 11 and 12

3%

See Fig. 13

1%

See Fig. 14

5%

Cost-optimal Packages of Energy Retrofit Measures

ERMs to reduce TED

Primary Generation Systems RESs Heating

Cooling

PV

IC

dPEC

dGC

DH

- TH = 20 ◦ C - TC = 27 ◦ C - ev = 0.1 - av = 0.9 - er = 0.25 - ar = 0.9 - tv = 0.08 m - tr = 0.08 m - tf = 0.06 m - low-e windows - TH = 20 ◦ C - TC = 27 ◦ C - ev = 0.1 - av = 0.5 - er = 0.6 - ar = 0.7 - tv = 0.10 m - tr = 0.12 m - tf = 0.08 m - low-e windows absent

air-source heat pump

air-cooled chiller

polycrystalline cov = 100%

917.81 kD

−130.93 kWh/m2 a

−125.94 D /m2

38.99%

air-source heat pump

air-cooled chiller

polycrystalline cov = 100%

940.43 kD

−131.29 kWh/m2 a

−197.10 D /m2

39.99%

air-source heat pump

air-cooled chiller

polycrystalline cov = 100%

456.01 kD

−126.00 kWh/m2 a

−90.48 D /m2

43.02%

Performance Indicators*

Legend: –configuration already present in the base building configuration (BB); – configuration non present in the BB (=ERM). *Performance indicators: IC: investment cost; dPEC: difference in primary energy consumption compared to BB; dGC difference in global cost compared to BB; DH: annual percentage of discomfort hours. Clearly, negative values of dPEC and dGC provide energy and economic savings, respectively.

Fig. 15. Cost-optimal curve (dGC vs dPEC) for the investigated 8 190 retrofit scenarios. The considered economic discount rate (r) for GC assessment is equal to 5%.

higher. Thus, in this case, impactful retrofit scenarios, with several ERMs that produce significant energy savings, become more cost-effective as shown by the cost-optimal solutions achieved for r = 3% and r = 1%. Indeed, these solutions include all three kinds of ERMs, namely those addressed to building envelope, energy systems and RESs’ exploitation, respectively. On the other hand, the opposite occurs when r increases. In this regard, the cost-optimal solution for r = 5% includes, merely, the installation of an air-source heat pump and of PV panels. Moreover, the described trend is clear in Figs. 13–15, where the clear points represent retrofit solutions that include ERMs for the building envelope, and thus characterized by higher values of IC and potential PEC savings, whereas the dark points represent the solutions without this kind of ERMs, i.e., lower values of IC and PEC savings. When r = 3% clear and dark points are almost overlapped (see Fig. 13), while, as expected based on the above observation, clear points are generally below (i.e., these solutions are more cost-effective) when r decreases (see Fig. 14) and

above (i.e., these solutions are less cost-effective) when r increases (see Fig. 15). For what concerns the discussion about the achieved numerical outcomes, it should be noted that all cost-optimal solutions include the installation of an air-source heat pump for space heating and of a full-roof PV systems, composed of panels in polycrystalline silicon. Therefore, these ERMs are, definitely, the most cost-effective ones. Concerning the heat pump, this is due to the magnitude of space heating demand given the investigated climatic location, while the high cost-effectiveness of PV panels depends on the huge entity of electricity demand for university/office buildings, as the considered case study. In the most pessimistic scenario (r = 5%), the other ERMs are not cost-effective. In particular, the replacement of the chiller is never effective because of the low entity of space cooling demand, mainly due to the university closure for a significant period of the cooling season. On the other hand, the improvement of the building envelope energy performance is not effective for r = 5%, because the BB’s envelope is already characterized by a good level of thermal insulation. However, as previously argued, in a neutral (r = 3%) or optimistic scenario (r = 1%), ERMs addressed to the envelope become cost-effective, and when r decreases from 3% to 1% the costoptimal insulation layer thicknesses increase (see Table 6) because this produces higher energy savings, which provide further GC savings for r = 1%. However, thicker insulating layers penalize the space conditioning demand and thermal comfort during the cooling season, because of the overheating effect, which can be high in low-energy buildings [57], especially in presence of high internal heat gains, such us in universities. This is partially compensated by the increase of thermal emissivity (e) and the decrease of solar absorptance (a) of external plasters, which can be observed in the cost-optimal retrofit solution when r passes from 3% to 1%. Lastly, in both cases, the cost-optimal choice for the windows consists of argon-filled double-glazing systems with low-emissive coatings, which cause a substantial reduction of space heating

318

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

demand. Solar control coatings are not effective because cooling demand is not significant. It is noticed that all ERMs for the thermal insulation of opaque and transparent envelope, included in both cost-optimal solutions for r = 3% and r = 1%, comply with the limit Uvalues required to have access to public incentives (see Table 3), because this gives higher GC savings. As expected, the implementation of ERMs addressed to the building envelope induces an improvement of thermal comfort (DH decreases compared to DHBB), since it determines more comfortable mean radiant temperatures of internal walls’ surfaces. Definitely, all these considerations demonstrate the reliability and consistency of outcomes and methodology. Finally, prudent stakeholders will decide to act only on energy systems by implementing the cost-optimal solution for r = 5%, whereas more optimistic and virtuous ones (e.g., public administrations that should play a demonstration role) will act also on the building envelope by implementing the cost-optimal solution for r = 3% or r = 1%, thereby achieving an increase of thermal comfort too. In any case, the potential energy and economic savings are substantial and the global building PEC will be lower than 12 kWh/m2 a. Therefore, the implementation of any cost-optimal solution delineated in Table 6 will make the investigated building a nZEB. This shows that the cost-effective energy retrofit of the existing building stock into nZEBs is not a dream, but it is possible and fundamental in the long path towards sustainability. Definitely, the proposed methodology is not user-friendly, since it requires expertise in building energy modelling as well as in the implementation of the employed optimization algorithms. Therefore, the potential users are building scientists and specialized working groups that own the know-how to use appropriate, reliable building performance simulation and optimization software. In this regard, the main difficulties they may face concern the coupling between the software, as well as the proper choice of design variables and simulation, optimization parameters. In this frame, in order to enhance the applicability and replicability of the methodology, future works will be focused on its simplification, also by means of the use of other software, in order to develop a more user-friendly tool for most building practitioners.

5. Conclusions The proposed investigation showed a multi-step and multiobjective optimization of the energy refurbishment process of an educational building, by means of a repeatable methodology, completely described, that allows to scientists, professionals and designers to optimize the energy retrofit of existing buildings. The core of the method consists in the combined use of a mathematical code for formulating the optimization problem, by means of genetic algorithms and of one of the most suitable software for the whole energy simulation of buildings, under transient conditions of heat transfer. First of all, according to the demonstrative and exemplary role of the public hand, being the edifice owned by a public Institution, and thus the University of Sannio, in Italy, the building model was carefully built and performed. In particular, the base model was carefully calibrated for what concerns both the input data (e.g., in situ surveys, such as heat flux measurements and use of infrared-thermography, analyses of documentations) as well as about the results of simulation. In detail, the base case condition (present building) was calibrated in order to have a basic scenario that well-represents the present energy demands, by verifying the mean bias error (MBE) and the coefficient of variation of the root mean square deviation (CV RMSE), by means of comparison of simulated and monitored data concerning the energy demand. Once achieved a calibrated numerical model, single measures of energy

efficiency and packages have been simulated, by selecting the most efficient and the most promising by means of the development of a new optimization method. This was aimed at reducing the number of simulations and generating, through genetic evolutions, the best combinations for what concerns the suitability according to a multi-objective approach of contrasting aims. In a first phase, three objective functions to minimize were identified, and thus: the a) indoor thermal discomfort, b) the thermal energy need for the space heating and c) for the space cooling. For what concerns the energy retrofit measures (ERMs), these were grouped into three macro-categories: a) the building thermal envelope), 2) the heat generation for the microclimatic control, c) the renewable energy source. More in detail: a Building envelope: use of special plasters, innovative coatings, thermal insulating layers, exploitation of thermal mass, new lowemissive windows and solar screens; b Heating and cooling systems: selection of new, high-efficient generation systems, such as efficient air-source heat pump, ground source heat pump (GSHP), water-cooled chiller with cooling tower; c Installation of photovoltaics modules. The aforementioned ERMs were characterized by several configurations, for instance by changing, among a defined number of discrete values, the thicknesses of the thermal insulation, the spectral characteristics of external building coatings (i.e., solar reflectance and thermal absorptance), the type and size of heat generators. Once achieved the three-dimensional surface of Pareto (that expresses all non-dominated configurations), an innovative criterion of choice was proposed, and thus the selection of the package of energy retrofit measures that allows the lower global cost. In particular, the lower overall expenditure, and thus the sum of investment and operating costs along the building lifetime, was identified, also by taking into account different discounting factors, in order to make really feasible the energy refurbishment. The main originality of this research, beyond the interesting outcomes, concerns the multi-step methodology of the multiobjective optimization, applied to the building energy engineering. This allows, among all the possible combinations of energy conservations measures, the sole simulations of the most profitable configurations, by reducing enormously the computational effort and time, and by evidencing, among the solutions of the Pareto front, also a valid criterion of selection based on economic criteria of feasibility. Finally, for what concerns the results of the specific case study, and thus an educational building located in the Mediterranean area (i.e., Italian backcountry), the most profitable configurations of energy retrofit include installation of an air-source heat pump for the space heating and of a full-roof PV system. More in deep, an efficient heat pump satisfies the magnitude of space heating and the cost effectiveness of PV is explained because of the huge request of electricity. On the other hand, the improvement that can be achieved by refurbishing the building envelope is significant too, but not feasible if prudential scenarios of economic trends are taken into consideration. Moreover, by implementing the best packages of energy conservation measures, the energy demand of the building can be drastically lowered, until a value of PEC (primary energy demand) of about 12 kWh/m2 a, and thus by achieving the target of nZEB in existing building. References [1] European Commission, Communication from the commission to the European Parliament, the council, the European economic and social

F. Ascione et al. / Energy and Buildings 144 (2017) 303–319

[2]

[3]

[4]

[5] [6] [7]

[8]

[9]

[10] [11] [12]

[13]

[14]

[15] [16]

[17]

[18]

[19]

[20] [21] [22]

[23]

[24]

[25]

[26] [27]

[28]

[29]

[30]

Committee and the committee of the regions An EU Strategy on Heating and Cooling, COM(2016) 51 final, Brussels, Belgium 2016. European Commission and Parliament Directive 2010/31/EU of the European Parliament and of the Council of 19 May 2010 on the Energy Performance of Buildings (EPBD Recast); Brussels, Belgium, 2010. J. Groezinger, T. Boermans, A. John, et al., Overview of Member States information on NZEBs, Project number: BUIDE14975, Ecofys by order of: European Commission, 08.10.2014, 2014. Italian Government Decree, Decreto interministeriale 26 giugno 2015, Applicazione delle metodologie di calcolo delle prestazioni energetiche e definizione delle prescrizioni e dei requisiti minimi degli edifici, 2017 [in Italian]. V. Corrado, I. Ballarini, Refurbishment trends of the residential building stock: analysis of a regional pilot case in Italy, Energy Build. 132 (2016) 91–106. International Energy Outlook 2016 - DOE/EIA-0484 (2016) I May 2016, 2016, Available online at https://www.eia.gov/forecasts/ieo/pdf/0484(2016).pdf. X. Zhou, J. Yan, J. Zhu, P. Cai, Survey of energy consumption and energy conservation measures for colleges and universities in Guangdong province, Energy Build. 66 (2013) 112–118. M.H. Chung, E.K. Rhee, Potential opportunities for energy conservation in existing buildings on university campus: a field survey in Korea, Energy Build. 78 (2014) 176–182. M. Santamouris, Innovating to zero the building sector in Europe: minimising the energy consumption, eradication of the energy poverty and mitigating the local climate change, Sol. Energy 128 (2016) 61–94. D. Li, L. Yang, J. Lam, Zero energy buildings and sustainable development implications – a review, Energy 54 (2013) 1–10. M. Kapsalaki, V. Leal, Recent progress on net zero energy buildings, Adv. Build. Energy Res. 5 (1) (2011) 129–162. D.K. Serghides, S. Dimitriou, M.C. Katafygiotou, M. Michaelidou, Energy efficient refurbishment towards nearly zero energy houses, for the mediterranean region, Energy Procedia 83 (2015) 533–543. M. Ferreira, M. Almeida, A. Rodrigues, Cost-optimal energy efficiency levels are the first step in achieving cost effective renovation in residential buildings with a nearly-zero energy target, Energy Build. 133 (2016) 724–737. G. Ciulla, A. Galatioto, R. Ricciu, Energy and economic analysis and feasibility of retrofit actions in Italian residential historical buildings, Energy Build. 128 (2016) 649–659. K. Kuusk, T. Kalamees, nZEB retrofit of a concrete large panel apartment building, Energy Procedia 78 (2015) 985–990. A. Passer, C. Ouellet-Plamondon, P. Kenneally, V. John, G. Habert, The impact of future scenarios on building refurbishment strategies towards plus energy buildings, Energy Build. 124 (2016) 153–163. M. Patterson, J. Vaglio, D. Noble, Incremental fac¸ade retrofits: curtainwall technology as a strategy to step existing buildings toward zero net energy, Energy Procedia 57 (2014) 3150–3159. M. Dabaieh, N.N. Makhlouf, O.M. Hosny, Roof top PV retrofitting: a rehabilitation assessment towards nearly zero energy buildings in remote off-grid vernacular settlements in Egypt, Sol. Energy 123 (2016) 160–173. M. Alirezaei, M. Noori, O. Tatari, Getting to net zero energy building: investigating the role of vehicle to home technology, Energy Build. 130 (2016) 465–476. A. Aksamija, Regenerative design and adaptive reuse of existing commercial buildings for net-zero energy use, Sustainable Cities Soc. 27 (2016) 185–195. A. AlAjmi, H. Abou-Ziyan, A. Ghoneim, Achieving annual and monthly net-zero energy of existing building in hot climate, Appl. Energy 165 (2016) 511–521. T. Niemelä, R. Kosonen, J. Jokisalo, Cost-optimal energy performance renovation measures of educational buildings in cold climate, Appl. Energy 183 (2016) 1005–1020. G. Semprini, C. Marinosci, A. Ferrante, G. Predari, G. Mochi, M. Garai, R. Gulli, Energy management in public institutional and educational buildings: the case of the school of engineering and architecture in Bologna, Energy Build. 126 (2016) 365–374. M. Tahsildoost, Z.S. Zomorodian, Energy retrofit techniques: an experimental study of two typical school buildings in Tehran, Energy Build. 104 (2015) 65–72. F. Ascione, N. Bianco, R.F. De Masi, F. de’Rossi, G.P. Vanoli, Energy retrofit of an educational building in the ancient center of Benevento. Feasibility study of energy savings and respect of the historical value, Energy Build. 95 (2015) 172–183. T. Dalla Mora, F. Cappelletti, F. Perona, P. Romagnonia, F. Bauman, Retrofit of an historical building toward NZEB, Energy Procedia 78 (2015) 1359–1364. C. Becchio, S.P. Corgnati, C. Delmastro, V. Fabi, P. Lombardi, The role of nearly-zero energy buildings in the transition towards Post-Carbon Cities, Sustainable Cities Soc. 27 (2016) 324–337. European Commission Commission Delegated Regulation (EU) No 244/2012 of 16 January 2012 Supplementing Directive 2010/31/EU of the European Parliament and of the Council on the Energy Performance of Buildings; Brussels, Belgium, 2012. C. Becchio, P. Dabbene, E. Fabrizio, V. Monetti, M. Filippi, Cost optimality assessment of a single family house: building and technical systems solutions for the nZEB target, Energy Build. 90 (2015) 173–187. J. Kurnitski, A. Saari, T. Kalamees, M. Vuolle, J. Niemelä, T. Tark, Cost optimal and nearly zero (nZEB) energy performance calculations for residential buildings with REHVA definition for nZEB national implementation, Energy Build. 43 (2011) 3279–3288.

319

[31] M. Ferrara, E. Fabrizio, J. Virgone, M. Filippi, Appraising the effect of the primary systems on the cost optimal design of nZEB: a case study in two different climates, Energy Procedia 78 (2015) 2028–2033. [32] P.M. Congedo, C. Baglivo, D. D’Agostino, I. Zacà, Cost-optimal design for nearly zero energy office buildings located in warm climates, Energy 91 (2015) 967–982. [33] J. Lizana, Á. Barrios-Padura, M. Molina-Huelva, R. Chacartegui, Multi-criteria assessment for the effective decision management in residential energy retrofitting, Energy Build. 129 (2016) 284–307. [34] M.M. Sesana, G. Salvalai, Overview on life cycle methodologies and economic feasibility for nZEBs, Build. Environ. 67 (2013) 211–216. [35] S. Carlucci, G. Cattarin, F. Causone, L. Pagliano, Multi-objective optimization of a nearly zero-energy building based on thermal and visual discomfort minimization using a non-dominated sorting genetic algorithm (NSGA-II), Energy Build. 104 (2015) 378–394. [36] EnergyPlus the Official Building Simulation Program of the United States Department of Energy, 2017 (Available online: http://www.eere.energy.gov/ buildings). [37] P. Penna, A. Prada, F. Cappelletti, A. Gasparella, Multi-objectives optimization of Energy Efficiency Measures in existing buildings, Energy Build. 95 (2015) 57–69. [38] M. Hamdy, A.T. Nguyen, J.L.M. Hensen, A performance comparison of multi-objective optimization algorithms for solving nearly-zero-energy-building design problems, Energy Build. 121 (2016) 57–71. [39] F. Ascione, N. Bianco, C. De Stasio, G.M. Mauro, G.P. Vanoli, A new methodology for cost-optimal analysis by means of the multi-objective optimization of building energy performance, Energy Build. 88 (2015) 78–90. ® [40] MATLAB - MATrixLABoratory, 7.10.0. User’s Guide MathWorks, 2010 (All information available online: http://it.mathworks.com). [41] M. Hamdy, A. Hasan, K. Siren, A multi-stage optimization method for cost-optimal and nearly-zero-energy building solutions in line with the EPBD-recast 2010, Energy Build. 56 (2013) 189–203. [42] F. Ascione, Energy Conservation and Renewable Technologies for Buildings to face the impact of the climate change and minimize the use of cooling, Sol. Energy (2017), http://dx.doi.org/10.1016/j.solener.2017.01.022 (Elsevier, article in press). [43] F. Ascione, N. Bianco, R.F. De Masi, G.M. Mauro, G.P. Vanoli, Design of the building envelope: a novel multi-objective approach for the optimization of energy performance and thermal comfort, Sustainability 7 (8) (2015) 10809–10836. [44] F. Ascione, N. Bianco, C. De Stasio, G.M. Mauro, G.P. Vanoli, Multi-stage and multi-objective optimization for energy retrofitting a developed hospital reference building: a new approach to assess cost-optimality, Appl. Energy 174 (2016) 37–68. [45] K. Deb, Multi-Objective Optimization Using Evolutionary Algorithms, John Wiley & Sons Chichester, UK, 2001. [46] European Committee for Standardization (CEN), 008 EN 15603 Standard: Energy Performance of Buildings – Overall Energy Use and Definition of Energy Ratings, 2017, pp. 2. [47] Design Builder v.4.7., 2016 http://www.designbuilder.co.uk/. [48] Italian Unification body (UNI), Standard UNI 15251: Indoor Environmental Input Parameters for Design and Assessment of Energy Performance of Buildings Addressing Indoor Air Quality, Thermal Environment, Lighting and Acoustics, 2008. [49] International Organization for Standardization (ISO), Standard ISO 9869-1: Building Elements, In-situ Measurement of Thermal Resistance and Thermal Transmittance, Part 1: Heat Flow Meter Method, 2014. [50] Italian Unification body (UNI), Standard UNI EN ISO 6946: Building Components and Building Elements – Thermal Resistance and Thermal Transmittance – Calculation Method, 2008. [51] U.S. Department of Energy Federal Energy Management Program, M&V Guidelines: Measurement and Verification for Federal Energy Projects Version 3.0, U.S. Department of Energy Federal Energy Management Program, Boulder, CO, 2008. [52] Decree of the President of the Republic (DPR), 26 agosto 1993. n. 412, Regolamento recante norme per la progettazione, l’installazione, l’esercizio e la manutenzione degli impianti termici degli edifici ai fini del contenimento dei consumi di energia, in attuazione dell’art. 4, comma 4, della L. 9 gennaio, 1991 (in Italian). [53] F. Ascione, N. Bianco, C. De Stasio, G.M. Mauro, G.P. Vanoli, Simulation-based model predictive control by the multi-objective optimization of building energy performance and thermal comfort, Energy Build. 111 (2016) 131–144. [54] Italian Government Decree, Decreto interministeriale 16 febbraio Aggiornamento Conto termico, 2016 (in Italian). [55] Italian Government Law, Legge 23 Dicembre 2014 n. 190, G.U. 29/12/2014 (Legge di Stabilità, 2015 (in Italian). [56] G.M. Mauro, M. Hamdy, G.P. Vanoli, N. Bianco, J.L.M. Hensen, A new methodology for investigating the cost-optimality of energy retrofitting a building category, Energy Build. 107 (2015) 456–478. ´ M. Lopuˇsniak, Impact of shading structure on energy demand [57] D. Katunsky, and on risk of summer overheating in a low energy building, Energy Procedia 14 (2012) 1311–1316.