Aivars Rubenis, Aigars Laizāns
The Urban energy systems face increasing pressure to decarbonise while maintaining security of supply and resilience to external disruptions. Rooftop photovoltaic (PV) systems on multi-apartment residential buildings offer a promising pathway to increase locally available renewable electricity in dense urban environments. However, reliable assessment of PV contribution requires a statistically robust understanding of residential electricity demand patterns at high temporal resolution. This article presents the first stage of a multi-layer study assessing the technical viability of rooftop PV deployment in multi-apartment buildings in Riga, Latvia, with a focus on energy availability and resilience. The object of the study is residential electricity demand in apartment buildings. Using hourly smart-meter data from a representative sample of 198 apartments (3.1 million observations), the study analyses whether households with different annual electricity consumption levels exhibit systematically different diurnal or seasonal demand structures once scale effects are removed. Electricity demand profiles were normalised by annual consumption and analysed using functional analysis of variance with permutation-based inference. The results show that mean hourly electricity consumption ranges from 0.09 kWh h⁻¹ in the lowest consumption group (<1000 kWh yr⁻¹) to 0.44 kWh h⁻¹ in the highest group (>3000 kWh yr⁻¹), while dispersion also increases markedly. However, after normalisation, neither diurnal nor seasonal load-shape functions differ statistically across consumption groups. The tests do not reject equality of diurnal profiles (p = 0.24) or seasonal profiles (p = 0.30). These findings demonstrate that higher residential electricity consumption in multi-apartment buildings reflects scale effects rather than distinct daily routines or seasonally concentrated usage. An assumption about winter-dominated electric heating in high-consumption apartments is not supported at a statistically significant level. From a practical perspective, the results support the use of representative, scalable demand profiles in subsequent modelling of rooftop PV generation, energy flows, and local energy availability, providing a robust foundation for resilience-oriented urban energy assessments.