NZEB and market-based renovation case study of an existing office building

The goal of decarbonizing the building stock in the EU requires a multi-fold increase of the current renovation rates. In Estonia, the non-residential building sector has had little or no public support to improve the energy efficiency. Therefore, it is essential to study the energy efficient and cost-optimal measures for non-residential building renovation to give guidance to real estate companies and other stakeholders about the renovation alternatives. Furthermore, crucial is to provide input to the government to develop the renovation grant and incentives for renovation. In this study, energy renovation measures and savings to improve the energy performance to NZEB level were identified in a large (16 990 m2 heated area) office building. For that purpose, energy use was measured, simulation model developed and calibrated, feasible and more comprehensive energy improvements and costs analysed. The improvement of lighting, AHU, heating, installation of a 69 kW PV system, and window replacement was needed to achieve the goal with a primary energy use of 163 kWh/m2. However, some of the applied measures had long payback times of 40-70 years and are not realistic to be implemented without renovation incentives.


Introduction
European Union (EU) has a goal to develop a sustainable, competitive, secure, and decarbonised energy system in EU Member States by 2050. The purpose and guidelines are brought out in the energy performance (EP) of building directive (EPBD). [1]- [3] Further steps are to develop measures that reduce the final energy use of buildings to reduce greenhouse gas emissions. Non-residential buildings account for 25% (m 2 ) of the total stock in Europe and office buildings are the second biggest category (23%). However, the specific energy use of non-residential buildings (covering all end-uses) is at least 40% greater than the equivalent value for the residential sector. [4], [5] Energy use of a building is evaluated with the energy performance certificate (EPC) that indicates the EP of a building calculated by a methodology adopted in accordance with Directive 2010/31/EU [3]. In Estonia it is defined as primary energy (PE) and the limits from [6] are in Table 1. Regarding the standard ISO 52000-1 [7], to calculate the PE use and EPC of an existing buildings, the measured energy use can be used. However, the electricity for other uses than energy performance of buildings (non-EPB) (e.g., small power plug loads, commercial activities, elevators, industrial kitchen) should be excluded from calculation. Nevertheless, in Estonian regulation [6], [8] the appliances are included to the energy calculation. Therefore, as shown in [4] and [11], today major electricity uses in the office are not * Corresponding author: helena.kuivjogi@taltech.ee necessarily considered by long-term renovation strategies.
Estonian long-term renovation strategy states that today the annual PE use is 225-270 kWh/m 2 (delivered heating 130 kWh/m 2 and electricity 70 kWh/m 2 ) in office buildings constructed before 2000. However, to achieve the major renovation NZEB level (class C in Table 1), the PE for an office building should be 136-160 kWh/m 2 (delivered heating 70 kWh/m 2 , electricity 45 kWh/m 2 ). Unfortunately, today building stock has several barriers that prevent to achieve the NZEB level renovation. One main barrier is the long, 20-30 year payback time of NZEB renovation, as real estate companies prefer the payback time around 10 years. [9]  This study is a continuation and clarification of the article [9], that has contributed to Estonian long-term renovation strategy [11]. The idea is to give an overview of the measured energy consumption and analyse the energy renovation measures to improve the energy performance of NZEB major renovation in office buildings. The reference building (Figure 1) is actively used, well maintained, and with good ventilation system 5 storey office building with heated area 16 990 m 2 built in 2008. The construction has a deep foundation and non-loadbearing exterior walls (EW) (wood frame with mineral wool thickness 0,15-0,18 m), external floor (EF) towards ambient air (concrete panel with wood frames and insulation thickness 0,25 m), EF towards the ground (concrete floor with insulation in perimeter), roof (loadbearing profile plate and mineral wool plates or concrete panel with 0,18 m mineral wool and 0,05 m EPS 65F). Building has an uncommon shape and WWR about 50%. We consider that the reference building has 3 elevators (one for each building section). On the first floor, the trade areas and canteen (respectively, 10% and 3% of the total area) are located, and the other floors are for offices (about 83% of the total area) and health care (5% of the total area). Heat supply comes from the local natural gas substation and electricity from the power grid. We collected the delivered energy of the reference building from the energy invoices measured by the main energy meters (look in Figure 2). Measured total delivered heating energy was 104 kWh/m 2 and electricity 126 kWh/m 2 (non-EPB included). Specific services were not measured, but assumed by calculations and other studies. Domestic cold water use was measured and the domestic hot water (DHW) percentage of this was assumed to be 40%. Using this data, the DHW use was calculated by Equation 1: (1) Where E: energy, kWh m: the mass of water, l Cp: specific heat of water, kJ/kg °C ΔT: temperature difference, °C ߩ -density of water in 50 °C DHW use was separated from heating energy use and the last one was weather normalized with degree day method. Table 2 shows the monthly weather normalized heating energy, domestic hot water (DHW) heating, and electricity energy use in the reference building in the years 2012-2014. The specific electricity for non-EPB use (Figure 2) has been estimated by further calculations. Cooling electricity and plug load electricity for the server room has been calculated by the installed power of fan coils in server room. This building has 18 electrical service or server rooms that are cooled by fan coils with different cooling power: 1 kW to 4 kW. The sum of fan coil cooling power is about 61 kW and it was assumed that the servers will work every day all over the year (8760h). Annual cooling power is about 31 kWh/m 2 . The SEER of chiller has been taken 3.5 and therefore, the electricity use is 9 kWh/m 2 . Outdoor lighting is fluorescent lights with bulb power 58 W. As there are about 25 lights, the annual electricity energy for outdoor lighting was 6.2 MWh (0.36 kWh/m 2 ). The electricity of the 3 lifts was about 13 kWh/m 2 , estimated by the study [12].

Energy simulation model
For further analyses, there was built an energy simulation model (further as the base model) of the reference building (Figure 3) in the energy and indoor climate simulation program IDA Indoor Climate and Energy 4.8 (IDA-ICE) by using project documentation and energy audit information. The test reference year of Estonia was used [13].   Table 3 represents the input data of the base model. In the calculation of heating energy use, the generation and distribution were considered with heat source efficiency factor and efficiency factor of the distribution and output of radiator (  Table 3). For cooling calculation, there was taken into account also the heat loss coefficient and Eq. 2 was used from [8].
where Qje -annual net energy of AHU cooling elements kWh; Qrs -annual net energy of zone units kWh; βje -the energy loss associated with AHU cooling elements βrs -the energy loss of cooling energy distribution to zone Base model imitates the real performance of the reference building and has occupancy, appliances, lighting, and AHU schedules according to real use (in Appendix 1, Figure 9-Figure 14).
Finally, the delivered energy has been correlated with primary energy factors as for natural gas 1.0 and for electricity 2.0 to calculate the PE. Table 3. Input data of the base model and for energy use calculations

Calibration method of energy simulation model
Energy simulation model has been calibrated with the corresponding and normalized energy use of the reference building ( Table 2), both data for year 2014.
The calibration was done and the modelling uncertainty was controlled by the coefficient of variation of root mean square error (CV(RMSE)) Eq. 3, (regarding to ASHRAE Guideline [14]) which indicates the uncertainty inherent in the model and the computer model shall have a CV(RMSE) of 15% relative to the monthly calibration data. Results are in Table 4.
The energy performance simulation in standard condition was done to calculate the energy performance certificate (EPC) and clarify whether the measures achieve the level of NZEB major renovation. The base model according to standard conditions uses the schedule in Appendix 1, Figure 9, every zone for occupancy, appliances and lighting (except in the trade area, where the lighting and appliances schedule is 0.55 at 7-21:00). The AHU works in full capacity from 6:00 to 19:00 and outside these hours is set to work 7.5% of the total capacity.

Improvement of energy performance
Several energy saving measures have already been implemented in this building. However, the improvement of automatics, rebuilding of the cooling system, improvement of free cooling of ventilation and cooling system did not achieve the NZEB level of this building regarding to energy audit. Furthermore, there has been constructed the combined heat and power (CHP) station and the photovoltaic (PV) panel system. Arithmetical calculation shows that these savings will achieve the C-level of this building, but needs further detailed investigation. This study was considering different measures to increase the energy use and achieve the C-level of the reference building. Several measures were selected by the suggestion of energy audits, the knowledge of the authors, and measures earlier studied in [9] were improved.
The measures were added to the base model individually (method described in Figure 4) and the saving of every measure has been calculated with the energy simulation software IDA-ICE 4.8.  Furthermore, the sequence of measures regarding the best saving of primary energy was composed and the models with measure packages were composed by adding each measure one by one to the model, so each package had one measure more than previous.
The costal savings and investment of each measure was estimated by calculations with data from project documentation and energy audits, and the experience of the authors of this study. The energy price has been taken from energy invoices in the years 2017-2018 as 0.031 €/kWh and 0.072 €/kWh, respectively, natural gas and electricity.

Results and discussion
Specific energy use for heating and electricity of the base model has brought out in   Composing the sequence according to the best saving of primary energy (Table 4) Addition of:   Table 5 were implemented in the base model and the energy savings were identified. The highest saving in primary energy was achieved by replacing the fluorescent lighting with LED -18.6 kWh/m 2 annually. That was also the cheapest measure with 0.47 €/kWh saved primary energy and the shortest 12.5-year payback time. Measures 2 and 3 have bigger investment, because the improvement of AHU fans or heat exchanger separately still needs a replacement of the entire air handling unit that costs about 20 €/m 2 . The payback time will be 26 years, if the measures are made together (AHU replaced with unit with better fans SFP and heat exchanger temperature efficiency). From Figure 6 we can see that measures like sun protection film and improvement of free cooling function were less worth measures and in further analyses they are not involved. The installation of the PV-panel system was an additional measure to achieve NZEB level.
The energy simulations were done with the base model and models with the four following packages (pack.) with similar payback time:  (Figure 7-Figure 8) -respectively, 19 kWh/m 2 and 8.4 kWh/m 2 . However, the standard use PE is 173 kWh/m 2 that stays in EPC D-level. Further packages did not improve the electricity performance, but reduced the heating energy use considerably. Adding measure 5, the canteen AHU control with CO2 sensors (package 2) reduced the heating energy for 7.2 kWh/m 2 and resulted the standard use PE 165 kWh/m 2 , that is close to NZEB level. With package 3, after replacing the windows, the heating energy reduces for 3.5 kWh/m 2 and resulted in NZEB level with standard PE use 160 kWh/m 2 . Using the night setback of indoor temperature in addition to previous measures (package 4) saved the heating energy for 4.3 kWh/m 2 and gave good PE use even without PV-panels -163 kWh/m 2 .
Nevertheless, to achieve NZEB major renovation, the installation of PV panel system is essential.
The payback time of individual measures is high and will be over 40 years for half of measures. However, if measures are made as packages, the payback time is smaller, but still in between 30-34 years. Furthermore, replacing the window change measure with the night setback measure will reduce the investment of package 3, but the pay-back time will be still about 28 years.

Conclusion
Analyse of energy efficiency and cost-optimal measures for non-residential building renovation is essential to achieve the goal and decrease greenhouse gas emissions. Therefore, this study focused on energy performance and renovation measures, citing one larger office building as an example. The measured energy consumption, the calibration method of energy simulation with energy consumption of the existing office building, and identification of energy renovation measures were the tools to discover the improvement of the energy performance to NZEB level of major renovation.
The cheapest measure, changing the lighting improved the electricity energy performance significantly and therefore, saved the primary energy for 18.6 kWh/m 2 . In general, the majority of measures were with good primary energy saving (5-18 kWh/m 2 ), but at the same time with long payback time for real estate companies -over 16 years, and even will reach over 60 years. Using the measures together in renovation packages reduced the payback time to 30 years, that is still not suitable for real estate companies.
Authors found that replacing fluorescent lighting to LED, upgrading the AHU with a better performance unit, and improving the automatics, using night setback for heating and replacing windows will result the primary energy use close to the NZEB as 163 kWh/m 2 . However, to reach this goal, the 69 kW photovoltaic panel system is needed.
The replacing of the windows in an office building is a big investment measure and will be under serious consideration, is it needed or not. Calculations show that replacing the window change measure with the night setback measure will result the PE use 161 kWh/m 2 , that is almost NZEB level, but it needs further investigation with some more case studies.
The limitation of this study is the calibration method, as the measured data has been weather normalized, but the simulation were done with test reference year. Further studies need to be done to investigate the impact of this difference and to develop the calibration methodology. Appendix 1 Figure 9. Schedule for occupants, equipment and lighting