Abstract
Frequency prediction after a disturbance has received increasing research attention given its substantial value in providing a decision-making foundation in power system emergency control. With the advancing development of machine learning, analysis power systems with machine-learning methods has become completely different from traditional approaches. In this paper, an ensemble algorithm using cross-entropy as a combination strategy is presented to address the trade-off between prediction accuracy and calculation speed. The prediction difficulty caused by inadequate numbers of severe disturbance samples is also overcome by the ensemble model. In the proposed ensemble algorithm, base learners are selected following the principle of diversity, which guarantees the ensemble algorithm’s accuracy. Cross-entropy is applied to evaluate the fitting performance of the base learners and to set the weight coefficient in the ensemble algorithm. Subsequently, an online prediction model based on the algorithm is established that integrates training, prediction and updating. In the Western System Coordinating Council 9-bus (WSCC 9) system and the Institute of Electrical and Electronics Engineers 39-bus (IEEE 39) system, the algorithm is shown to significantly improve the prediction accuracy in both sample-rich and sample-poor situations, verifying the effectiveness and superiority of the proposed ensemble algorithm.
Highlights
With the construction of the alternating current (AC)/direct current (DC) hybrid power grid, grid operating characteristics have undergone fundamental changes
The trade-off between accuracy and calculation speed in the power system frequency prediction is resolved by the described ensemble algorithm
When the frequency prediction, in the case of inadequate samples, was investigated, it was shown that the proposed algorithm tended to be more accurate under those conditions
Summary
With the construction of the alternating current (AC)/direct current (DC) hybrid power grid, grid operating characteristics have undergone fundamental changes. Both methods involve a trade-off between calculation speed and precision Different from these two methods, machine learning, as a model-free method, is devoted to analyzing the numerical relevance among operation state variables and research targets using data science and computer science [4]. To solve the problem caused by the inadequacy of severe disturbance samples, an ensemble algorithm based on cross-entropy is proposed to predict the minimum frequency after a disturbance. The outputs of various base learners are merged to fully excavate the information contained in a sample and realize fast, reliable prediction of power system frequency in the event of sample scarcity Base learners, such as artificial neural networks (ANNs) and support vector regression (SVR), are applied to the fitting problem in several steps, including data acquisition, training and verification [23].
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have