TimeGPT in Load Forecasting: A Large Time Series Model Perspective

Machine learning models have made significant progress in load forecasting, but their forecast accuracy is limited in cases where historical load data is scarce. Inspired by the outstanding performance of large language models (LLMs) in computer vision and natural language processing, this paper aims to discuss the potential of large time series models in load forecasting with scarce historical data. Specifically, the large time series model is constructed as a time series generative pre-trained transformer (TimeGPT), which is trained on massive and diverse time series datasets consisting of 100 billion data points (e.g., finance, transportation, banking, web traffic, weather, energy, healthcare, etc.). Then, the scarce historical load data is used to fine-tune the TimeGPT, which helps it to adapt to the data distribution and characteristics associated with load forecasting. Simulation results show that TimeGPT outperforms the benchmarks (e.g., popular machine learning models and statistical models) for load forecasting on several real datasets with scarce training samples, particularly for short look-ahead times. However, it cannot be guaranteed that TimeGPT is always superior to benchmarks for load forecasting with scarce data, since the performance of TimeGPT may be affected by the distribution differences between the load data and the training data. In practical applications, we can divide the historical data into a training set and a validation set, and then use the validation set loss to decide whether TimeGPT is the best choice for a specific dataset.

Power Response and Modelling Aspects of Power Electronic Loads in Case of Voltage Drops

In this paper, the power response of power electronic loads in case of voltage drops are measured and their dynamics are analysed. Based on this, dynamic simulation models are derived which can be used for voltage stability investigations. For this, four loads with different power factor techniques are considered. In addition, the impact of the grid impedance and input filter on the power response are measured in the laboratory. Based on the measurements, the simulation models are described. It is also outlined under which aspects the components of the loads are normally dimensioned if no detailed information is available. A comparison with the measurements demonstrates that the simulation models capture the main dynamics. At the end of the paper, the load models are compared to a constant power load in a short-term voltage stability use case. The results indicate that the power electronic loads have a more positive influence on short-term voltage stability in case of voltage drops. Overall, the contributions of the paper are the identification of the basic power dynamics of power electronic loads for different voltage drops and a subsequent derivation of suitable simulation load models for voltage stability investigations.