Abstract:In order to identify power production and demand in realtime for efficient and dependable management for diverse renewable energy systems, precise and intuitive renewable energy predictions are required. Deep learning can be exploited to handle a variety of operations and maintenance improvement challenges, as well as develop better methods and perspectives for medium- and long-term energy prediction. This paper provides a detailed literature and bibliometric review of deep learning models for effective renewable energy forecasting. To begin, data was gathered via the Web of Science (WoS) library to access a large amount of articles and journals. In WoS, a total of 276 publications were extracted, including five different types, including Articles (261), Reviews (13), Early Access (10), Proceedings Paper (5), Editorial Material (2), and Data Paper (1). Then, literature statistics analyzed top 10 productive country, author, institution, journal. Overall keyword analysis are explored and discussed from various aspects like most frequency keywords, keywords analysis by year, keywords co-occurrence analysis, keywords co-occurrence graph, topic evolution-accumulation, topic evolution weighted. In addition, literature statistics of renewable energy are evaluated for wind energy, solar energy, ocean energy, hydrogen energy. Deep learning models can be leveraged to anticipate underneath a variety of uncertainty arising from renewable energy sources that are fluctuating. Furthermore, the estimated prediction accuracy requirements are given, and the keywords analysis of the deep learning forecasting models iare demonstrated from the perspectives of SAE (Stacked AutoEncoder), DBN(Deep Belief Network), CNN(Convolutional Neural Networks), GAN(Generative Adversarial Networks), and RNN(Recurrent Neural Network). Due to various shifting weather conditions as well as other variables, the forecasting model responds differently by various types of datasets. Deep learning methods offer intriguing potential discoveries in the field of energy forecasting. The relevant aspects and suggestions for future research were highlighted in the conclusion to conquer the projected barriers.

Deep learning versus traditional machine learning methods for aggregated energy demand prediction

Using Bayesian Deep Learning to Capture Uncertainty for Residential Net Load Forecasting

Evaluation of Deep Learning-based prediction models in Microgrids

Data-Driven Techniques for Short-Term Electricity Price Forecasting through Novel Deep Learning Approaches with Attention Mechanisms

Residential energy consumption forecasting using deep learning models

Advances in Deep Learning Techniques for Short-term Energy Load Forecasting Applications: A Review

Deep Learning for Energy Time-Series Analysis and Forecasting

Application of Deep Learning Long Short-Term Memory in Energy Demand Forecasting

Dynamic energy prices for residential users based on Deep Learning prediction models of consumption and renewable generation

A deep learning model for short-term power load and probability density forecasting

A comparative assessment of deep learning models for day-ahead load forecasting: Investigating key accuracy drivers

Deep Learning for Energy Markets

Deep learning based predictive analysis of energy consumption for smart homes

Deep and Machine Learning Models to Forecast Photovoltaic Power Generation

Enhancing Electrical Power Demand Prediction Using LSTM-Based Deep Learning Models for Local Energy Communities

Deep Learning for Short-term Instant Energy Consumption Forecasting in the Manufacturing Sector

Renewable energy power generation forecasting using deep learning method

Advanced Ensemble Methods Using Machine Learning and Deep Learning for One-Day-Ahead Forecasts of Electric Energy Production in Wind Farms

Highly-Accurate Electricity Load Estimation via Knowledge Aggregation

Deep learning for renewable energy forecasting: A taxonomy, and systematic literature review

Machine and deep learning approaches for forecasting electricity price and energy load assessment on real datasets