**1. Introduction**

The implementation of zero-emission plans in Europe poses serious challenges for EU countries. One is that new buildings must be designed to be near zero energy [1]. However, it seems that modernizing existing facilities is even more difficult. In this case, pre-war tenement buildings with the highest energy demand appear to be the greatest challenge. In these buildings, it is necessary to change the heating systems to low-emission ones, and first to reduce their energy consumption. Each EU country faces different challenges in this area.

In Poland, an additional problem with pre-war buildings is the significant number of flats still equipped with solid-fuel heating systems. Research carried out in Wroclaw, Poland, in 2019 showed a high concentration of such heat sources in buildings built before 1945, i.e., downtown tenement houses. It turned out that as many as 64% of the apartments heated with solid fuel in the city (12,000 out of 18,700) are included in this group of buildings [2]. Carbon-based heat sources are the direct cause of emissions of substances such as particulate matter (PM2.5, PM10), benzo(a)pyrene (BaP), nitrogen oxides, sulfur dioxide, and carbon monoxide. The emission level is also directly related to the energy consumption of buildings. Estimating this energy provides a basis for building models related not only to air pollution, but also to energy modeling of municipal buildings (UBEM) [3]. Actions should be taken on many levels and interrelated goals should be defined: improving air quality in cities, reducing the energy consumption of buildings in urban agglomerations, and improving the quality of life of their inhabitants.

To conduct large-scale activities, it is necessary to properly identify the housing structure, which will allow for long-term planned research and investments. The current research directions are aimed at developing a methodology for modeling the energy demand of buildings at various scales. The micro scale includes individual premises and buildings, the meso scale determines the energy consumption of districts or building quarters, and the macro scale applies to entire cities [4]. The basic input data for modeling are meteorological data, 3D models of cities, and nongeometric properties of buildings. Such modeling also requires the prioritization of housing stock into building archetypes. Their construction is based mainly on the age, shape, method of use, and heating systems used. The process of creating archetypes is a crucial part of building models and making them reliable. The main problem indicated by the researchers is the fact that UBEM designers do not have access to the measured final energy consumption in premises or buildings. Limited access to such data and generally insu fficient knowledge about the thermal properties of buildings in di fferent age categories often make it impossible to simulate uncertainty and reduce the inaccuracies in model construction [3]. Therefore, it is necessary to provide data resulting from the calculation of energy demand, as well as on actual energy consumption in selected well-recognized buildings. There is a need to develop a way to consider user behavior that significantly a ffects the real energy consumption of buildings and apartments. In addition, researchers point out the need for a detailed approach to the problem of predicting energy demand profiles of building resources with detailed time resolution. These profiles should take into account geographic location, user behavior, and the e ffectiveness of various energy systems and technologies in cities [5].

To conclude, the literature lacks information supporting the determination of energy demand considering all of the above conditions on a micro scale, especially in older buildings. Currently, methods for energy calculation and saving mainly focus on new buildings and algorithmic determination of energy consumption, hence they often fail for existing buildings [6]. The share of historic buildings in the centers of European cities is large, which means that accurate information on their energy consumption is critical in the process of modeling and planning changes aimed at local improvement of air quality, but also to a ffect progressive climate change. Researchers who have access to data on the actual energy consumption of older buildings emphasize that they may use less energy than expected, pointing to the fact that the greater energy demand of these buildings leads to more conscious use of the systems [6]. Collecting as much data as possible on this topic will allow researchers to construct more accurate energy models for cities.

The actual energy consumption of a building often di ffers significantly from the computed value, even if it is obtained using advanced complex software for dynamic simulated energy performance. This phenomenon is commonly known and described in the literature as the energy performance gap (EPG) [7–9]. Many authors use this indicator to assess the energy situation in relation to real measurements for residential, nonresidential, and single-family buildings [10]. The energy performance gap is defined depending on the situation. For example, de Wilde [9] defines three types of gaps: between initial predictions and measurements, between machine learning and measurements, and between predictions and display certificates in legislation. The article, however, is mostly related to the design of new buildings. In the literature, this coe fficient also appears in studies that describe the e ffects of the thermal modernization process of buildings. For example, in [11], EPG is defined

as the difference between actual and design consumption as a fraction of the design consumption. Interesting research is presented in the article [8]. The size of the EPG was determined for different refurbishment solution. The scientists analyzed the impact of the occupants' interviews and surveys, and the consequential feedback to the occupants about the correct behaviors and use of the heating system. The results show that, for the buildings in which occupants learned how to optimally use the system, EPG was equal or smaller than 0% and occupant behaviors has been identified as one of the causes of the gap.

In the context of modernized buildings, it is worth noting that the literature describes a certain tendency of users to increase their needs after higher technological standards are achieved in the building, the so-called rebound effect. This effect is also known as the Jevons paradox [12]. It was found that the assumed economic use of energy in such cases is not always confirmed, and the truth is often the opposite. The rebound effect is described in two forms: indirect and direct. The direct rebound effect is described as increased efficiency and the associated reduced cost of a product or service, resulting in its increased consumption. The indirect rebound effect describes the situation where savings from reduced efficiency cost enable spending more income on other products and services [13]. A complementary concept is described in [14], in which the authors introduced the term "prebound effect". This effect measures the discrepancy between the measured and calculated energy requirements of existing buildings that have not been thermomodernized. Researchers have observed that these buildings consume less energy than calculated using the methods in energy certification. A higher prebound effect in combination with low income may be a sign of energy poverty. The authors emphasize that further research is needed to understand motivations and practices in households that show a high prebound effect.

The actual level of final energy demand of a building is influenced by users' behaviors, which is related to the tenants' individual characteristics, such as their age and habits [15–19]. For example, the heating schedule for the residences of retirees will differ significantly from those of young working people or families with children. Since the former stay in their apartments more and usually have higher temperature requirements, a correspondingly longer heating time is needed. Kashif et al. [20] stated that both time and environmental factors create a certain context, which means that residents must perform certain activities to adapt their surroundings to their needs. As inhabitants have different approaches to the balance between indoor comfort and energy consumption [21–25], they behave differently despite being in the same environment. It can be concluded that the existing studies confirm the discrepancies between expected and observed energy parameters of buildings. The energy performance gap is mainly due to differences between the assumptions for engineering calculations and reality, including, in a significant way, the behaviors of users: the savings effect or the effect of increased demand. This paper enriches this topic by providing information on the size of the difference between real and computed energy consumption in pre-war tenement houses, indicating the reasons. As it turns out, users' behaviors are very closely related to the characteristics of heating systems in buildings and reflect more problems than those mentioned in the literature so far.

#### **2. Materials and Methods**

#### *2.1. Aim of the Research*

The aim of the research was to measure the energy gap between the real and computed energy consumption of residential premises located in pre-war tenement houses and to indicate possible reasons for its occurrence. The calculations were made separately for energy for space heating and for the preparation of domestic hot water (DHW). The research was extended with a detailed analysis of the actual situation in apartments and their surroundings. Not only was energy consumption measured in the examined premises, but also the parameters of indoor and outdoor air and other factors determining the energy consumption of apartments. The collected data allowed for model development using field measurements. The model was then used to simulate energy consumption under various conditions. This in-depth analysis was aimed at assessing the impact of users' behaviors and real weather conditions on the energy consumption of tenement houses in Wroclaw.

### *2.2. Subject of Study*

Measurements were carried out in 15 apartments (labeled A1–A15) located in pre-war tenement houses in the selected quarter of the city of Wroclaw (Poland) from 14 January to 9 March 2020. The technical standards of the facilities are varied: there are buildings that require extensive renovation (very bad condition), buildings in slightly better condition (bad condition), neglected buildings often found in Wroclaw (average condition), and those in which partial or full thermal modernization has been carried out (good condition). The facades of exemplary buildings, showing their typical technical conditions, can be seen in Figure 1. All apartments are in the same type of building (built before 1920), but they were selected due to di fferences on the micro scale. They di ffer in size, location in the building, heat source, and the tenants' situation and lifestyle. The premises were selected for research based on a pilot study conducted in 2018 [26] covering 410 residential premises. Basic data on the examined premises are presented in Table 1. Additionally, the issues important for the test results should be emphasized: there was no bathroom in apartment A9 and no DHW preparation system in apartment A7.

**Figure 1.** Four specified technical conditions of tenement houses in Wroclaw (according to Table 1 and described in the text): (**a**) good; (**b**) average; (**c**) bad; (**d**) very bad.


**Table 1.** Basic information about the apartments.


**Table 1.** *Cont.*

SF, solid fuel; EE, electrical energy; DH, district heating; NG, natural gas.
