Datasets on Energy Simulations of Standard and Optimized Buildings under Current and Future Weather Conditions across Europe

Abstract

The building sector has a strategic role in the clean energy transition towards a fully decarbonized stock by mid-century. This data article investigates the use of different weather datasets in building energy simulations across Europe. It focuses on a standard performing building optimized to a nearly-zero level accounting for climate projections towards 2060. The provided data quantify the building energy performance in the current and future scenarios. The article investigates how heating and cooling loads change depending on the location and climate scenario. Hourly weather datasets frequently used in building energy simulations are analyzed to investigate how climatic conditions have changed over recent decades. The data give insight into the implications of the use of weather datasets on buildings in terms of energy consumption, efficiency measures (envelope, appliances, systems), costs, and renewable production. Due to the ongoing changing climate, basing building energy simulations and design optimization on obsolete weather data may produce inaccurate results and related building designs with an increased energy consumption in the coming decades. Energy efficiency will become more crucial in the future when cooling and overheating will have to be controlled with appropriate measures used in combination with renewable energy sources.

Datasets: D’Agostino, Delia; Parker, Danny; Epifani, Ilenia; Crawley, Dru; Lawrie, Linda (2022), Weather datasets and energy simulation files related to standard and optimized buildings across Europe, Mendeley Data, doi:10.17632/6krybfjfsg.1, Direct URL to data https://data.mendeley.com/drafts/6krybfjfsg (accessed on 14 April 2022).

Dataset License: CC-BY 4.0

1. Summary

The need for a clean energy transition reducing dependence on fossil fuels has become more urgent in the light of the recent geopolitical situation [1]. At the European level, key principles of the Green Deal are prioritizing energy efficiency, improving the energy performance of buildings, and developing a power sector based largely on renewable sources [2]. Globally, the agreed target for the coming decades is to decarbonize the energy sector, making buildings more efficient [3]. Within this framework, nearly zero energy buildings (NZEBs) have a strategic role in improving energy efficiency and renewables using a cost-optimal approach [4]. NZEBs are the mandatory target for new buildings since January 2021, as established in the Energy Performance of Building Directive [5].

A powerful way to investigate the design of NZEBs is to carry out building energy simulations with different goals, such as for the establishment of technologies [6], the prediction of energy and stock performance [7,8], the analysis of comfort and indoor air quality [9,10], the assessment of certification [11,12], and the estimation of future savings scenarios to test policy options [13,14].

Previous research by the authors determined the optimal design of NZEBs in different European climates using commonly available weather datasets [15,16]. This work showed how to reach the NZEB target with cost-optimality obtaining the best combination of efficiency measures, equipment, appliance, and renewables.

However, over the last decade, we have experienced usual and extreme weather events, especially summer heat waves [17]. At the boundary between the outdoors and indoors, it is certain that climate change will also affect buildings and occupants [18]. Nevertheless, not all the climate change effects are currently known and several possible effects need to be carefully investigated (e.g., changes in heating and cooling loads and consequent increased greenhouse gas emissions, shifts in thermal operational conditions, heating, ventilation, and air conditioning—HVAC—capacity mismatch).

Therefore, considering the climate change implications for buildings that must be properly built or renovated, it is crucial to take a forward-looking perspective [19]. Accordingly, recent research by the authors, to which the datasets of this paper refer, showed how the energy balance will be altered in European buildings in future. In more detail, heating will drop by 38–57%, while cooling will rise by +99–380% in the different tested locations [20]. Moreover, efficiency measures to reduce cooling and overheating will become more important (e.g., roof insulation, window type, solar shading, envelope finishes). Energy efficiency will have a strategic role within a changing climate scenario.

The data linked to this paper provide quantitative information on the design of NZEBs across Europe in past, present, and future climatic conditions projected to the year 2060. Currently used datasets in energy simulations can be misleading since past data drivers predict energy loads in the future. Hourly weather datasets that are frequently used in building energy simulations are provided for different locations.

The provided data account for future climate change in the simulations. These are weather morphed files for 2060 in the IPCC anticipated warming potential. We chose a 50% percentile, meaning that half of the models show a temperature offset minor or equal to that set in the scenario, with a representative concentration pathways RCP-8.5, meaning that an additional radiation of 8.5 W/m² is foreseen towards the year 2100) [21,22,23,24]. The linked data emphasize an all-electric residential building prototype modeled to investigate how climate change will impact buildings in terms of design, efficiency measures, loads, and renewable production.

The purpose of this paper is to provide data and methods on how climate change may also impact the design of NZEBs, heating and cooling loads, the selection of cost-optimal energy efficiency measures, and the production of renewable energy in different European climates. The target stakeholders of this paper and related dataset are mainly related to the scientific and policy community dealing with buildings and energy. The potential benefits are huge considering how the selection of weather datasets is crucial in modeling [25,26,27]. Another benefit is the support to EU policies related to technologies to be favored in buildings, energy performance calculations, design trends of NZEBs, and climate change impact on future buildings. The impact of the paper relates to the inclusion of climate change projections to predict the long-term performance and optimization of NZEB design. This inclusion considerably changes the energy balance in buildings and the efficiency measures needed to reduce cooling needs and overheating. Climatic parameters have evolved considerably over recent decades, and past weather datasets should not be used in order to avoid unreliable simulation results. A major finding is that improving energy efficiency will be even more crucial within a climate change scenario.

2. Data Description

The shared datasets include both the weather datasets and the building simulations files that can be used to investigate the impact of climate on energy performance and NZEB design as comprehensively illustrated in [20]. Eight locations are included in the datasets:

• Stockholm (Sweden)
• Milan (Italy)
• Vienna (Austria)
• Madrid (Spain)
• Paris (France)
• Munich (Germany)
• Lisbon (Portugal)
• Rome (Italy)

The locations were selected to have a good coverage of the European climatic variability based on heating and cooling degree days. The hourly files used in building energy simulations are available.

Yearly weather datasets (lsx) were collected and analyzed from 2003 to 2018 (folder “hourlyWeatherDatasets”). The folder contains the hourly weather datasets for all locations, heating and cooling loads (detailed per use) in: 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2020, 2017, 2018, 2018 TMY (typical metrological year), lWEC (International Weather for Energy Calculations), IWEC2, 2060. Sensitivity analysis is available for Milan and Stockholm.

In more detail, in the analysis, IWEC and TMY datasets were included as they are commonly used in building simulations. The IWEC and IWEC2 hourly datasets refer to the average weather observed, or TMY, over the last 15–25 years (IWEC 1984–2001, IWEC2 1994–2011) [28,29,30]. Single years are able to give the tendency of the ongoing conditions [31,32]. More recent TMYs are currently also available [33], together with data collected since 2011 based on satellite data [34].

In order o include climate change and obtain the 2060 projection, we considered the IPCC 5th assessment applying a down-scaling methodology to the most recent TMY file (2004–2018) [21,22,23,24,25,26,27,28,29,30,31,32,33,34,35]. The method is summarized in [36]. We referred to the commercially available Weather-Shift implementation of the Belcher et al. calculation [21]. We selected a business-as-usual scenario with limited mitigation (RCP 8.5 median 50%) including a sensitivity analysis of 10% and 90% lower and upper bound cases [37,38].

The shared datasets also relate to the building simulation files and related outputs in different climates. The data were elaborated using the open source tool BEopt [39]. BEopt is powered by the EnergyPlus and TRNSYS simulation engines. The first derives hourly heating, cooling, water heating, and appliance loads. The second estimates the renewable production for photovoltaic electric and solar water heating. Starting from the standard configuration, a sequential search technique is used to optimize the design. This allows calculation of the annual loads and resulting costs. The methodology, characteristics of the standard building, and assumptions are set out fully in [20].

The shared data allow us to identify, starting from a standard building, the optimal building design to reach NZEBs at the lowest cost in different climates. The data provide information on efficiency measures, envelope characteristics, equipment and systems, lighting, renewables, and costs in new residential buildings, both in current and future climate scenarios. The data include energy consumption, energy savings, and implemented measures in the base and optimized configurations.

The folder “building simulation file” contains the Beopt (.BEopt) and EnergyPlus (epw) files for the standard and optimized buildings in all locations, with present and future climates, the IDF (input data file), and XML executable files. The XML file provides data mapping between EnergyPlus and the external interface. The appliance schedules for the hourly simulation are contained in 1_sched.csv. The domestic hot water (DHW) schedules and quantities are contained in the DHW_3bed_unit0_10 min.csv file.

The folder “simulation file outputs” contains the annual simulation file results for all the studied locations and weather datasets. Heating and cooling loads (detailed per use) are given for: 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2020, 2017, 2018, 2018 TMY, IWEC2, lWEC, 2060. The variances from the 15-year average and weather averages are also given.

The subfolder “Standard and Optimized building_outputs_2018TMY and the 2060TMY” details the results (e.g., simulated electrical end-uses in kWh for: miscellaneous, ventilation fan, appliances, lights, cooling fan/pump, heating fan/pump, cooling, heating, hot water, total, PV, net) for all locations in terms of standard and optimized buildings using the TMY 2018 and TMY 2060 weather datasets.

The heating system comprises an air source heat pump so that there is energy use for the indoor fan blower, the outdoor compressor, and supplemental (Suppl.) heat which uses resistance heat when outdoor temperatures are very low and the compressor no longer has sufficient capacity for heating. An important part of the study is the consideration of PV changes in a climate change scenario. Forecasts indicate a warmer climate and a reduction in prevailing clouds. Furthermore, on-site short-term electrical storage systems are also considered for the first time as projections indicate that this system will be widely available and increasingly affordable in coming decades.

3. Methods

Comprehensive weather datasets are given in the provided data as detailed in Section 2. A preliminary analysis of the weather datasets was carried out to check for missing data and make relative corrections.

The weather data for Milan are 403K hourly data points (e.g., dry and wet bulb temperature, dew-point, relative humidity, wind direction and speed, global and diffuse horizontal radiation, precipitation, and sky cover) from 1973 to 2020. The average, variance, standard deviation, maximum, and minimum can be visualized as boxplots to show the distribution of the climatic variables using dispersion and position indexes.

As a data visualization example, Figure 1 shows the temperature and monthly boxplots, while Figure 2 shows the temperature boxplot, medians, and inter-quartile binned over five years from 1973 to 2018.

Figure 2 shows an increase in the outdoor average temperature, particularly after 1985. The most recent bin shows higher values, minimums, and medians. Apart from the median, minimum, and maximum values, boxplots also allow the visualization of the first and third quartiles, and outliers (Figure 3).

According to an ANOVA/regression analysis, the temperature in Milan grew over the period, with an average increase rate of 0.56 (±0.21) °C per decade, at a 95% confidence interval. The cross correlation among main climatic parameters can also be investigated as reported in

Table 1. Correlation among investigated variables in 2000 (linear correlation index −1 means an inverse correlation, 0 no correlation, +1 direct correlation) (T = dry bulb temperature (°C), RH = relative humidity (%), GHI = global horizontal irradiance (W/m²), direct (beam) normal irradiance (W/m²), diffuse horizontal irradiance (W/m²), infrared radiation downwards (W/m²), wind speed (m/s), wind direction (Deg.), air pressure (Pa)).

	T	RH	GHI	DNI	DHI	IRD	WS	WD	P
T	/	−0.24	0.46	0.32	0.43	0.85	−0.07	−0.008	−0.15
RH		/	−0.29	−0.22	−0.28	−0.21	−0.07	−0.02	0.15
GHI			/	0.89	0.77	0.33	0.16	−0.11	−0.006
DNI				/	0.47	0.16	0.11	−0.11	0.07
DHI					/	0.37	0.15	−0.05	−0.09
IRD						/	0.04	0.11	−0.33
WS							/	−0.06	−0.16
WD								/	−0.07
P									/

.

Table 1 confirms that the variables related to irradiance (in red) are strongly related to each other and to temperature.

Other statistical methods applied on weather datasets, temperature, and buildings’ performance variation are reported in Appendix A.

To evaluate the building design in current and future climates, a simulation-based optimization model is developed as described in [20]. A standard residential building prototype has been optimized to NZEBs to derive the most cost-effective design.

The methodology followed to produce the data includes a consideration of climate change both in the model set-up and in the weather datasets used to carry out the energy simulations. Among the important novelties:

• the building is all-electric, in line with the EU strategy of a future electrification of the sector;
• an updated library of energy efficient options (envelope, appliances, systems) is included, in particular the building envelope generally has a long-term impact due to the differing lifetime time horizon;
• the cooling set points are set upwards to 25.6 °C for the control air node operative temperature in compliance with ISO Standard 7730;
• a standard air-source heat pump is selected with electric resistance and a seasonal coefficient of performance (SCOP = 2.4) evaluated at multiple temperature conditions in EN 14511 and EN 14825. The operating SCOP varies depending on both the prevailing temperature conditions and building characteristics;
• a more efficient heat pump (SCOP = 3.1) is also available. This system is often selected, particularly in extreme heating or cooling climates.

We improve the standard building to reach the NZEB target (90% reduction in primary energy) at the lowest possible cost in specific locations.

Once the model has evaluated the standard building performance, all options are compared in a series of parametric evaluations, as reported in (1).

ESavings_i,n = (Base energy_n − Measure energy_n,i)

(1)

where:

• ESaving_n,i = energy savings within optimization iteration ‘n’ evaluated for option ‘i’.
• Base energy_n = calculated energy use of the standard building at the beginning of iteration ‘n’.
• Measure energy_n,i = estimated energy use of the base building with measure ‘i’ installed within iteration ‘n’.

The optimal range is assessed, analysing the primary energy and global costs related to the tested measures. The lowest point of the curve that belongs to the band is indicative of the optimal configuration [40,41,42]. The total costs over the life are then annualized to an annual cost of energy and additional mortgage linked to the added incremental costs [43].

At each iteration, the single most cost-effective option (lowest annualized cost of investment and energy costs) is selected after having evaluated all measures. Within the optimization process, the standard building evolves, and it is modified by adding the selected option at the end of an iteration before proceeding to the next. All remaining options are then re-evaluated together with the next selected option. This process continues until the performance target is reached, until zero energy is achieved using renewables, or the available cost-effective options are exhausted (Figure 4).

The sequential search technique allows us to reach the established target locating the least expensive path. It also locates intermediate optimal points. Thus, it is possible to identify what the 50% reduction level is along the path to reach a NZEB target.

For each location, the software ran around 200–400 simulations in 15–25 iterations to reach the target. A total of 19 weather datasets (e.g., IWEC, IWEC2, TMY, yearly datasets from 2003 to 2018, TMYshift_2060) were included for all locations (152 simulations for the standard building).

Global costs (C_G), in terms of net present value (NPV), for each combination of measures are derived using UNI EN 15459 [44].

The cost-effectiveness of each option is derived by estimating the NPV of the cost of the improvement or change over the life of the building. This is compared with the cost of the changing standard building:

NPV_n,i = I(V_n,a_n)

(2)

$$\mathrm{PV}=\mathrm{I}\left({\mathrm{V}}_{\mathrm{n}}{\mathrm{a}}^{\mathrm{n}}\right)+\sum _{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{a}}^{\mathrm{j}}\left({\mathrm{M}}_{\mathrm{j}}+{\mathrm{R}}_{\mathrm{j}}\right)+\sum _{\mathrm{k}=1}^{\mathrm{H}}\sum _{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{P}}_{\mathrm{k}}{\mathrm{Q}}_{\mathrm{k}}{\mathrm{b}}^{\mathrm{j}}$$

(3)

where:

• PV = total present-value of life-cycle costs before taxes;
• I = total first costs associated with energy saving measure;
• V_n = residual or salvage value at year n, the last year in the evaluation (50 years);
• a = single-present-value formula from j = 1 to n, and discount rate d; i.e., a^j = (1 + d)^−j;
• M_j = maintenance costs in year j;
• R_j = repair and replacement costs in year j;
• P_k = the initial price of the kth type of conventional energy carrier for energy types k = 1 to H;
• Q_k = the quantity required of the kth type of energy;
• b^j = a formula for finding the present value of an amount in the jth year, escalated at a rate Θ_k, where k denotes the kth type of energy carrier, and discounted at a rate d; i.e., b^j = [(1 + Θ_k)/(1 + d)].

The costs related to building elements not influencing the energy performance are omitted. The calculation considered an initial investment C_I and a yearly cost for every year i (referred to the starting year) for each component or system j, and a final value. Global cost (C_G) takes into account the calculation period τ according to (4):

$$C_G\left(\tau \right)=C_I+\sum _j\left[\sum _{i=1}^{\tau }\left(C_{a,i}\left(j\right)\times R_d\left(i\right)\right)-V_{f,\tau }\left(j\right)\right]$$

(4)

The final value V_f,τ (j) of a component is found by a straight-line depreciation of the initial investment until the calculation period end in reference to the start. If the calculation period τ exceeds the lifespan τ_n (j) of the considered component (j), the last replacement cost is accounted for in the straight-line depreciation as:

$$V_{f,\tau }\left(j\right)=V_0\left(j\right)\times {\left(1+R_p/100\right)}^{n_{\tau }\left(j\right)\ast {\tau }_n\left(j\right)}\times \left[\frac{\left(n_{\tau }\left(j\right)+1\times {\tau }_n\left(j\right)-\tau \right.}{{\tau }_n\left(j\right)}\right]\times R_d\left(\tau \right)$$

(5)

where:

$$V_0\left(j\right)\times {\left(1+R_p/100\right)}^{n_{\tau }\left(j\right)\ast {\tau }_n\left(j\right)}$$

(6)

is the last replacement cost, when considering the development rate of the price for products (R_p);

$$\frac{\left(n_{\tau }\left(j\right)+1\times {\tau }_n\left(j\right)-\tau \right.}{{\tau }_n\left(j\right)}$$

(7)

is the straight-line depreciation of the last replacement cost;

$$R_d\left(\tau \right)=\left(\frac{1}{1+R_R/100}\right)$$

(8)

i the discount rate at the end of the calculation period, and the real interest rate, depending on the market interest rate R and on the inflation rate R_i.

In relation to the economic parameters, the cost-effectiveness calculations of individual measures are based on the present value of life-cycle costs considering projections over 30 years. The methodologies for life-cycle calculations are in [45,46].

The assumed costs, service lives, and maintenance fractions for each of the hundreds of efficiency measures considered are given in an Excel sheet linked to the simulation. They are based on recommended guidelines supplementing Directive 2010/31/EU [4]. The energy costs for electricity are taken from [47,48].

No financial incentives have been assumed for either efficiency or renewable energy sources. However, a differing lifetime is specified for each measure considering data from a number of sources, including Standard EN-15459 [44].

The energy price inflation rate approximates the EU Emissions Trading Scheme with carbon pricing of 25 EUR/tCO₂ in 2020 to 39 EUR/tCO₂ in 2020. It is possible to alter the input parameters to consider very long time horizons and/or higher energy inflation rates. The optimization can also be limited to non-equipment options, providing a better evaluation of one-time interventions, such as those related to envelope insulation [49,50].

In relation to the modeled building, the data include an updated library of energy efficient options [51,52]. This includes technical data, climate, energy parameters, operation, maintenance, and replacement costs of measures related to the envelope, appliances, and systems [53,54].

The provided data include model inputs and outputs [55,56] (Figure 5).

The cost-optimal curve showing global costs (EUR/m²) and energy consumption (kWh/m²y) can be visualized [57,58]. The outputs can be graphed to show energy consumption, savings, selected efficiency measures, costs, and renewable production in present and future climates, before and after the optimization, and in current and future conditions [59].

Examples of data visualization from the provided dataset are shown in Figure 6 and Figure 7 in relation to Rome and Stockholm (data available from the folder “simulationFilesOutputs”). The figures show how climate change impacts the balance of heating and cooling in buildings in Rome and Stockholm from 2018 to 2060. The results also show how NZEB efficiency improvements help to control these differences.

Table 2. Standard building performance in Milan for the TMY2018 weather dataset (E = electricity). Raw data provided within the datasets.

End Use	Annual kWh
Misc. (E)	762.0
Vent Fan (E)	454.3
Lg. Appl. (E)	1591.5
Lights (E)	551
Cooling Fan/Pump (E)	187.6
Heating Fan/Pump (E)	442.6
Cooling (E)	524.6
Heating (E)	2230.4
Heating, Suppl. (E)	381.1
Hot Water (E)	2086.8
Total	9212
PV	7623
Net (Total PV)	1589

shows the outputs for the building located in Milan for the TMY2018 weather file with a standard (Std) building.

The simulation outputs show different results in the energy uses of the building prototype and different optimization results of the selected energy efficiency measures in the studied locations which are due to the changed climate data [60,61]. This dataset does take climate change into account, and this is evident in the altered building loads and, in particular, in the increased cooling needs of the residential prototype [62]. The selected options differ in many cases based on the altered weight of heating and cooling based on the morphed climate files. It is possible to visualize the selection of main efficiency options related to: building orientation, walls, ceilings, roofs, foundation, thermal mass, windows, airflow, and space conditioning. Different technical and cost data have been defined and are available within the provided files.