23 research outputs found
Real-Time Big Data Analytics with Computational Intelligence Approaches for Energy Load Forecasting
No full textIn a real-time scenario of load forecasting, it is crucial to determine the future electric energy consumption in power distribution electrical networks. The electric energy forecasting models need to be updated with real-time trends of energy consumption as the analyzed energy consumption data exhibits high variability between historical and current data. This work proposes a multi-stage supercomputing-based big data analytics service for parallel and real-time load forecasting. Moreover, theoretical and experimental perspectives are proposed for multi-core parallel short-term load forecasting. Additionally, the knowledge from existing load forecasting based on deep learning models is used to innovatively develop highly accurate transfer learning models at different distribution nodes. Transfer learning models present practical applicability and productive possibilities in cases when sufficiently large data is not available. A novel approach based on deep neural network models is employed for load forecasting. Firstly, the electrical distribution nodes are grouped into different clusters with an aim to decrease the number of deep learning models to be trained. Secondly, network architecture information, weights, and biases are transferred from the first developed clustered model to subsequent models with an aim to reduce the training time of a large number of clustered models. And incremental learning is employed to incorporate newer data points for real-time processing and improving the forecasting accuracy of the clustered models on individual distribution points. Furthermore, parallel pool-based processing is employed to make efficient utilization of computing cores and to reduce the model development time further. The proposed big data real-time analytics methodology is evaluated on real-world energy consumption data collected from 105,148 Spanish electrical distribution transformers. The proposed methodology aims to reduce the number of trained models, training time, and execution time while still maintaining high prediction accuracy
Integration of Multisensor data and Deep Learning for realtime Occupancy Detection for Building Environment Control Strategies
No full text
A Review of Wind Energy Conversion Systems
No full text© 2022 IEEE.In the last decade, wind energy as a renewable energy source has become increasingly popular, and the establishment of large-scale wind energy conversion systems (WECS) and its connection to the electricity grid has become common. However, conventional power systems are not directly compatible with the characteristics of wind turbines. In this article, different topologies and classification of wind turbine systems are examined and different wind energy conversion systems are discussed. The article focuses on the speed-based, output-based, generator-type-based and orientation-based classification of WECS. The typical structure and information of WECS are explained in detail. Fixed and variable WECS are compared and contrasted in the context of network stability. An overall review and comparison of different wind turbine generators in WECS are discussed. Finally, the balance problem for different wind turbine energy conversion systems in the grid network is presented and possible different mitigation methods and solutions are suggested
Privacy Preservation of Data-Driven Models in Smart Grids Using Homomorphic Encryption
No full textDeep learning models have been applied for varied electrical applications in smart grids with a high degree of reliability and accuracy. The development of deep learning models requires the historical data collected from several electric utilities during the training of the models. The lack of historical data for training and testing of developed models, considering security and privacy policy restrictions, is considered one of the greatest challenges to machine learning-based techniques. The paper proposes the use of homomorphic encryption, which enables the possibility of training the deep learning and classical machine learning models whilst preserving the privacy and security of the data. The proposed methodology is tested for applications of fault identification and localization, and load forecasting in smart grids. The results for fault localization show that the classification accuracy of the proposed privacy-preserving deep learning model while using homomorphic encryption is 97–98%, which is close to 98–99% classification accuracy of the model on plain data. Additionally, for load forecasting application, the results show that RMSE using the homomorphic encryption model is 0.0352 MWh while RMSE without application of encryption in modeling is around 0.0248 MWh
Short-term Power Forecasting Model Based on Dimensionality Reduction and Deep Learning Techniques for Smart Grid
No full text
Performance Evaluation of Distributed Machine Learning for Load Forecasting in Smart Grids
No full text
