Analyzing grid frequency behavior in response to weather variations: A statistical approach
The transition towards carbon neutrality in the Nordic Power System (NPS) by 2035/40 intensifies the integration of Renewable Energy Sources (RES), especially wind and solar power. This change introduces significant challenges in maintaining the balance between electricity demand and generation, imp...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2024 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/425939 |
| Acceso en línea: | https://hdl.handle.net/2117/425939 |
| Access Level: | acceso abierto |
| Palabra clave: | Electric power system stability Renewable energy sources Energy policy Sistemes de distribució d'energia elèctrica--Estabilitat Energies renovables Política energètica Àrees temàtiques de la UPC::Energies::Recursos energètics renovables |
| Sumario: | The transition towards carbon neutrality in the Nordic Power System (NPS) by 2035/40 intensifies the integration of Renewable Energy Sources (RES), especially wind and solar power. This change introduces significant challenges in maintaining the balance between electricity demand and generation, impacting grid frequency stability, a critical parameter that indicates the reliability of the power system. Despite extensive research on RES integration, there is a lack of studies examining the direct impact of weather conditions on grid frequency and frequency containment reserve (FCR) activation in the NPS. This study investigates how weather phenomena can influence grid frequency stability and the activation of FCR, aiming to model these relationships using statistical algorithms. A literature review explored the theoretical background of NPS frequency control services, weather-dependent generation sources, prominent weather parameters, and existing statistical modelling approaches. The review identified wind speed and air temperature as crucial weather parameters in the NPS influencing generation variability and demand fluctuations. Besides, the research suggested the need for advanced models to capture complex, nonlinear relationships between these weather variables and grid frequency. A dual-phase analysis was involved in the methodology, utilising historical data from 2017 to 2022. Long-term analysis involved collecting and preparing datasets on grid frequency, wind speed, and air temperature, ensuring data quality through synchronisation, cleaning, and handling missing values. The methodology combined descriptive statistical measures, exploratory data analysis (EDA), and time-series modelling using AutoRegressive Integrated Moving Average with eXogenous variables and Seasonal AutoRegressive Integrated Moving Average with eXogenous variables (ARIMAX/SARIMAX) within the Box-Jenkins framework. In the short-term analysis, Non-Deterministic Frequency Deviations (NDFDs) from 2022 were focused on utilising high-resolution data to capture disturbance events. Criteria for disturbances were defined, and key frequency quality indicators such as the Rate of Change of Frequency (RoCoF) were calculated. Reliability analysis was conducted by calculating the Time to Failure (TTF) and Mean Time Between Disturbances (MTBD), allowing for the assessment of the grid’s resilience during different seasons and Frequency Containment Reserve for Disturbance (FCR-D) activation. Throughout both phases, various probability distribution functions were fitted to the frequency deviation data, with the Johnson SU distribution emerging as the best fit. This distribution effectively captured the skewness and heavy tails observed. The study shows that extreme weather conditions, especially during winter, cause increased frequency deviations and frequent FCR-D activations, indicating that traditional time-series models are insufficient for forecasting. The results emphasise the need for advanced modelling methods, real-time monitoring, and adaptive frequency control strategies due to the variability caused by RES integration. This research provides practical options for maintaining grid reliability and stability in the evolving energy context of the NPS. |
|---|