Abstract:The world has embarked on a road to sustainable energy production. As a result, countries have turned to microgrid developments. This article aims to study the feasibility of renewable sources such as solar PV and wind power for integrating a microgrid campus, taking the example of a case in East Africa, precisely the case of the University of Djibouti. We applied the weather parameters to evaluate the solar and wind potential with the Decision Tree method for analyzing and classifying the degrees of solar radiation and the consistency of wind speed. These data are spread over eight years to establish and capture seasonal changes and prove the accessibility of renewable sources in a specific site. The results were compared to Random Forest, Logistic Regression, K-Nearest Neighbors, Support Vector Machine, and Naïve Bayes classifiers, which showed that the performance of classifying the Decision tree outperformed all other methods with an accuracy of 0.99321. The second work of this article explored the forecasting of the possible powers predicted with the LSTM deep learning method by the generation of the Solar PV array and wind turbines which were simulated on PVLib and Windpowerlib. The results are favorable, and the LSTM has performed well on the different hyperparameters. With the combination of machine learning and deep learning, it was possible to theoretically conclude the integration of renewable energies since we investigated all the potential possibilities in evaluating meteorological parameters and power predictions. Finally, decision scores from the Decision Tree architecture and the LSTM features were integrated to form a hybrid Tree-LSTM fusion method. It introduces a novel architectural concept that can effectively address sequential data and harness the non-linear capabilities of decision trees. The proposed model was validated by tuning the hyperparameters. Enhancing the maximum depth of the model increases the performance at a certain point, and conversely, reducing the minimum sample split improves the model performance. These contributions will help to create sustainable energy systems and increase the transition to a clean CO2 environment.

An Aggregated Decision Tree-Based Learner for Renewable Integration Prediction

Assessing the Feasibility of Integrating Renewable Energy: Decision Tree Analysis for Parameter Evaluation and LSTM Forecasting for Solar and Wind Power Generation in a Campus Microgrid

A Short-Term Photovoltaic Power Prediction Model Based on the Gradient Boost Decision Tree

An integrated method with adaptive decomposition and machine learning for renewable energy power generation forecasting

TransOS-ELM: A Short-term Photovoltaic Power Forecasting Method Based on Transferred Knowledge from Similar Days

Solar power generation prediction based on deep Learning

Short-term power prediction for photovoltaic power plants using a hybrid improved Kmeans-GRA-Elman model based on multivariate meteorological factors and historical power datasets

Deep and Machine Learning Models to Forecast Photovoltaic Power Generation

LTPNet Integration of Deep Learning and Environmental Decision Support Systems for Renewable Energy Demand Forecasting

A hybrid approach based machine learning models in electricity markets

Short-Term Photovoltaic Output Prediction Based on Decomposition and Reconstruction and XGBoost under Two Base Learners

Short-term PV power data prediction based on improved FCM with WTEEMD and adaptive weather weights

A novel GBDT-BiLSTM hybrid model on improving day-ahead photovoltaic prediction

Adaptive Encoder-Decoder Model Considering Spatio-Temporal Features for Short-Term Power Prediction of Distributed Photovoltaic Station

Prediction and aggregation of regional PV and wind generation based on neural computation and real measurements

Achieving wind power and photovoltaic power prediction: An intelligent prediction system based on a deep learning approach

Short-term photovoltaic power prediction based on modal reconstruction and hybrid deep learning model

Short-Term Power Prediction for Renewable Energy Using Hybrid Graph Convolutional Network and Long Short-Term Memory Approach

A Photovoltaic Power Prediction Approach Based on Data Decomposition and Stacked Deep Learning Model

Gradient boosting dendritic network for ultra-short-term PV power prediction