Automated biomass recycling management system using modified grey wolf optimization with deep learning model. (February 2023)

• Record Type: Journal Article
• Authors: Althubiti, Sara A; Kumar Sen, Sanjay; Altaf Ahmed, Mohammed; Laxmi Lydia, E.; Alharbi, Meshal; alkhayyat, Ahmed; Gupta, Deepak
• Is Part Of: Sustainable energy technologies and assessments. Volume 55(2023)
• Publication Date: 2023-02
• Subjects: Biomass recycling -- Solid waste management -- Deep learning -- Computer vision -- Artificial intelligence -- Object detection; Renewable energy sources -- Periodicals; Energy development -- Technological innovations -- Periodicals; Electric power production -- Periodicals; Energy storage -- Periodicals; 333.79
• DOI: 10.1016/j.seta.2022.102936
• Languages: English
• ISSNs: 2213-1388

Abstract

Abstract: Biomass residues encompass non-recyclable municipal solid waste, crop wastes, sewage effluents and sludges, domestic and industrial greywater, etc. Numerous wastes to energy conversion technology use biomass to generate various kinds of renewable energy to reduce environmental issues. The recycling rate seems to rise continuously, but reports reveal that humans are creating more waste than before. Machine learning (ML) can be used that offers a structure to take as a structural enhancement of the fact without being programmed. This study proposes an automated biomass recycling management system using modified grey wolf optimization with deep learning (ABRM-MGWODL) model. The presented ABRM-MGWODL technique aims to effectually identify and categorize the waste objects to enable effectual biomass recycling. The ABRM-MGWODL method would follow 2 major processes: waste object detection and waste object classification. For the waste object recognition and detection process, the YOLO-v4 model is exploited in this work. Next, the graph convolution network (GCN) method can be used for classifying recognized waste objects. Finally, hyperparameter tuning of the GCN model is effectually carried out using the MGWO algorithm, thereby enhancing the ABRM-MGWODL method's classification outcome. A widespread set of simulations were performed to ensure the superior waste classification efficacy of the ABRM-MGWODL model. The simulation outcomes demonstrate the improvements of the ABRM-MGWODL method to other DL models with increased accuracy of 99.01%.