Multistage Optimization toward a Nearly Net Zero Energy Building Due to Climate Change

Abstract

Climate change is one of the major problems of the planet. The atmosphere is overloaded with carbon dioxide caused by fossil fuels that are burned for energy. Almost 40 percent of the total energy worldwide is used by the building sector, which comes from non-renewable sources and contributes up to 30% of annual greenhouse gas emissions globally. The building sector in Iran accounts for 33.8% of Iran’s total energy usage. Within the building sector, the energy consumption of Iranian educational buildings is 2.5 times higher than educational buildings in developed countries. One of the most effective ways of reducing global energy consumption and greenhouse gas emissions is retrofitting existing buildings. This study aims to investigate whether a particular energy-optimized design under the present climate conditions would respond effectively to future climate change. This can help designers make a better decision on an optimal model, which can remain optimal over the years based on climate change. For methodological purposes, multistage optimization was used to retrofit an existing educational building. Specifically, the non-dominated sorting genetic algorithm (NSGA-II) was chosen to minimize the cooling and heating load, as well as consider investment costs for present and future weather files, using the jEPlus tool. Furthermore, the TOPSIS method was used to identify the best set of retrofit measures. For this purpose, a four-story educational building in Tehran was modeled on Design Builder software v7.0.0.116 as a case study to provide a better understanding for researchers of how to effectively retrofit a building to achieve a nearly zero energy building considering climate change. The results show that the optimized solution for the present weather file does not remain the optimized solution in 2080. Moreover, it is shown that to have an optimized building in regard to future weather files, the model should be designed for the future weather conditions. This study shows that if the building becomes optimized using the present weather file the total energy consumption will be reduced by 65.14% and 86.18% if using the future weather file. These two figures are obtained by implementing active and passive measures and show the priority of using the future weather file for designers. Using PV panels also, this building is capable of becoming a nearly net zero building, which would produce about 90% of its own energy demands.

1. Introduction

The outdoor climate and indoor environment of a building are inextricably bound up with each other; so, a building’s energy consumption is closely related to outdoor climate factors. The atmosphere is overloaded with carbon dioxide caused by fossil fuels being burned for energy. The Intergovernmental Panel on Climate Change (IPCC) estimates that global greenhouse gas (GHG) emissions increased by 70% between 1970 and 2004 (IPCC, 2007) [1]. According to the International Energy Agency (EAI), to keep global warming below 2 °C, CO₂ emissions need to be reduced by 77% by 2050. Climate change is estimated to have an effect on the environment [2]. A building provides a comfortable indoor environment, and the change in outdoor climate will affect building energy consumption. Many studies have shown the effects of global climate change on the energy consumption of buildings around the world. Wan et al. examined the impact of climate change on the energy consumption of buildings in different climate zones in China for two emission scenarios. The results show that more energy consumption would result in larger emissions, and consequently larger emissions would lead to climate change and global warming. The impact of climate change on building energy use could be reduced by energy conservation measures [3]. Ciancio et al. examined the relative impact of global warming for the future years 2050 and 2080 across 19 European cities of different Koppen-Geiger climate classes to compare the current and estimated future energy requirements of a hypothetical house [4]. The results show that, due to the progressive increase in average temperatures in 2050 and 2080, the heat energy demand for heating will decrease, while the demand for electricity for cooling will increase over this period of time.

Almost 40 percent of total energy worldwide is used by the building sector, which comes from non-renewable sources and contributes up to 30% of annual greenhouse gas emissions globally [1,5]. The energy consumption of Iranian educational buildings contributed up to 33.8% of total energy consumption and 23.37% of total carbon dioxide emissions in 2014 [6]. Over the past few decades, there has been an average growth rate of 8% per year in energy consumption in Iran [7]. The building sector is capable of affecting almost 50% of annual CO₂ emissions if the produced carbon footprint is eliminated by optimizing energy consumption and generating renewable energy [8]. The energy efficiency of a building is inextricably linked to the subsequent greenhouse gas emissions (GHGs) during its life cycle [9]. If energy efficiency improvements are not made in the building sector, energy consumption could increase by 50% by 2050 [10].

According to the International Energy Agency (EAI), global energy demand rose 2.3% in 2018, and global energy-related CO₂ emissions rose 1.7% to 33 gigatons (Gt) in 2018 [11] Energy demand due to the growing population and developing urbanization results from the development of sustainable buildings [6,12]. A low or nearly zero energy building (NZEB) has a substantial effect on many aspects including global energy efficiency, energy conservation, and the integration of renewable energy systems [12].

Some mitigation and adaptation strategies are necessary to consider the impact of buildings on the environment [13]. In this regard, some policies have been implemented throughout the world to control the effect of rising temperature and add to the efficiency of buildings. For instance, the implementation of the European Union 2020 energy and climate targets has triggered the transformation of Europe’s neighborhoods to net zero energy districts [14] The U.S. target is to achieve zero energy by 2050 for all commercial buildings [5]. The EU set three key targets by 2020 to include a 20% decrease in greenhouse gas emissions (compared to 1990 levels), providing 20% of EU energy from renewable sources, and a 20% improvement in energy efficiency [14].

Due to the higher energy consumption and user density of an educational building, as well as atypical daily/annual usage behavior and high heat susceptibility of their users, these buildings belong to the critical category. Poor room conditions can endanger both the health and productivity of students and staff. There are many types of buildings built in different locations to supply any country’s future plans [15]. Within the building sector, schools have educational responsibility. According to the U.S. Energy Information Administration commercial buildings consume 18% of the total usage in the USA [16]. Universities account for almost 12% of energy consumption in the commercial sector and are energy-intensive buildings. Due to the limited personal responsibility of the residents, they are in a precarious situation with regard to energy savings [17].

Educational buildings in Iran consume energy about 2.5 times more than in a developed country [6]. Public sector buildings are energy intensive in many developing countries [18] Various research shows that higher energy savings would occur if schools were efficiently adapted to real energy needs, which would lead to a reduction in CO₂ emissions [19]. Poor thermal comfort conditions, moreover, can hurt children because of their age, body, and differences in metabolism, compared to adults. Different resource consumption patterns, moreover, can be seen in educational buildings compared with non-commercial ones. These differences can be caused by the level of activity in the building, clothing requirements, and age. Classrooms typically experience high internal gains and excessive indoor overheating as they operate at full capacity during the hottest hours of the day. Although, the number of schools in Tehran as the capital city of Iran is the highest, Figure 1 shows that there have been no particular investigation or guidelines for Net Zero Energy or an energy-efficient design of school buildings in the Mediterranean climate of Tehran with a hot summer and cold winter.

The first aim of retrofit is to reduce the energy consumption of the remaining useful life of the building, while upgrading thermal comfort; the second should be to restore, sustain, or increase the useful life of existing habitations [20].

Since most Iranian buildings are old and energy-saving measures have only been implemented in a few schools, these buildings waste a lot of energy. Only by retrofitting these old buildings can the energy footprint of the building stock be significantly reduced. The thermal comfort of the residents, reduction in energy consumption, and climate change mitigation can be the result of retrofitting of the existing building [6]. In the world of retrofitting, installing energy-efficient heating, ventilation, and air conditioning (HVAC) systems, LED lighting, and building management systems can help to build managers reduce energy consumption in buildings [20]. According to the European Energy Performance of Buildings Directive (EPBD) of 2003, there are three key objectives for improving the energy efficiency of the building stock: (1) mitigating climate change; (2) ensuring energy security; and (3) eradicating energy poverty [21].

Generally, sustainability is the major issue in which the construction industry is engaged [22]. In order to satisfy sustainability objectives [23], building design and operation should be optimized to tackle climate change and greenhouse gas emissions. However, there is not adequate comprehensive research on the optimization of buildings under climate change conditions and only a few studies on the impact of climate change on building energy consumption. Only a few of these have suggested optimized retrofitting measures for an existing building to have an acceptable performance during their lifespan, with regard to climate change. Figure 2 shows the percentage of the building types among the studied cases. Despite the fact that climate change and variations in stimulation parameters during the building’s lifetime may affect the optimized design, most of the building energy optimizations are generally carried out under current climatic conditions with fixed simulation.

The aim of this study is to investigate whether a certain energy-optimized design would react effectively to future climate change under the current climatic conditions. Accordingly, a simulation-based optimization is developed in this article that uses climate models and NSGA-II optimization to compare the energy-optimized designs under current and future climates. This optimization is applied to a typical school building in Tehran. For these reasons, this study aims to investigate whether a particular energy-optimized design under present climate conditions would respond effectively to future climate change scenarios. The thermal energy performance of the building, moreover, will be considered for both present and future conditions.

The paper is divided into a description of energy retrofit measures, methods and materials, results, discussions, and conclusions.

Energy Retrofit Measures

There are significant challenges in the energy sector worldwide that become even more serious every day. The introduction of modern technologies has affected a fundamental transformation of the construction industry in many aspects [22], and energy efficiency measures are widely spread. These days the most important issue is to determine the more effective and reliable proven measures for the long term. Moreover, buildings’ lifespans are increasing due to the wide application of building technologies, so the building owner tends to take retrofit measures to improve building performance during its lifespan.

To reach the best possible solution among different proposed measures, decision makers have to take into consideration environmental, financial, and social factors, as well as energy consumption. The chosen measure should ensure both the maximization of the energy efficiency of a building, satisfaction, and the building’s final user/occupant/owner demands.

There are various measures that could be taken into consideration for improved energy efficiency in buildings. The basic categories are as follows [24,25]:

• Measures to improve the building envelope, such as supplementing or improving the insulation, changing colors, placing heat-insulating door and window frames, increasing the thermal mass, super-insulated building envelopes, building shape and orientation, etc.
• Reducing the heating and cooling loads of buildings using some strategies (i.e., incorporating passive heating and cooling techniques such as insulation of the building fabric (i.e., roof, wall, etc.), retrofitting of windows (i.e., multiple glazing, shading, etc.), reinforcements, airtightness, etc.).
• Using renewable energy such as solar thermal systems, building integrated photovoltaic, hybrid systems, wind energy, biomass systems, geothermal energy, etc.
• Using “intelligent” devices for managing energy consumption, i.e., advanced sensors, energy control (zone heating and cooling) and monitoring systems, light controlling system, etc.
• Measures for improving the indoor quality of the room with a view to minimizing energy requirements (increasing ventilation performance, use of mechanical ventilation with heat recovery, improvement of boilers and air conditioning systems, efficient use of multifunctional devices).
• Use of energy-efficient appliances and compact fluorescent lighting.
• Improvement of human-related factors such as energy consumption patterns, including comfort requirements, occupancy regimes, management and maintenance, occupant activities, access to controls, etc.
• Energy-efficient equipment and low energy technologies, i.e., control upgrade, natural ventilation, lighting upgrade, thermal storage, energy-efficient equipment and appliances, heat recovery, etc.

2. Materials and Methods

2.1. Overview

This study pursued a simulation-based assessment approach for the renovation of an educational building under different weather scenarios. The research employed DesignBuilder and EnergyPlus as a building simulation tool. It is important to guarantee the correctness of the model and results, so verification was performed after the initial simulation.

In this research, the feasibility of the application of multiobjective optimization techniques was investigated, and the improvement of the energy efficiency in building passive strategies was incorporated. Passive strategies were applied to the case in order to reduce the building’s heating and cooling consumption. PV panels, moreover, were installed on the roof to provide as much of the electricity demand of the building as possible. The PV panels were sloped by 45 degrees to achieve the best performance from the southern orientation.

The energy model was based on the building envelope retrofit optimization. A building envelope is used to separate the building from the outside environment. This separation has a substantial effect on a building’s thermal behavior, because its construction determines heating and cooling requirements for the present and future conditions, and is considered as a current design model.

The availability of materials in the studied city is an important factor for making decisions about strategies. Consequently, insulation material, window glazing construction, and the thickness of insulation were selected as passive strategies.

It is important for nearly net zero-energy buildings to have thermal insulation. Applying insulation is particularly important for buildings in colder climates. The use of proper insulating material with proper thickness is essential for providing good thermal resistance to the building envelope. Commonly used insulation materials are soft, such as glass wool and rock wool, or rigid, such as extruded polystyrene insulation (XPS) or expanded polystyrene insulation (EPS) or PVC.

The high thermal resistance or R-value of these materials is calculated by multiplying the thickness of the material by the thermal conductivity. Reduction in cooling load by using natural ventilation is one of the most effective design strategies in future buildings to combat climate change [26]. In this study, both natural and mechanical strategies were incorporated.

A window is capable of controlling the environment by reducing the need for cooling, heating, and lighting. Various glazing types with different window gas materials were used in this study as another passive strategy. Double clear glazing (DblClr) filled with argon or air and ranging from 1 mm to 10 mm thickness was used.

The chosen glazing types were: (1) glazing type with an outermost pane thickness of 13 mm and innermost pane thickness of 6 mm or 3 mm, filled with argon; (2) glazing type with outermost pane thickness of 13 mm and innermost pane thickness of 6 mm or 3 mm, filled with air; and (3) glazing type with outermost pane thickness of 6 mm and innermost pane thickness of 6 mm or 3 mm, filled with air. Passive measures, furthermore, were used to reduce the external cooling load, which is caused by the direct solar heat gain and heat conduction through the building, so window shading was helpful in this case.

2.2. Constructing Hourly Future Weather Years

Jentsch et al. developed a Microsoft Excel-based software tool to generate EPW (EnergyPlus/EPSr Weather) and TMY2 (Typical Meteorological Year Data) weather files for every place in the world for the A2 (medium-high) emission scenario for two future time slices, the 2050s and 2080s [27]. The UK Met Office Hadley Center Coupled Model 3 HadCM3 was used to model future climate scenarios. The Climate Change World Weather File Generator (CCWorldWeatherGen) is used to generate climate change weather files based on the model summary data from the Intergovernmental Panel on Climate Change (IPCC) Third Assessment Report of the HadCM3 A2 experiment.

For data evaluation purposes, the weather station Tehran-Mehrabad was selected. This station was chosen, mainly because the educational buildings in this climate have not been highly regarded; on the other hand, a ‘present day’ EPW file was available only for a few cities in Iran. Figure 3 compares the dry temperature of Tehran for the present weather file and 2080. The darker color shows the present dry temperature, while the lighter one shows the 2080 dry temperature. This graph shows that the dry temperature will increase in the future. So, it is necessary to propose an optimized solution for both climate conditions.

2.3. Study Platform

Based on the climate classification of the Köppen, Iran has hot–humid, hot–dry, mild and humid, and cold climate. A typical educational building (high school) was evaluated in the hot summer and cold winter zone of the Iranian capital Tehran. In this study, Design Builder was used to create the simulation platform with a building model, HVAC systems, and a renewable energy system model for this city.

Figure 4 shows the Köppen climate classification of meteorological stations in Iran and Figure 5 shows the temperature change in Tehran over 12 months.

The building had three floors, with three stories above the ground and one story under the ground.

Table 1. Key design data of the building envelope.

Building Component	U Value (W$/{\mathbf{m}}^2-\mathit{K})$
Roof	1.713
Ceiling	0.814
Internal wall	2.093
External wall	2.17
Basement wall	4.038
Window glazing	2.665

shows the key design data of the building envelope. The total gross area of the building was 2132.84 m², the height of the underground and ground floor was 3.2 m, and the height of the first and the second floor was 4.4 m. Figure 6 illustrates the first-floor plan as well as the east view of the studied building.

The building was representative of a typical educational building (high school) in Tehran, built in 2008. The building was heated by radiators (COP = 0.8), which produced hot water using a condensing gas boiler. Furthermore, a water cooler (COP = 7.8) was used as an air-conditioning system in hot weather. In addition to the water cooler as a mechanical cooling system, the building was ventilated naturally using windows. Other end-uses included lighting, teaching facilities that used electricity, and domestic hot water, which used natural gas. The prime energy demand was for space heating, because the school was closed in summer. The energy model was designed for space heating and cooling, and calibrated with a total monthly gas and electricity consumption of the building.

UPVC windows and a wooden door were included in each classroom, office, kitchen, and other related rooms. Figure 7 shows the classroom view. T8 (25 mm diam) fluorescent halo phosphate was used as a lighting system. A VPH-10S2S ventilator was used for the toilet and kitchen in each story, and the number of air changes per hour (ac/h) for this ventilator was 10.29.

2.4. Building Simulation

In order to minimize the energy consumption of the building, it is important to optimize the shell. The aim of using passive strategies was to obtain the measures that would deliver the optimized annual cooling load, annual heating load, and investment costs. It was examined to what extent the passive measures could reduce the annual cooling and heating load through these retrofit measures. For the assessment, the building was simulated with the DesignBuilder software using the heating and cooling system. The simulated model, which was carried out by DesignBuilder, is shown in Figure 8.

2.5. Building Calibration and Validation

The next step was to guarantee the correctness of the model and results, so in this research, the energy consumption data (from June 2018 to December 2019) were collected, and the building was monitored over this period of time. The building energy model was calibrated. Verification was performed after the initial simulation, using the measured energy consumption in 2018–2019. The initial simulation results were analyzed and compared with the monthly energy cost (bills) survey data as the verification reference. Adjustments were carried out if a comparison showed a higher initial simulation result value. Then, the simulation was reimplemented, and the monthly energy cost results were compared with the survey data.

To explain that a model is calibrated, the mean bias error (MBE) and the coefficient of variation of the root mean square error [CV (RMSE)] must be calculated [28]. Mean bias error (MBE) is primarily used to estimate the average bias in the model and to decide if any steps need to be taken to correct the model bias. Mean bias error (MBE) captures the average bias in the prediction and is calculated as:

$$MBE(\%)=\frac{{\Sigma }_{\mathit{Period}}{(S - M)}_{\mathit{Interval}}}{{\Sigma }_{\mathit{Period}}{(M)}_{\mathit{Interval}}}\times (100),$$

(1)

where M is the measured kilowatt-hours or fuel consumption during the time interval, and S is the simulated kilowatt-hours or fuel consumption during the same time interval.

Root mean square error (RMSE) is the square root of the mean of the square of all of the error. The use of RMSE is very common, and it is considered as an excellent general-purpose error metric for numerical predictions. The degree of accuracy of each model was evaluated using RMSE, which is a good indicator to represent the quality of the model and its ability to describe the real behavior of each system. RMSE is given by:

$${RMSE}_{Period}=\sqrt{{\displaystyle \sum }\frac{{{\left(S-M\right)}_{Interval}}^2}{N_{Interval}}},$$

(2)

where (N)_Interval is the number of time intervals in the monitoring period. The mean of the measured data for the period is calculated given by the equation below:

$$A_{Period}=\frac{{\Sigma }_{\mathit{Period}} M_{Interval}}{N_{Interval}}$$

(3)

The $C_{V\left(RMS{E}_{Period}\right)}$ is calculated using the following equation:

$$C_{V\left(RMS{E}_{Period}\right)}=\frac{{\mathit{RMSE}}_{\mathit{Period}}}{A_{Period}} \times \left(100\right)$$

(4)

The results showed that the real energy consumption was highly matched to the simulated model, which was simulated using the IWEC TMY2 weather file from the Energy Plus website. The calculated calibration results showed that the MBE was 4.1% for power consumption and 4.7% for gas consumption. In addition, the RMSE was 14.27% for electricity consumption and 14.81% for gas consumption. The results of the calibration met the ASHRAE guideline 14-2014 requirement, which suggests that MBE and RMSE should be under the limit values of 5% and 15%, respectively.

2.6. Optimization Methods and Algorithms

The design of buildings is a multicriteria optimization problem, where a compromise must always be made between capital expenditure, operating costs, and the thermal comfort of the occupants. Due to a large number of retrofit measures, it is important to make decisions about the choice of the right lens. More recently, researchers have used multiobjective criteria techniques to overcome this type of decision making. Multicriteria decision-making methods (further MCDM) are widely used in various fields of science. MCDM is very useful when evaluating alternatives according to different criteria. Multicriteria decision-making problems can be described by a set of criteria. In construction projects, decision making is based on interdependent criteria and even on both qualitative and quantitative criteria. Criteria may have different units of measurement and a different optimization direction. Gero et al. were among the first to implement a multicriteria model in the building design process to investigate the relationship between building thermal output, capital costs, and the usable area of the building [29]. Antucheviciene et al. (2010) applied the TOPSIS (Technique for the Order Preference by Similarity to Ideal Solution) method and integrated the Mahalanobis distance while analyzing building redevelopment in Lithuanian rural areas. If opposing goals have to be fulfilled at the same time, a single goal function cannot fulfill the requirements for describing the problem, and multicriteria procedures arise. The multitarget optimization method is used to optimize the base model. Finding an optimal building model, adapting it to the harsh climatic conditions of the future, and evaluating its resilience and reduction is the main goal of this study. Optimization algorithms have a major impact on increasing efficiency. Several numerical experiments have shown that metaheuristic search methods should be the primary ones [30,31,32,33,34]. Deb et al. suggested a non-dominated and crowding distance sorting genetic algorithm (NSGA-II) to solve multiobjective optimization problems [35]. NSGA-II is one of the widely used multiobjective evolutionary algorithms [35,36]. To achieve the optimization purpose, the NSGA-II algorithm was selected since it was capable of achieving the best convergence among other algorithms [30].

After that, the simulated model was optimized using the jEPlus tool, which was used to interact consistently with EnergyPlus. The jEPlus+EA tool, moreover, was used for determining the Pareto front. In this regard, 300 passive measures were optimized considering over 100 generations. The rate of gene mutation in the population was set at 0.2%, and each generation had 50 different individuals; the crossover probability was 1%. Since the case was located in climate B3, heating consumption was more important than cooling consumption.

2.6.1. Choosing Objectives

The proposed decision model focused on the retrofit phase of the buildings. Three objectives were chosen in this study. Achieving environmental and economic objectives as well as social objectives is the main purpose of this study. Sustainable architecture aims to minimize the negative environmental impact of buildings, and is designed to limit humanity’s impact on the environment. Some of the characteristics of sustainable architecture are as follows:

• Overall focus on reducing the human environmental impact on the environment;
• Buildings that produce at least as much energy as they consume for a net-zero effect; and
• Limiting wasteful energy consumption with the use of renewable energy sources such as solar panels and natural heating, cooling, and ventilation systems.

Minimizing investment costs is considered an economic objective. Due to the inflation rate, currency fluctuations, and low energy cost, this study addressed just the investment cost, including providing the required material for retrofitting the building. Minimizing both heating and cooling load as well as thermal comfort were considered as environmental and social goals, respectively.

2.6.2. Passive Measures

The passive measure evaluation was used to obtain information for choosing the appropriate combination of passive measures, following energy consumption and economic criteria. The selected passive measures aimed to find equilibrium between achieving a minimum cooling and heating load with the minimum investment cost. The best passive measure was chosen from the different packages of passive measures base on the TOPSIS method. In this method cooling load, heating load, and construction cost were considered as decision criteria.

For finding the optimal solutions among the current options presented in the Pareto frontier, jEplus was used as the main tool. First, the weather files, the RVI/MVI files, and the model template file, which was exported from Design builder and launched using EnergyPlus, was prepared as the contents of a jEPlus project. Next, the parameter tree was made by defining different parameters and their values. Window glazing, insulation thickness, and insulation type were regarded as the main parameters. The insulation thickness varied between 0.001 m and 0.01 m with 0.001 m steps. The selected options are shown in

Table 2. The selected passive retrofitting options.

	Type	Cost (USD)
Glazing type	XDblClr 6 mm/13 mm Arg DblClr 3 mm/13 mm Arg	7.04 6.2
	DblClr 3 mm/13 mm Air XDblClr 6 mm/13 mm Air	4.4 4.92
	DblClr 3 mm/6 mm Air DblClr 6 mm/6 mm Air	2.4 3.2
Insulation material	EPS XPS Rock wool-92 Rock wool-100 PVC Glass wool	4.8 5.2 2.4 2.6 3.2 0.68

. Three objectives were chosen in this study. Achieving the environmental and economic objectives of the sustainable building regarding the social objectives was the main purpose of this study. Minimizing investment costs was considered an economical objective. Due to the inflation rate, currency fluctuations, and low energy cost, this study addressed investment cost, including providing the required material for retrofitting the building. Minimizing both heating and cooling load were considered environmental goals. The average pricing data were obtained by asking sellers, or searching websites, which is based on USD.

At the following stage, this project was used in jEplus+EA, which was based on the NSGA-II algorithm, as an input. In addition to that, the population size, crossover rate, and mutation rate were defined. In this regard, 300 passive measures were optimized considering over 100 generations. The rate of gene mutation in the population was set at 0.2% and each generation had 50 different individuals; the crossover probability was 1%. The whole simulation took about 48 h, and after 25 simulations it was close to convergence.

Making good decisions is important for finding the best option. In almost all such problems, with the multiplicity of criteria for judging, the decision maker aims to attain more than one objective or goal, while satisfying various constraints related to environment, processes, and resources. In this study, the TOPSIS method was used to solve a multiple objective decision-making problem. The Technique for Order of Preference by Similarity to Ideal Solution is a multicriteria decision analysis method, which is commonly used to compare a set of alternatives by identifying weights for each criterion, normalizing scores for each criterion, and calculating the geometric distance between each alternative and the ideal alternative, which is the best score in each criterion. The alternative should have the furthest distance from the negative-ideal solution and the shortest distance from the ideal solution. After simulation, the selected retrofit solutions were assessed with the TOPSIS method. This method was used to indicate the best optimized solution among the ones in the Pareto front. The whole calculation related to the TOPSIS method was carried out in Microsoft Excel. Investment cost, annual heating, and cooling load were selected as the three objectives for present weather optimization, while total site energy, PMV, and investment cost were selected as the other three objectives for future weather files.

PMV is an index that aims to predict the mean value of votes of a group of occupants on a seven-point thermal sensation scale. The reason for selecting this parameter in the second stage was that the selected options in the first stage (passive measures) could not meet the PMV limitation of −5 and 5. Therefore, this study addressed the thermal comfort problem as the second step to meet all the occupants’ needs.

Multiple iterations based on the three main categories (window glazing type, insulation material, and insulation thickness) were made. Additionally, 600 simulations were performed considering the present weather file and the future one. An algorithm was implemented on the model using jEPlus+EA to find the best combination of building parameters that provide the minimum energy demand and minimum cost for each scenario. The optimization process included two phases and was carried out on a 16-core PC that could run 16 simulations at the same time, using around 20 per weather file. First, the algorithm found the best possible combination of both current and future energy demand considering investment cost, cooling, and heating load consumption.

2.6.3. Active Measures

Heating, Ventilation, and Air Conditioning (HVAC) systems consume a higher percentage of the total energy use, and are used to ensure the thermal comfort and safety of occupants. Heating and air conditioning components help to control the indoor climate and airflow, ensuring that the occupants do not freeze or sweat. HVAC systems have the potential to use more than 40% of the total electricity consumed in any building. Four different HVAC systems were considered in this study. For this purpose, all the HVAC systems were designed in DesignBuilder using detailed HVAC. This option enabled us to connect these HVAC systems to solar panels, and use solar energy to achieve a nearly net-zero energy building. The best solution was selected based on total energy consumption, thermal comfort, and investment cost. The TOPSIS method was used for selecting the best solution. The average pricing data were obtained by asking sellers, or searching websites, as shown in

Table 3. The selected active retrofitting options.

HVAC Type	Cost (USD)
FCU 4-pipe, Air-cooled Chiller	4600
District Heating and Cooling FCU 4-pipe	4600
Zone Water to Air Heat Pump Supplied by Ground Heat Exchanger	6500
VAV Reheat, Air-cooled Chiller	4600

, and is based on USD.

2.6.4. Nearly Net Zero Energy Strategies

After implementing both active and passive retrofit measures, the consumed energy decreased considerably. As a result, the model was converted to a nearly net zero energy building with a lower price. This roof was split into three different areas, as shown in Figure 9.

All of the calculations were performed using DesignBuilder and EnergyPlus. To make a net-zero energy building, PV panels were installed based on the DesignBuilder’s assumption. Considering PV panels with 1.939 m² each and 18% efficiency, it was possible to install 107 panels in area 1 oriented to the south, leading to an installed area of 207.473 m², 54 panels in area 2 oriented to the south, which provided 104.706 m², and 15 panels in area 3 oriented to the south, which provided 29.085 m². These panels were installed every 2.65 m to avoid shading on other panels. The suitable spaces between panels were chosen, as shown in Figure 10, based on the selected orientation. The designed PV system was simulated using DesignBuilder.

To achieve the best performance from a southern orientation, the PV was sloped by 45 degrees. After a detailed energy audit, the building was modeled and simulated under the EnergyPlus environment using Design Builder.

3. Results

3.1. Climate Variation

Of all climatic variables, the focus was on global horizontal solar radiation, relative humidity, and dry bulb temperature, as these elements can be effective in terms of thermal comfort, energy consumption, and the design of PV panels [37]. A comparison was then made between the base scenario and the 2080 scenario, which clarified the change in the abovementioned variables. This study mainly took into account the predictions of outdoor temperature and solar radiation. Achieving the most energy savings and the least energy consumption for installing solar panels, local shading for four main directions, and lighting control were added to the model. Solar radiations were extracted by the METEONORM software 8.0.4 and are shown in Figure 11. Then the model was ready to install solar panels. In the second phase, panels, which provide 341.264 square meters, were installed on the roof. The panels had a 45-degree slope to achieve the best performance.

3.2. Performance of Optimized Building under Different Weather Files

Changing the active heating or cooling systems for minimizing energy consumption should be proceeded by passive optimization, as this can lead to a smaller load. The results of the passive analysis for present and future weather files are shown in Figure 9. Obtaining the measures or combination of measures that provide the optimized annual cooling load, annual heating load, and the investment cost are the objectives of applying passive strategies. It was explored to what extent the passive measures were able to reduce the annual cooling and heating load through these retrofit measures. For the evaluation, the building was simulated by Design Builder software considering the heating and cooling system. After that, the simulated model was optimized using the jEPlus tool to determine the Pareto front. Since the case is located in climate B3, heating consumption is more important than cooling consumption.

To develop the passive measure assessment, the building simulation was carried out for the base case in order to know the actual situation, and the assessed parameters of heating load, cooling load, and investment costs were evaluated for the various passive measure packages.

All of the graphs in Figure 12 represent the results, and each point represents one simulation. The color of the points shows various measures of the Pareto frontier based on the relevant criteria (present and future weather files). The base model is also depicted in these graphs to provide the possibility of comparison.

3.3. Passive Measures Selection

The passive assessment of measures was used to obtain information for the selection of the suitable combination of passive measures according to energy consumption and economic criteria. The selected passive measures tried to find an equilibrium between achieving a minimum cooling and heating load with the minimum investment cost. The best passive measure was chosen from different packages of passive measures based on the TOPSIS method. In this method, cooling load, heating load, and construction cost were considered as decision criteria. The same priority was assumed for each criterion. The selected items are shown in

Table 4. The selected measures based on the present weather file.

Glazing Type	Insulation Type	Insulation Thickness
DblClr 3 mm/13 mm Arg	Rockwool-100	0.1 m

.

The selected passive measures caused 11.83% reduction in annual gas consumption and over 1.15% cooling reduction in comparison with the base model. The investment cost for achieving this situation was USD 2412.71. To choose the most appropriate passive measures, the best solution for 2080 was chosen from different packages of passive measures base on the TOPSIS method and the Pareto front. The selected items are shown in

Table 5. The selected measures based on the 2080 weather file.

Glazing Type	Insulation Type	Insulation Thickness
DblClr 6 mm/13 mm Air	PVC	0.1 m

.

The selected passive measure caused over 40.4% reduction in annual gas consumption and a 4.84% increase in cooling load in comparison with the base model. The investment cost for achieving this situation was USD 2817.02.

3.4. Active Measures Selection

The annual PMV and PPD of a building after the implementation of passive measures were about 0.52 and 23.41%, respectively (the PPD threshold of a building in thermal comfort mode is 20%). This means that building users would be dissatisfied 717.08 h of the year. As a result, taking active measures can improve the thermal comfort of building users.

Table 6. HVAC systems used in current weather conditions.

Solution Priority	HVAC Type	Total Site Energy (kWh)	PMV	Cost (USD)
1	FCU 4-pipe, Air-cooled Chiller	38,032.88	−0.26	4600
2	District Heating and Cooling FCU 4-pipe	58,222.83	−0.26	4600
3	Zone Water to Air Heat Pump Supplied by Ground Heat Exchanger	39,817.28	−0.299	6500
4	VAV Reheat, Air-cooled Chiller	66,857.37	−0.55	4600

and

Table 7. HVAC systems used in 2080 weather conditions.

Solution Priority	HVAC Type	Total Site Energy (kWh)	PMV	Cost (USD)
1	FCU 4-pipe, Air-cooled Chiller	31,410.8	0.05	4600
2	District Heating and Cooling FCU 4-pipe	54,773.91	0.05	4600
3	Zone Water to Air Heat Pump Supplied by Ground Heat Exchanger	101,230.75	0.007	6500
4	VAV Reheat, Air-cooled Chiller	55,574.07	−0.34	4600

show the proposed HVAC system for the current years and 2080, respectively. According to the tables, 4-pipe air-cooled chillers were selected for both climatic conditions based on the TOPSIS method. The best solution was selected based on total energy consumption, thermal comfort, and investment cost.

The optimized solution, which was an FCU 4-pipe, Air-cooled Chiller, could cause 65.14% saving of total energy consumption for the current weather file and 86.18% for the future weather file. This option met not only the energy consumption demands but the PMV criteria.

3.5. Net Zero Energy Strategies

The planned system guaranteed an annual energy production of 48.26 kWh/m², which corresponds to 90.66 percent of total site energy, which is equal to 48.26 kWh/m². A higher generation level would also be possible, as several other roof areas could be used. However, such a generation stage could only cover a total electricity requirement of 8.77 kWh/m². Therefore, 39.49 percent of the generated electricity would have to be injected into the grid. Despite the higher demand in the building during the higher generation hours, the building has low demand on the weekends and in July and August due to the summer holidays, which leads to a generation surplus during these periods.

4. Discussion

This study was conducted in three phases, two of them were active and passive measures optimization (three-objective optimization) and one was applying renewable energy to the studied case. In the first phase, the existing building was simulated in Design Builder and three measures (insulation material, insulation thickness, and window glazing type) were chosen to reduce building energy consumption. In this regard, the reduction in cooling load, heating load, and investment cost was considered as the three optimization objectives. EPS, XPS, Rockwool-100, Rockwool-92, and PVC were examined in this study. Considering the same priority for each objective, the selected solution for the present weather was Rockwool-100 (100 density) with 0.1 m thickness and DblClr 3 mm/13 mm Arg as the window glazing type. Rockwool-100 and XPS had almost the same performance, but Rockwool-100 was more cost-effective. Due to the high value of EPS and XPS, the Pareto frontiers were created by use of Rockwool-100, Rockwool-92, and PVC items. This passive solution with a USD 2412.71 investment cost, resulted in an 11.83% reduction in heating load and over 1.15% in cooling load. Although there were more energy-saving items, they had greater investment costs. The selected measures are shown in

Table 8. Selected optimized measures for the present and future weather files.

Solution Priority	Building Construction Cost (USD)	Annual Heating Gas (kWh)	Annual Cooling Electricity (kWh)
Base model	0	9081.63	42.5
Optimized present model	2412.71	8006.67	42.01
Optimized 2080 model	2817.02	5405.52	44.56

.

For making the best decision about the most optimized solution between the present and the future weather file, both optimized models were assessed in this study. The reduction in cooling load, heating load, and investment cost were considered as the three optimization objectives for the 2080 weather file for this purpose. The selected solution for the 2080 weather conditions was PVC with 0.1 m thickness and DblClr 6 mm/13 mm Air as the window glazing type. Although this solution suggested a larger investment cost, it resulted in higher energy savings. This passive solution with a USD 1977.81 investment cost, resulted in a 40.4% reduction in heating load and a 4.84% increase in cooling load.

Active measures that included different kinds of HVAC systems were investigated in the third phase. The best solution was selected based on total site energy consumption, HVAC system cost, and the annual PMV. The acceptable value for PMV is between −0.5 and 0.5. It can be seen that only the first, second, and third items met the allowable limit. Although the first and second items had the same amount of PMV and value, the first item was selected due to the lower annual energy consumption.

Achieving the most energy savings and the least energy consumption for installing solar panels, local shading for four main directions, and lighting control were added to the model. Then the model was ready for installing solar panels. In the second phase, 176 panels that provided 341.264 m² areas were installed on the roof. The panels had a 45-degree slope to achieve the best performance from the southern radiation. Although a 20% reduction in energy consumption was enough to provide a nearly net-zero energy building, these panels create over a 90% reduction in energy consumption.

After achieving the optimized model, a comparison between the proposed and ASHRAE baseline model was made. The optimized model was simulated and assessed for both present and future weather files. In this regard, the CCWorldWeatherGen tool was used to create 2080 weather data.

5. Conclusions

The substantial impact of climate change, as well as some basic construction properties, such as the type of window glazing and insulation, has been shown to impact the energy efficiency of educational buildings in the climate context of Tehran. The optimization of the insulation material and the window glazing in the renovation phase took into account a warmer climate, and this building could lead to a reduction in the annual energy requirement.

In the present weather file, a 43.38% reduction in comparison with the ASHRAE baseline model occurred due to the various retrofit measures. This means that the proposed building had more energy savings than the baseline building. The window shading and lighting control system were added automatically to the baseline model by DesignBuilder to make the comparison more reasonable. So, the baseline total energy decreased from 56,680 kWh to 55,473.8 kWh considering the existing building, and the total energy consumption reduced from 109,129.3 kWh to 31,410.8 kWh, which represents an over 247% reduction.

To consider the impact of 2080 weather data, these models were simulated under new conditions. As can be seen, the present optimized model (DblClr3 mm/13 mm Arg, Rockwool-100, 0.1 m) can achieve better energy savings in both the present and the future weather. The selected option will achieve a 52% reduction in the future and a 49% reduction in the present in comparison with the ASHRAE baseline. Due to the lower energy consumption, the nZEB will perform as an NZEB after implementing active measures especially using the future file. The current optimized model would not remain optimized in the long run; thus, it is important to propose a different optimized model based on the building’s needs and lifespan. Considering the future weather file is essential in this regard. Applying different retrofit measures such as changing setpoint temperature and using sustainable retrofit technologies such as a green roof and Trombe wall is suggested for the future.