Introduction
An energy-efficient building is called a structure that provides minimum carbon emission by using energy effectively consumed for heating, cooling, air conditioning, and lighting. The most important stage in the construction processes of energy-efficient buildings is the design step. The building needs less energy through the measures taken and the decisions made during the design phase.
The most basic strategy for less energy consumption and carbon emissions is to design an energy-efficient building envelope. The energy-efficient building envelope consists of components that provide thermal insulation and indoor comfort. However, thermal insulation systems are not preferred in developing countries such as Turkey because of the higher cost of purchase and installation of insulation (1). Therefore, the building (residential/commercial) sectors in Turkey need to spend a lot of money on heat suppression every year (2). In addition, the most basic factor affecting the choice of materials and equipment in energy-efficient buildings is the location of the building. Since it is not appropriate to use the same building envelope designs in different climatic zones, climatic characteristics also have an impact on the heating and cooling energy consumption and carbon emissions of buildings. In addition to the climatic characteristics, the earthquake characteristics of the regions where the buildings will be constructed should also be taken into account in the thermal energy consumption. Since, depending on the specific geographical region where it is situated, a building can potentially be attacked by being exposed to natural hazards such as earthquakes, it may suffer from different levels of structural deformation (3). This situation may cause the thermal balance of the indoor environment to be lost due to deterioration in the building envelope. Naturally, more energy consumption and more carbon emissions will be required in order to restore the thermal balance. In order to prevent more energy consumption, possible repair costs of the building envelope should also be considered, taking into account the possibility of an earthquake.
Countries that are developing and in need of energy imports, such as Turkey, must make different laws and regulations to reduce their energy consumption. Moreover, in a country with different climate and earthquake zones such as Turkey again, heating and cooling energy consumption cannot be the same in every zone. Therefore, in recent years, energy-efficient building designs have become one of the most important strategic activities of governments that need energy import.
Once the studies in the literature are reviewed, the lifecycle cost assessment (LCA) is generally used for energy efficient building designs. The cost parameters that occur during the economic life of the buildings are taken into account through the LCA approach. Accordingly, different parameters including energy consumption, material cost, and environmental impacts are included in LCA. In addition, numerical calculations, simulation programs, or artificial intelligence approaches are preferred to obtain LCA parameters.
Caglayan et al. (4) analyzed the heating energy consumption and material costs for four different climate zones in Turkey, taking into account the window type, wall insulation, ceiling insulation, and basement floor insulation material thicknesses. They developed an optimization tool by using numerical calculation formulas in a genetic algorithm (GA). In addition, they performed a sensitivity analysis for window, wall, ceiling, and basement insulation materials. Himmetoglu et al. (5) applied attribute reduction to obtain the climate characteristics affecting the heating and cooling energy consumption of a public building for two different climate zones in Turkey. They also proposed a structure called PSACONN mining to determine the most suitable building envelopes that give the minimum heating and cooling energy consumption. Acar et al. (6) took into account the orientation, wall insulation material, roof insulation material, glazing type, and window thickness for the residential buildings in Turkey. They used the EnergyPlus simulation program and non-dominated sorting genetic algorithm-II (NSGA-II) together to analyze building envelope alternatives that minimize the total thermal energy demand and investment cost. Delgarm et al. (7) analyzed envelope alternatives that minimize heating, cooling, and lighting energy consumption for four different climate zones, by considering shading specifications, window size, glazing, and wall material. They aimed to scan the entire solution space in a shorter time by developing an EnergyPlus-based particle swarm optimization (PSO) algorithm. Chantrelle et al. (7) considered exterior wall type, roof type, ground floor type, intermediate floor type, partition wall type, and window type as decision variables. They analyzed energy consumption, thermal comfort, investment costs, and environmental impacts using the TRNSYS simulation program and the NSGA-II approach together. Karmellos et al. (8) aimed to determine the most suitable building envelope combination that minimizes annual energy consumption and investment costs for two different climate zones. They have developed a MATLAB-based tool for decision-makers, taking into account door type, window type, wall type, energy systems, lighting systems, and electrical appliances. Echenagucia et al. (9) aimed to minimize the heating, cooling, and lighting energy demands for different climatic regions by considering the number of windows, window position, window shape, window type, wall thickness, and glazing. They used EnergyPlus and NSGA-II together. Gossard et al. (10) analyzed the annual energy consumption and comfort levels for two different climate zones, taking into account the thermo physical properties of the external wall. They proposed an approach including TRNSYS and NSGA-II approaches, taking into account thermal conductivity and volumetric specific heat for the wall and the roof as decision variables. Ascione et al. (11) proposed an approach that minimizes the percentage of heating/cooling energy demand and thermal discomfort hours for two cities with the same climate features. They considered window type, insulation thickness, wall density, solar absorptance, and thermal emissivity by using the EnergyPlus simulation program integrated into the NSGA-II approach. Wang and Wei (12) analyzed building envelope designs that minimize building energy loads and construction costs for tropical and subtropical climate zones. By integrating numerical calculations into quantum GA, they used the wall material, roof material, window sizes, glazing, window shading, orientation, and the number of windows as decision variables. Albatayneh (13) aimed to minimize heating and cooling loads to provide thermal comfort by using EnergyPlus and GA together for a climate zone. A sensitivity analysis was performed by using regression analysis for the decision variables of orientation, wall insulation thickness, roof insulation material, partition construction, window/wall ratio, window type, window shading, glazing type, infiltration rate, and natural ventilation rate. Chegari et al. (14) aimed to minimize heating and cooling energy consumption by using TRNSYS, NSGA-II, and ANN approaches together. They considered exterior wall materials, roof materials, window materials, glazing, shading, and air changing as the decision variables. Bre et al. (15) analyzed heating and cooling performances for decision variables of roof type, exterior wall type, interior wall type, solar orientation, solar absorptance, window size, window type, window shading, and infiltration rate by using EnergyPlus, the artificial neural network (ANN), and NSGA-II together. Huang et al. (16) analyzed the heating energy consumption by proposing a mathematical model using numerical formulations. Insulation thicknesses, orientation, window/wall ratio, and window type were taken into account as decision variables. Lu et al. (17) applied LCA through EnergyPlus by considering heating energy consumption, cooling energy consumption, CO2 emission, material cost, and heat transfer coefficient. Window type, wall insulation type, roof insulation type, and insulation thickness were used as decision variables. Yuan et al. (18) proposed an LCA approach to minimize material cost, heating energy consumption, and cooling energy consumption using numerical calculations. They considered doors, windows, exterior walls, partition walls, and roof materials as decision variables. Lin et al. (19) used the NSGA-II approach to minimize the building envelope costs and CO2 emissions. They considered wall material, roof material, glass curtain material, window size, number of windows, number of glasses, window sunshade shape, window sunshade type, and six different air conditioning parameters as the decision variables. Kim et al. (20) applied feature subset selection with the C4.5 decision tree method, taking into account parameters such as material type, insulation thickness, and air gap for walls and roofs. For more detailed literature on envelope design and material analysis in energy-efficient buildings, the review paper published by Kheiri (21) may be reviewed. To the best of our knowledge, there is only one study evaluating the effects of earthquakes on energy-efficient building envelope materials. In the study presented by Liu and Mi (22), damages that occur only on the windows due to earthquakes (drift rate) are taken into account along with the thermal energy consumption of the building, CO2 emission, and material cost.
This study has a wider perspective than the above studies in terms of its scope. The most important contribution of this study is to consider the effects of climate and earthquake zones in energy-efficient building designs, as well as the importance levels of the window, exterior plaster, insulation (wall and roof), wall, and interior plaster materials used in the building envelope. The second important contribution is to take into account the heating energy consumption, cooling energy consumption, building carbon emissions, embodied carbon, material costs, and repair costs caused by the earthquake effect. Accordingly, in the first step of the study, the parameters that most affect heating, cooling, and CO2 emissions from the climate and earthquake zone parameters along with the building envelope attributes are determined. In the second step of the study, an LCA analysis including heating energy consumption, cooling energy consumption, CO2 emission, embodied carbon, material cost, and earthquake repair cost is proposed. ANN models based on the EnergyPlus simulation program are developed to predict heating, cooling, and CO2 emissions. The proposed approach was performed in a small-sized case study.
This study is organized as follows. In Section “Research Elaborations,” the approaches used for the proposed methodology are presented. In Section “Results and findings for a case study,” a case study is presented. Conclusion are given in section “Conclusion.”
Research elaborations
The proposed methodology consists of three main steps. In the first step, the feature subset selection is performed in order to analyze the features affecting the heating energy consumption, cooling energy consumption, and CO2 emission. In the second step, the predictive models are developed in order to separately forecast heating energy consumption, cooling energy consumption, and CO2 emission according to the features obtained in Step 1. In the last step, an LCA analysis including the material cost, embodied carbon, and seismic repair cost along with the parameter values estimated in Step 2 is performed.
Feature subset selection
The proposed approach analyzes the effects of building envelope material features and regional characteristics on heating energy consumption, cooling energy consumption, and CO2 emission. The importance and effect of regional characteristics and material characteristics may not be the same for the mentioned parameters. Therefore, the decision variables affecting each parameter should be evaluated separately.
In this step, “the correlation-based feature subset selection algorithm for machine learning” (CfsSubsetEval) approach proposed by Hall (23) was preferred for the feature selection process. The CfsSubsetEval is an approach that evaluates the worth of a subset of attributes by considering the individual predictive ability of each feature along with the degree of redundancy between them (23). The random forest algorithm presented by Breiman (24) was used to analyze the accuracy of the CfsSubsetEval approach. Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest (24).
In the proposed approach, applying the feature subset selection to the decision variables for heating energy consumption, cooling energy consumption, and CO2 emission parameters aims to make the estimation structure work better. In addition, it will enable the discovery of hidden patterns between the mentioned parameters and decision variables.
The predictive modeling
Energy simulation programs are useful for the energy analysis of buildings in the design phase (5). A wide variety of building simulation programs have been developed (25). Examples of these programs are BLAST, EnergyPlus, eQUEST, TRACE, DOE2, and ECOTECT (20). Crawley (26), Sadineni (27), and Mirsadeghi (28) present a detailed review of building simulation programs used in the literature.
EnergyPlus is a well-known, free software program developed by the U.S. Department of Energy that can be used to perform whole-building energy analysis (25). It has two main libraries that are component and template libraries. It also can analyze many features, from material alternatives to hourly air values, from metabolic rates to storage. It provides convenience to decision makers in the design processes in every region through the detailed climate files of almost every region of the world. However, no single energy simulation program offers sufficient capabilities and flexibilities to analyze integrated building systems and to enable rapid prototyping of innovative building and system technologies (29). Once the number of alternatives increases in the building design process, the time to enter data into EnergyPlus and the time to evaluate the results increases. Therefore, there is a need for effective and practical approaches that mimic the working mechanism of EnergyPlus. Using simulation programs along with artificial intelligence techniques will increase the efficiency of building design processes. In this study, ANN models, which learn the working structure of the EnergyPlus simulation program, are developed.
Artificial neural network models consist of six basic elements. These are layers, weights, neurons, network structure, training algorithm, and transfer functions. ANN models generally consist of three layers: input, hidden, and output layers. Some neurons hold information in each layer. Each neuron in each layer has a certain weight value. According to this weight value, the values of the output neurons vary. Determining the neuron weights is the most important step in ANN models. Weight calculation processes are carried out through training algorithms. Information transmission between the layers is provided by transfer functions. The network structure of ANN models determines the form of information transmission. Detailed technical information and the basic concept of an ANN can be found in Refs. (30–32).
In order to obtain EnergyPlus-based ANN models, a sample input data set representing the whole alternative solution space is generated by considering the different values of the decision variables. According to this data set, heating and cooling energy consumptions and CO2 emissions are calculated with the EnergyPlus program, and a sample output data set is obtained. ANNmodels are developed separately for heating energy consumption, cooling energy consumption, and CO2 emission by using input and output sample data sets. In the study, EnergyPlus based ANN models are proposed for cases where there are intensive calculations and many alternatives.
Life cycle cost assessment (LCA) analysis
Life-cycle cost assessment is an economic analysis technique that takes into account the investment cost and the periodic (e.g., monthly and annual) costs that will occur during the economic life of the building for the project management processes. It is very effective in making the most appropriate decision, especially for energy-efficient building designs, taking into account the investment costs and annual costs.
In this study, heating energy consumption, cooling energy consumption, CO2 emission, embodied carbon, material cost, and earthquake repair cost are considered as parameters of the LCA. Although prediction models are created for heating energy consumption, cooling energy consumption, and CO2 emissions, there is no need to create prediction models for embodied carbon, material cost, and earthquake repair cost. Since the embodied carbon, material cost, and earthquake repair cost are material- oriented, there is no need for estimation since the related parameter values can be calculated directly with a simple calculation.
Results and findings for a case study
Building definition
In this section, the proposed methodology was applied to a small-sized case study. A one-story structure was designed with residential building features for the case study. The building has an area of 25 × 25 m2. It also has four flats, two elevators, two warehouses, a staircase, and a fire escape. The flats are symmetrical. Each flat consists of four living rooms, a bathroom, a kitchen, a toilet, and a hall. The materials used on the floor are sand-cement plaster (12.5 mm), polystyrene rigid foam (20 mm), reinforced concrete (150 mm), screed (50 mm), and ceramic (20 mm). Partition walls are in the form of brick (105 mm) and both- side plaster (12 mm). The building height is 3 m.
The building model designed in DesignBuilder, which is an interface program that provides data entry to EnergyPlus, is shown in Figure 1. The case study building was designed to be cooled to 24°C when the temperature rises above 28°C and be heated to 22°C when the temperature drops below 18°C.
Figure 1. DesignBuilder single-story building model (not scaled).
In the proposed approach, applying the feature subset selection to the decision variables for heating energy consumption, cooling energy consumption, and CO2 emission parameters aims to make the estimation structure work better. In addition, it will enable the discovery of hidden patterns between the mentioned parameters and decision variables.
Feature subset selection
In the case study, two main input attributes are taken into account. These are building envelope variables and regional decision variables. The building envelope decision variables are interior plaster, masonry material, masonry insulation material, exterior plaster, roof insulation material, the thickness of these materials, window type, and window/wall rate. For regional decision variables, climatic zones and earthquake zones are taken into account. The effects of these decision variables on heating energy consumption, cooling energy consumption, and CO2 emission are analyzed and hidden patterns are investigated. The alternatives used for the building envelope decision variables are presented in Table 1. For regional decision variables, four different climate and earthquake zones in Turkey are taken into account. Four different regions are selected to represent each climate and earthquake zone in Turkey. General information about the pilot regions is given in Table 2.
Table 1. Building envelope decision variables for the case study.
Table 2. The feature of the pilot regions.
It is applied to the building in the case study by generating a hundred different combinations covering each alternative in each decision variable. That is, heating energy consumption, cooling energy consumption, and CO2 emission values are obtained through EnergyPlus according to a hundred different combinations, considering the alternatives and pilot regions selected from the building envelope decision variables. There are more than ten million alternative combinations in total, including pilot regions. While determining a hundred different combinations, alternatives representing all spaces should be determined, taking into account the worst and best scenarios. Once the number of alternatives is increased, the time spent on EnergyPlus will increase. Therefore, its number should be kept at a reasonable level.
According to the results obtained with EnergyPlus, the most appropriate input decision variables are determined for each related parameter by applying the CfsSubsetE- val approach. A comparison of results with and without CfsSubsetEval is shown in Table 3. Random Forest is applied to compare the results. The decision variables determined as a result of the tree structures obtained are shown in Table 4. Decision variables obtained as a result of feature subset selection will be used to develop ANN models.
Table 3. The feature of the pilot regions.
Table 4. Reduced decision variables with feature subset selection.
EnergyPlus-based ANN modeling
The number of attributes is reduced by means of the feature subset selection. Thus, it is possible to generate simpler models with fewer inputs for ANN models. Notably, since the effect of the decision variable on each input parameter may not be the same, the same input decision variables are not taken into account in ANN models. Therefore, instead of generating a single ANN model, three different models are produced for heating energy consumption, cooling energy consumption, and CO2 emission.
The model structures
According to Table 4, five, four, and seven input neurons are used for heating energy consumption, cooling energy consumption, and CO2 emission, respectively. For the ANN training processes, a hundred-data set determined in the previous step is used. For each ANN model, feedforward MLP is used as the network structure. As the training algorithm, the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm is preferred. A single hidden layer is used. Linear, logistic, tanh, and exponential functions are considered for the transfer functions. The number of neurons in the hidden layer is relaxed between 3–11, 3–10, and 4–12 for heating energy consumption, cooling energy consumption, and CO2 emission, respectively. In order to determine the best ANN model for each output parameter, the training algorithm is run 1,000 times using the STATISTICA64 package program. The properties of the ANN models generated by obtaining the most appropriate weights are shown in Table 5.
Table 5. The features of ANN models.
Sensitivity analysis
The significance levels of the decision variables for each ANN model are analyzed and given in Table 6. Here, it is observed that earthquake zones are as important as climatic zones. In fact, it is interesting to see that earthquake zones are more important than insulation criteria.
Table 6a. The sensitivity analysis results of the decision variables for the heating energy consumption.
Table 6b. The sensitivity analysis results of the decision variables for the cooling energy consumption.
Table 6c. The sensitivity analysis results of the decision variables for the CO2 emission.
The building envelope LCA analysis
In this step of the study, the LCA approach is applied for energy-efficient building envelope designs. The LCA parameters are the cost of materials used in the building envelope, earthquake-based repair cost, embodied carbon, heating energy consumption, cooling energy consumption, and CO2 emission for the case study. Since the analysis of many alternatives and criteria with energy simulation programs is time-consuming, ANN models are developed. In contrast, since the cost of materials, earthquake-based repair cost, and embodied carbon values are material- oriented and can be easily calculated, there is no need to develop an estimation model. The material cost consists of building envelope and window costs. Exterior plaster, insulation, masonry, and interior plaster materials affect the cost of the building envelope. Since window/wall is a decision variable, the building envelope and window surface area are not constant (Table 1). For 30, 35, and 40% window/wall ratios, the building envelope surface area is 210 m2, 195 m2, and 180 m2, and the window surface area is 90 m2, 105 m2, and 120 m2, respectively. Building envelope and window costs are investment costs. Other LCA parameters are costs incurred over the economic life cycle of the building. Since the LCA approach is cost-based, each parameter must be converted to cost. The embodied carbon is carbon emissions that occur during the entire life cycle of materials (from production to consumption). The earthquake-based repair cost depends on the material cost, the magnitude of the earthquake, and the probability of earthquakes that could damage the building envelope. If we take into account the risk of earthquakes for each year during the lifespan of the building, the earthquake-based repair cost should also be converted to present value. In addition, the costs of energy consumption, CO2 emissions, and embodied carbon are also converted to present value. Therefore, the present worth factor (PWF) should be calculated. Eqs. (1) and (2) are used to calculate the PWF. The LCA equation is given in Eq. (3). The parameters used for the LCA approach are shown in Table 7.
Table 7. LCA parameters.
Where i* is the interest rate adapted for inflation, N is the lifespan, i is the interest rate, and g is the inflation rate.
For the earthquake-based cost, four different earthquake zones are determined in Turkey. The earthquake zones are defined on the expected peak ground acceleration (PGA) for a return period of 475 years (10% exceedance in 50 years) (33). In 2018, Turkey’s earthquake zone map was renewed by Turkey Disaster and Emergency Management Authority (AFAD) (34). In the new map, the earthquake zones are separated according to the PGA values. In this study, for the PoE, the number of earthquakes with Mx > 4 between 1970 and 2020 (50 years) is taken into account in each region. Since it has been observed that earthquakes whose magnitudes are 4 and greater than 4 have damaged buildings in Turkey, Mx > 4 has been considered. The SRM is assumed as 20, 50, 80, and 100% for significant local damages of many components (4 < Mx < 5), extensive damage to many components (5 < Mx < 6), extensive widespread damages (6 < Mx < 7), and complete widespread damages (7 < Mx), respectively [81]. For the application, the probability of at least one earthquake occurring in a year is calculated. Since earthquakes commonly follow Poisson distribution (35), the earthquake probability for the magnitude j in each region is calculated with Eq. (4). Accordingly, for instance, the probability of at leastoneearthquakeis1-PoE1,Mx > 4(X = 0) = 0.8892for region 1.
Where e is Euler’s number. A represents the average (expected) number of earthquakes in a unit period. x is the number of earthquakes occurring in a unit period.
As a result, a thousand different alternatives were generated to calculate the accuracy of the results obtained with the LCA approach. A thousand different combinations were applied manually in EnergyPlus. Since it is almost impossible to manually enter all combinations into EnergyPlus, randomly selected thousand different combinations were evaluated. A summary of the results obtained is presented in Table 8.
Table 8. Summary table.
Conclusion
Energy-efficient building designs have an important strategic position for developing countries such as Turkey that need energy imports. The most important component of energy-efficient building designs is energy-efficient building envelopes. A broad perspective is presented that takes into account energy consumption, indoor comfort, and environmental effects. By applying LCA analysis, heating, cooling energy consumption, CO2 emission, material cost as well as embodied carbon, and earthquake-based cost are also taken into account. Interior plaster, wall insulation, roof insulation, wall, exterior plaster, material thickness, window, window/wall, climate, and earthquake zones are considered the decision variables. Significance levels of decision variables for heating energy consumption, cooling energy consumption, and CO2 emissions were determined. According to the results obtained, it is observed that the earthquake zones have a remarkable effect. In future studies, the scope of the study can be expanded by using metaheuristic approaches such as GA and PSO, which can scan the entire alternative solution space.
Author contributions
All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
References
1. Uygunoglu T, Kegebas A. LCC analysis for energy-saving in residential buildings with different types of construction masonry blocks. Energy Build. (2011) 43:2077–85.
2. Kecebas A, Kayveci M. Effect on optimum insulation thickness, cost and saving of storage design temperature in cold storage in Turkey. Energy Educ Sci Technol Part A. (2010) 25:117–27.
3. Venture SJ. State of the art report on systems performance of steel moment frames subject to earthquake ground shaking. Washington, DC: FEMA 355C (2000).
4. Caglayan S, Yigit S, Ozorhon B, Ozcan-Deniz G. A genetic algorithm-based envelope design optimisation for residential buildings. Proceedings of the Institution of Civil Engineers- Engineering Sustainability. (Vol. 173), (2020). p. 280–90.
5. Himmetoglu S, Delice Y, Aydogan EK. PSACONN mining algorithm for multi-factor thermal energy-efficient public building design. J Build Eng. (2021) 34:102020.
6. Acar U, Kaska O, Tokgoz N. Multi-objective optimization of building envelope components at the preliminary design stage for residential buildings in Turkey. J Build Eng. (2021) 42:102499.
7. Chantrelle FP, Lahmidi H, Keilholz W, El Mankibi M, Michel P. Development of a multicriteria tool for optimizing the renovation of buildings. Appl. Energy. (2011) 88:1386–94.
8. Karmellos M, Kiprakis A, Mavrotas G. A multi-objective approach for optimal prioritization of energy efficiency measures in buildings: model, software and case studies. Appl Energy. (2015) 139:131–50. doi: 10.1016/j.apenergy.2014.11.023
9. Echenagucia TM, Capozzoli A, Cascone Y, Sassone M. The early design stage of a building envelope: multi-objective search through heating, cooling and lighting energy performance analysis. Appl Energy. (2015) 154:577–91. doi: 10.1016/j.apenergy.2015.04.090
10. Gossard D, Lartigue B, Thellier F. Multi-objective optimization of a building envelope for thermal performance using genetic algorithms and artificial neural network. Energy Build. (2013) 67:253–60. doi: 10.1016/j.enbuild.2013.08.026
11. Ascione F, Bianco N, De Masi RF, Mauro GM, Vanoli GP. Design of the building envelope: a novel multi-objective approach for the optimization of energy performance and thermal comfort. Sustainability. (2015) 7:10809–36. doi: 10.3390/su70810809
12. Wang Y, Wei C. Design optimization of office building envelope based on quantum genetic algorithm for energy conservation. J Build Eng. (2021) 35:102048. doi: 10.1016/j.jobe.2020.102048
13. Albatayneh A. Optimisation of building envelope parameters in a semi-arid and warm Mediterranean climate zone. Energy Rep. (2021) 7:2081–93. doi: 10.1016/j.egyr.2021.04.011
14. Chegari B, Tabaa M, Simeu E, Moutaouakkil F, Medromi H. Multi-objective optimization of building energy performance and indoor thermal comfort by combining artificial neural networks and metaheuristic algorithms. Energy Build. (2021) 239:110839. doi: 10.1016/j.enbuild.2021.110839
15. Bre F, Roman N, Fachinotti VD. An efficient metamodel-based method to carry out multi-objective building performance optimizations. Energy Build. (2020) 206:109576. doi: 10.1016/j.enbuild.2019.109576
16. Huang J, Lv H, Gao T, Feng W, Chen Y, Zhou T. Thermal properties optimization of envelope in energy-saving renovation of existing public buildings. Energy Build. (2014) 75:504–10. doi: 10.1016/j.enbuild.2014.02.040
17. Lu S, Wang Z, Zhang T. Quantitative analysis and multi-index evaluation of the green building envelope performance in the cold area of China. Sustainability. (2020) 12:437. doi: 10.3390/su12010437
18. Yuan Z, Zhou J, Qiao Y, Zhang Y, Liu D, Zhu H. BIM-VE-based optimization of green building envelope from the perspective of both energy saving and life cycle cost. Sustainability. (2020) 12:7862. doi: 10.3390/su12197862
19. Lin YH, Lin MD, Tsai KT, Deng MJ, Ishii H. Multi-objective optimization design of green building envelopes and air conditioning systems for energy conservation and CO2 emission reduction. Sustain Cities Soc. (2021) 64:102555. doi: 10.1016/j.scs.2020.102555
20. Kim H, Stumpf A, Kim W. Analysis of an energy efficient building design through data mining approach. Autom Constr. (2011) 20:37–43. doi: 10.1016/j.autcon.2010.07.006
21. Kheiri F. A review on optimization methods applied in energy-efficient building geometry and envelope design. Renew Sustain Energy Rev. (2018) 92:897–920. doi: 10.1016/j.rser.2018.04.080
22. Liu MM, Mi B. Life cycle cost analysis of energy efficient buildings subjected to earthquakes. Energy Build. (2017) 154:581–9. doi: 10.1016/j.enbuild.2017.08.056
23. Hall MA. Correlation-based feature selection for machine learning, Ph.D. Thesis. Hamilton: The University of Waikato (1999).
25. Alonso MJ, Dols WS, Mathisen HM. Using Co-simulation between EnergyPlus and CONTAM to evaluate recirculation-based, demand-controlled ventilation strategies in an office building. Build Environ. (2022) 211:108737. doi: 10.1016/j.buildenv.2021.108737
26. Crawley DB, Hand JW, Kummert M, Griffith BT. Con-trasting the capabilities of building energy performance simulation programs. Build Environ. (2008) 43:661–73. doi: 10.1016/j.buildenv.2006.10.027
27. Sadineni SB, Madala S, Boehm RF. Passive building energy savings: a review of building envelope components. Renew Sustain Energy Rev. (2011) 15:3617–31. doi: 10.1016/j.rser.2011.07.014
28. Mirsadeghi M, Costola D, Blocken B, Hensen JL. Review of external convective heat transfer coefficient models in building energy simulation programs: implementation and uncertainty. Appl Therm Eng. (2013) 56:134–51. doi: 10.1016/j.applthermaleng.2013.03.003
29. Treka M, Hensen JL, Wetter M. Co-simulation of innovative integrated HVAC systems in buildings. J Build Perform Simul. (2009) 2:209–30. doi: 10.1080/19401490903051959
30. Bui DK, Nguyen TN, Ngo TD, Nguyen-Xuan H. An artificial neural network (ANN) expert system enhanced with the electromagnetism-based firefly algorithm (EFA) for predicting the energy consumption in buildings. Energy. (2020) 190:116370. doi: 10.1016/j.energy.2019.116370
31. Bui DK, Nguyen T, Chou JS, Nguyen-Xuan H, Ngo TD. A modified firefly algorithm-artificial neural network expert system for predicting compressive and tensile strength of high-performance concrete. Constr Build Mater. (2018) 180:320–33. doi: 10.1016/j.conbuildmat.2018.05.201
32. Nguyen T, Kashani A, Ngo T, Bordas S. Deep neural net-work with high-order neuron for the prediction of foamed concrete strength. Comput Aided Civil Infrastruct Eng. (2019) 34:316–32. doi: 10.1111/mice.12422
33. Pohoryles DA, Maduta C, Bournas DA, Kouris LA. Energy performance of existing residential buildings in Europe: a novel approach combining energy with seismic retrofitting. Energy Build. (2020) 223:110024. doi: 10.1016/j.enbuild.2020.110024
34. Turkey Building Earthquake. Regulation, Turkey disaster & emergency management authority. (2022). Available online at: https://www.resmigazete.gov.tr/eskiler/2018/03/20180318M1.pdf (accessed April 24, 2022).
35. Ture O, Qobanoglu I, Gul M, Karacan E. Seismic record approach for the evaluation of natural hazards: a key study from SW Anatolia/Turkey. Environ Earth Sci. (2021) 80:1–16. doi: 10.1007/s12665-021-09779-0
© The Author(s). 2024 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.