Search

Due to the limited availability of historical data, forecasting the wind power of newly-built wind farms poses a significant challenge. Transfer learning methods offer a potential solution by transferring knowledge from the source domain to the target domain, thereby reducing the data requirements for wind power forecasting. However, the difference in dataset distribution between the source and target domains causes domain shift. To address this issue, we propose a multi-domain adversarial network (MDAN). MDAN uses multi-domain datasets for adversarial learning, which maps numerical weather prediction (NWP) data to an adaptive latent space, thereby reducing domain shift. Additionally, we propose a data fusion-based incremental learning method to mitigate catastrophic forgetting. Through comprehensive case studies, MDAN provides accurate short-term wind power probabilistic forecasts in zero-shot learning. The incremental learning method enhances forecast accuracy in few-shot learning. Moreover, visualization analysis using t-stochastic neighbor embedding (t-SNE) shows that MDAN successfully reduces the domain shift between the source and target domains. • A multi-domain adversarial network is proposed for probabilistic forecasting of newly-built wind farms. • The attention-based feature extraction network is employed to map multi-domain data to an adaptive latent space. • The data fusion-based incremental learning method is employed to mitigate catastrophic forgetting. [ABSTRACT FROM AUTHOR]

The existence of ac and dc grid leads to a new concept of hybrid microgrid in which both grids are tied by an interlinking converter (ILC). Various control schemes of ILC have been discussed in literature but there is a lack of a generalized, autonomous control which considers the tie line capacity between ac and dc grids. In addition, the majority of schemes involve switching between control modes which increases power losses. To address these issues, this paper presents a generalized droop scheme of ILC for power sharing and control in a standalone hybrid ac/dc microgrid. A 3-dimensional plane is proposed where the axis correspond to dc voltage, ac frequency and converter power. The converter utilizes frequency droop at ac terminal and voltage droop at dc terminal so any change in voltage or frequency results in active power flow from an underloaded to overloaded grid. A corresponding mathematical model is also presented which determines the impact of load change on frequency and dc voltage. The proposed scheme enables autonomous power transfer in both directions without any form of communication. A comparison between the traditional normalization scheme and proposed scheme is also presented where it is observed that this method reduces the frequency deviations and transient response time, thereby giving a better performance. Both the simulations and experimental results confirm the validity of the proposed model. [ABSTRACT FROM AUTHOR]

Microgrid is becoming an attractive concept to meet the increasing demands for energy and deal with air pollutions. Distributed energy sources (DERs) in Microgrid are usually interfaced with the utility grid by inverters, so the characteristics of Microgrid stability are much different from that of a traditional grid. However, the classifications, guidelines, and analysis method of Microgrid stability are well behind of the Microgrid development. In this paper, a Microgrid stability classification methodology is proposed on the basis of the of Microgrid characteristics investigation, which considers the Microgrid operation mode, types of disturbance and time frame. Then a comprehensive review of the body of research on Microgrid stability is presented in order to identify and advance the field. Finally, some challenges and suggestions of Microgrid stability for further researches are discussed. [ABSTRACT FROM AUTHOR]

Abstract: Manganese extraction from a low-grade ore was investigated using cane molasses as a reducing agent in dilute sulfuric acid medium. The effects of concentrations of cane molasses and sulfuric acid, leaching temperature as well as reaction time were discussed. The results showed that high manganese recovery with low Fe and Al extraction yield could be obtained by analyzing the leaching efficiencies of Mn, Fe and Al during the leaching process. The optimal leaching condition was determined as 1.9 mol/L H2SO4 and 60.0 g/L cane molasses for 120 min at 90 °C while using particles smaller than 0.147 mm. The leaching efficiencies were 97.0% for Mn, whereas 21.5% for Al and 32.4% for Fe, respectively. [Copyright &y& Elsevier]