Frequency in power systems is a real-time information that shows the balance between generation and demand. Good system frequency observation is vital for system security and protection. This paper analyses the system frequency response following disturbances and proposes a data-driven approach for predicting it by using machine learning techniques like Nonlinear Auto-regressive (NAR) Neural Networks (NN) and Long Short Term Memory (LSTM) networks from simulated and measured Phasor Measurement Unit (PMU) data. The proposed method uses a horizon-window that reconstructs the frequency input time-series data in order to predict the frequency features such as Nadir. Simulated scenarios are based on the gradual inertia reduction by including non-synchronous generation into the Nordic 32 test system, whereas the PMU collected data is taken from different locations in the Nordic Power System (NPS). Several horizon-windows are experimented in order to observe an adequate margin of prediction. Scenarios considering noisy signals are also evaluated in order to provide a robustness index of predictability. Results show the proper performance of the method and the adequate level of prediction based on the Root Mean Squared Error (RMSE) index.

This study presents the impact on power system frequency control in small power systems based on different generator topologies with a large penetration (50%) of variable speed wind turbines. The impact of a proposed controller is investigated versus various wind speeds. In particular, wind speeds with an average wind speed just below rated wind speed proves to cause the worst frequency fluctuations regardless of the type of backup generation topology investigated during 50% wind‐penetration ratio. For this wind‐speed session and a hydro‐based system, the proposed control system improves the frequency duration outside of the specified range [49.9, 50.1]Hz from 81 to 53% while reducing its delivered energy by only 6%. Furthermore, the proposed control reduces the absolute value of requested reserve by 49% from the hydro unit responsible for primary frequency control.

This work presents a comparison of alternating current (AC) and direct current (DC) distribution systems for a residential building equipped with solar photovoltaic (PV) generation and battery storage. Using measured PV and load data from a residential building in Sweden, the study evaluated the annual losses, PV utilization, and energy savings of the two topologies. The analysis considered the load-dependent efficiency characteristics of power electronic converters (PECs) and battery storage to account for variations in operating conditions. The results show that DC distribution, coupled with PV generation and battery storage, offered significant loss savings due to lower conversion losses than the AC case. Assuming fixed efficiency for conversion gave a 34% yearly loss discrepancy compared with the case of implementing load-dependent losses. The results also highlight the effect on annual system losses of adding PV and battery storage of varying sizes. A yearly loss reduction of 15.8% was achieved with DC operation for the studied residential building when adding PV and battery storage. Additionally, the analysis of daily and seasonal variations in performance revealed under what circumstances DC could outperform AC and how the magnitude of the savings could vary with time.

This study investigates the frequency behaviour in the Nordic power system (NPS) utilising phasor measurement units (PMUs) from different locations in the system. Different metrics are selected or proposed to quantify the system frequency response in case of major disturbances and of daily normal operations. Based on the collected PMU data, there are on average 17.5 major, generator disconnection related, disturbances per year. The maximum absolute rate of change of frequency exceeds 0.1 Hz/s for 20% of these disturbances. The average frequency nadir falls to 49.7 Hz, and the average time to nadir is about 8.7 s. If wind turbines were to be requested to provide inertia support to the system, these four metrics indicate how often, how fast, for how much and for how long the wind turbines should act. Moreover, during daily normal operations, the system has in general longer duration of over‐frequency ( f > 50.1 Hz) than under‐frequency ( f < 49.9 Hz). These frequency deviations are observed to occur around the hourly clearing of the Nord Pool Spot market and, especially during the hours between 05:00 and 08:00. In order to improve the frequency quality, the Nordic system operators tested and implemented an automatic frequency restoration reserve (aFRR) in addition to the existing manual one. Based on a frequency evaluation, this aFRR is needed most during 4:45–08:15 and 22:00–23:30 o'clock in 2012 in the NPS.

This paper investigates and compares different methods of primary frequency control for wind power during high wind speed range. The paper also identifies the models of pitch regulated wind turbines in a power system with hydro power dominated. Short-term extra active power, inertia emulation and direct pitch angle regulation as a primary frequency control are evaluated when there is a power imbalance. In addition, the combinations of these three methods are evaluated. Simulation results show that the combination of temporary extra active power and inertia emulation method has the highest value of temporary minimum frequency. However it is slower than other proposed methods and has second frequency drop. The advantage of using direct pitch angle regulation separately or in combining with the two other methods is that, the rotor speed does not decrease below the value that it has before the power imbalance happens.

This study presents the modelling of transformers for voltage disturbances. The focus in this study is on the security of the power supply for sensitive loads, for instance, the auxiliary systems for a nuclear power station feeding, for instance, pumps for emergency systems. Both steady‐state operation as well as transient conditions have been analysed. A five‐limb as well as a three‐limb 4 kVA transformer was designed in order to verify theoretical assumptions with experimental results. The results show that the theoretical models provided results in conformity to the actual experiments with an average discrepancy of 3%. Moreover, it is shown that a five‐limb transformer gives lower inrush currents, especially in the highly saturated region. In addition, it was demonstrated that an inductive load gives much higher inrush currents compared with a resistive load. A final observation was that the current drawn by the two transformers in their saturated state differ severely depending on the topology of the core.