**3. Results**

## *3.1. Buildings With District Heating*

This section shows the specific power demand of heating and electricity use for all building configurations that use district heating. Both the original and the retrofitted cases are presented. Figure 4 (left side) shows the hourly duration curves of district heating use in all applicable buildings. The blue lines show the district heating demand in the reference cases, highlighting the improvements in energy efficiency along with the tightening building code (new buildings consume less energy). The green lines show the district heating demand for buildings that have been retrofitted to the minimum cost level, D. This includes the use of an air-to-air heat pump, a slightly improved building envelope and some solar thermal capacity. For the older building types SH1 and SH2, a major drop in both peak DH demand (−33% and −23%, respectively) and average DH demand is seen. The differences are the smallest where the demand is the lowest. Finally, the red lines show the DH demand for buildings that have been significantly retrofitted to level B. This includes a further improved envelope and a large solar thermal capacity. For half of the year, no district heating is needed at all because of abundant solar energy. The peak DH demand goes down by 30 to 50% in all cases, as generally does the DH demand.

The duration curves give no indication on the temporal distribution of power demand. Having several hourly chronological lines would make the figure too cluttered to read, which is why Figure 4 (right side) shows the weekly maximum and minimum DH power consumption instead. However, instead of the single highest demand hour, the sustained peak power demand, that is, the average of the top (or bottom) 5% of weekly demand is shown. Each week was sorted according to hourly power demand and then the 8 hours (5% of 168 weekly hours) with the highest (or lowest) demand were averaged to obtain the plotted values. Together the values show how much the power demand varies within shorter periods during the whole year and give an indication on the required energy storage capacities. In the B cases, the summertime heating demand goes to zero, thanks to solar energy.

Similar to the heating demand, Figure 5 presents the electricity consumption of all district heated buildings. It can be seen on the left side that the minimum electricity consumption for the Reference and D cases remains slightly positive. However, in case B, the solar PV capacity grows so high that the exported excess solar power exceeds the maximum purchased power as much as six times. This increases the load on the grid, but could support the energy needs of other users in the grid. However, such an influx of power might require strengthening of the local grid, which was not taken into account in the optimization. The use of air-source heat pumps in the retrofitted buildings increased the peak electric power demand by as much as 60% compared to the reference case, but the absolute value of the increase is much lower than the power exported from the PV system. The reported values include only the heating use of AAHP, as cooling energy was not considered in this study. The right side of the figure shows the sustained peak (and bottom) power demand. This is the average power of the 5% of weekly hours (8 h) with the highest (or lowest) demand.


**Table 2.** The building configurations and other properties of the oldest single-family house Table 1976. SH1. Ref is an unmodified case while D and B are retrofitted cases [13]. For ground-source heat pump (GSHP) capacity, the number in parentheses is the ratio of HP thermal capacity to the maximum space heating demand. For


**Table 3.** The building configurations and other properties of the 1976–2002 single family house type, SH2. Ref is an unmodified case while D and B are retrofitted cases [13]. For GSHP capacity, the number in parentheses is the ratio of HP thermal capacity to the maximum space heating demand. Mech E stands for mechanical


**Table 4.** The building configurations and other properties of the 2003–2009 single family house type, SH3. Ref is an unmodified case while D and B are retrofitted


**Table 5.** The building configurations and other properties of the post 2010 single family house type, SH4. Ref is an unmodified case while D and B are retrofitted

**Figure 4.** District heating use of all district heated houses. On the left: The hourly duration curve of district heating (DH) power demand for the whole year. On the right: The sustained peak/bottom DH power demand. Solid lines depict the weekly top 5% of demand and dotted lines the weekly bottom 5% of demand.

**Figure 5.** Electric power demand of the district heated buildings. Negative values represent exports of excess electricity back to the grid. On the left: The hourly duration curve of power demand for the whole year. On the right: Sustained peak/bottom power demand. Solid lines depict the weekly top 5% of demand and dotted lines the weekly bottom 5% of demand.

## *3.2. Buildings with Wood Boilers*

The wood boiler use in the wood heated buildings closely matches that of district heating in the previous section. The electricity use for buildings with wood-based heating is also similar, as shown in the duration curves in Figure 6. The retrofitted cases have higher maximum power demand, because part of the wood-based heating was shifted to the air-source heat pumps in both the D and B cases. In the minimum demand side, we see that the B cases have very large exports of electricity due to the large solar panel arrays installed in the high investment cases.

The sustained weekly peak power demands are also shown in Figure 6 (right side). The sustained peak electricity demand in summer is the lowest for the B cases, due to self-consumption of solar electricity. In winter, the B cases have the highest peak demand, because of higher capacity air-source heat pumps. Solar panels produce very little power in the Finnish winter and do not influence peak demand.

#### *3.3. Buildings with Direct Electric Heating*

Figure 7 shows the duration curves of electricity use for buildings with direct electric heating. Electric heating in the oldest building (SH1) increases the peak demand over ten times compared to the non-electrically heated buildings. However, with electric heating the retrofits significantly reduce electricity demand, unlike in the DH and wood boiler cases. Here the absolute values of the maximum demand and the maximum solar energy exports are of the same scale in the retrofitted cases. Qualitatively, all the different age classes have similar sets of duration curves.

Figure 7 also shows the sustained peak and minimum power levels (the average of 5% of weekly max/min hours) for all cases with direct electric heating. During summer, the peak demand in the retrofit B case is similar to the minimum demand in the reference case for all age classes. In winter, the sustained peak demand in SH1 was 70, 50 or 40 <sup>W</sup>/m<sup>2</sup> for the Ref, D and B cases, respectively, showing a grea<sup>t</sup> reduction due to the retrofits. The difference was smaller for the newer building, such as SH4, where the sustained peak demands were 45, 36, or 30 <sup>W</sup>/m2, for the Ref, D and B cases, respectively.

#### *3.4. Buildings with Ground-Source Heat Pumps*

Use of the GSHP produced a major decrease in peak electric power demand compared to direct electric heating. Duration curves of the GSHP cases are shown in Figure 8. Comparing the original GSHP systems to the retrofit scenarios, in the case of SH1, the D level retrofit reduced peak demand from 55 to 43 <sup>W</sup>/m<sup>2</sup> while the capacity ratio (HP power vs. space heating demand) was 68% for both cases. In SH2, SH3, and SH4 there was no difference in peak power when comparing the reference buildings with original GSHPs and the D level retrofits. The power demand increased significantly during the peaks compared to the base level due to capacity constraints of the GSHP. The systems did not cover 100% of heating demand and electric backup heating was needed, thus significantly increasing demand during peak hours. With the level B retrofits, which included significantly improved thermal insulation of the building envelope, the heat pump size was sufficient (93% of space heating demand in SH2 and over 100% for the rest) and peak power was in check. As in the other cases, the PV arrays were oversized and the power exported to the grid was comparable to and even higher than the demand from the grid.

The seasonal variance of the electric power demand is shown on the right side of Figure 8, which shows the weekly top and bottom 5% of power flow. In SH1, retrofit D reduced sustained winter peak demand from 50 to 36 <sup>W</sup>/m2, while in SH2 and SH3 there was no difference. In SH4, retrofit D actually had higher peak demand, because the GSHP was sized down from 83% capacity ratio to 69% and thus the electric backup heater saw more use. Retrofit B shows major decreases in power demand during the entire heating season for all building age classes. It also significantly lowers the absolute variance between the peak demand in high and low demand time periods. For example, during weeks 1 to 5 in the case of SH2, the peak demand changes by 15 <sup>W</sup>/m<sup>2</sup> in the reference and retrofit D, but in retrofit B the change is only 5 <sup>W</sup>/m2. This is due to the sizing difference of the GSHP systems. In the retrofit B cases the GSHP capacity is close to peak demand, which means that electric backup heating with high power demand is not needed. In both the retrofit levels D and B, the exports of surplus solar electricity happen at high power. Especially in level B, the peak power in exports in summer is 1.5 to 3 times as much as the winter demand peak. Depending on the strength of the distribution grid, the optimal solution may not be feasible after all.

**Figure 6.** Electric power demand of the wood-heated buildings. Negative values represent exports of excess electricity back Table 5 of demand, and dotted lines depict the weekly bottom 5% of demand.

**Figure 7.** Electric power demand of the buildings with direct electric heating. Negative values represent exports of excess electricity back to the grid. On the left: The hourly duration curve of power demand for the whole year. On the right: the sustained peak/bottom power demand. Solid lines depict the weekly top 5% of demand, and dotted lines depict the weekly bottom 5% of demand.

**Figure 8.** Electric power demand of the buildings with ground-source heat pumps. Negative values represent exports of excess electricity back to the grid. On the left: The hourly duration curve of power demand for the whole year. On the right: Sustained peak/bottom power demand. Solid lines depict the weekly top 5% of demand, and dotted lines depict the weekly bottom 5% of demand.

## *3.5. E*ff*ect on Building Stock*

This section shows the e ffect these energy renovations could have on the whole building stock, namely the electricity consumption levels. Some retrofit actions increase electricity consumption, while others decrease it.

Significant changes in both district heating and electrical power levels were realized with energy retrofits in all the building age categories. In the previous study [13], it was found that switching to electrified heating resulted in significant emission reductions, when monthly average emission factors of Finnish electricity generation were used. However, as the number of retrofitted buildings goes up, the changes in consumption patterns start to influence the national grid. With increased electricity use, the average emission factors may no longer be reasonable. Thus, the changes in DH and electricity demand need to be quantified on the building stock level.

To estimate the potential influence of retrofits on a larger scale, assumptions about the retrofit levels were made in reference [13]. Of buildings that use wood or oil boilers for heating, 50% switch to GSHP and the other 50% switch to or keep using a wood boiler, while also doing other improvements. Buildings already equipped with GSHP are not renovated. In the other buildings, the main heating system remains unchanged while other retrofit actions are performed as described in Tables 2–5 and reference [13]. The buildings are retrofitted to either the minimum cost level D or the costlier but high impact level B.

The total annual electricity demand in each scenario is shown in Figure 9. Buildings with direct electric heating consumed most of the electricity in the base scenario. In both of the retrofit scenarios, the electricity consumption of ground-source heat pumps increased significantly. However, the improved energy e fficiency of buildings with direct electric heating more than compensated for the increase in heat pump use and the total electricity consumption went down in both the scenarios D (−11%) and B (−38%). The total PV capacity of the retrofitted building stock was 4400 MW in scenario D and 5600 MW in scenario B. It was assumed that no PV panels were installed in the reference buildings, although at the end of 2018 there was actually a total installed solar electric capacity of 120 MW in Finland [38]. Solar electricity produced in the retrofitted buildings significantly exceeded how much could be used in detached houses without energy storage technologies. Self-consumed solar electricity has been subtracted from the demand values presented, but the surplus amounts are shown separately as negative demand. A small part of the surplus could be utilized in other detached houses (Usable solar), which have unmet electricity demand, but the majority of it is excess energy that needs to be used in some other sector or by other building types (Excess solar).

The electric power levels in the whole building stock are shown in Figure 10. It shows the weekly top and bottom 5% of power as well as the hourly duration curves. Like in the individual building level, the excess solar power production was significant in the whole building stock as well. Notably, in retrofit scenario D, the peak demand during winter grew compared to the reference scenario, even though the total electric energy demand went down (as shown in Figure 9). In retrofit scenario B, the peak demand remained almost the same as in the reference scenario, even though the annual energy demand was significantly lower. This is elaborated in the duration curve of Figure 10, where the power demand of scenario B is lower than the original scenario for every moment after the peak. The largest di fferences are seen after hour 6570, when the large PV arrays of the retrofitted scenarios result in significant excess power. Solar electricity reduces power demand mainly when the demand is not very high anyway and has no e ffect on the peak demand. On the other hand, the large PV capacities greatly increased the power flow during the summer. In scenario B, the summertime export power reached 5 GW, while the imported peak power was only 4 GW.

**Figure 9.** Annual electricity demand for the whole detached house stock. Also shown is the excess solar electricity produced in the buildings. Usable solar is surplus production from some detached houses that could be used in other houses, while Excess solar is surplus that needs to be used in some other sector (values presented).

**Figure 10.** Sustained peak and bottom electric power (5% of weekly hours) in the whole building stock with the original systems and in the retrofit scenarios D and B. Also shown are the hourly duration curves of the same original data.
