Hasilnya menunjukkan bahwa generator profil beban yang dimodifikasi berhasil menghasilkan keluaran pada berbagai resolusi sub-jam, menawarkan pandangan yang jauh lebih terperinci tentang konsumsi energi rumah tangga dibandingkan dengan model per jam asli. Analisis menunjukkan kemampuan alat yang ditingkatkan untuk menangkap pola operasional berdurasi pendek dan puncak beban untuk berbagai peralatan, yang biasanya dirata-ratakan dalam profil per jam. Fitur opsional untuk mengonversi energi keluaran (Wh) menjadi daya rata-rata (W) juga ditambahkan untuk memfasilitasi analisis beban yang relevan dengan perencanaan jaringan.
Kesimpulannya, karya ini menghasilkan generator profil beban yang ditingkatkan dengan fleksibilitas dan akurasi temporal yang lebih baik. Dengan mendukung langkah waktu sub-jam hingga satu menit, peningkatan granularitas ini sangat penting untuk aplikasi yang membutuhkan data beban terperinci, seperti analisis stabilitas jaringan, perencanaan respons permintaan, dan penilaian akurat kontribusi permintaan puncak dari beban dinamis.
German electricity network operators face increasing complexity in forecasting household customer behaviour due to the rise of decentralized generation (like photovoltaics) and new loads such as electric vehicles and heat pumps. Traditional planning methods relying on standard load profiles aggregated at hourly intervals are often insufficient to capture the dynamic nature of modern energy consumption and prevent grid issues. While the Department of Electrical Power Systems (eES) at the University of Duisburg-Essen developed a Python-based bottom-up model to simulate hourly household loads considering various socioeconomic factors and technologies, its fixed hourly resolution limits its ability to accurately represent short-duration load peaks and rapid fluctuations in many households appliances.
This thesis aimed to enhance the flexibility and temporal accuracy of the existing load profile generator by enabling user-selectable sub-hourly time resolutions. The core methodology involved modifying the Python codebase to support dynamic time steps (including 30, 15 and 1-minute intervals). Key technical enhancements included implementing interpolation and extrapolation methods for weather data alignment, restructuring the core domestic device calculation algorithm to incorporate randomized start times and improved multi-phase device representation, and adapting the modules for electric vehicles (using emobpy), photovoltaics (using PVlib), and heat pumps to ensure compatibility and proportional energy distribution across finer time scales. Efficiency improvements, such as vectorization, and code clarity enhancements were also implemented.
The results demonstrate that the modified load profile generator successfully produces outputs at multiple sub-hourly resolutions, offering a significantly more granular view of household energy consumption compared to the original hourly model. Analysis shows the enhanced tool's improved capability to capture short-duration operational patterns and peak loads for various appliances, which are typically averaged out in hourly profiles. An optional feature to convert output energy (Wh) to average power (W) was also added to facilitate load analysis relevant to grid planning.
In conclusion, this work delivers an enhanced load profile generator with improved temporal flexibility and accuracy. By supporting sub-hourly time steps down to one minute. This increased granularity is crucial for applications requiring detailed load data, such as grid stability analysis, demand response planning, and accurate assessment of peak demand contributions from dynamic loads. "