A coupled artificial neural network with artificial rabbits optimizer for predicting water productivity of different designs of solar stills. (January 2023)

• Record Type: Journal Article
• Authors: Alsaiari, Abdulmohsen O.; Moustafa, Essam B.; Alhumade, Hesham; Abulkhair, Hani; Elsheikh, Ammar
• Is Part Of: Advances in engineering software. Volume 175(2023)
• Publication Date: 2023-01
• Subjects: Solar still -- Productivity prediction -- Industrial development -- Artificial intelligence -- Multi-layer perceptrons -- Artificial rabbits optimizer; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2022.103315
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: A coupled multi-layer perceptrons model with artificial Rabbits optimizer is developed. The model was employed to predict the water productivity of different designs of solar stills. The investigated solar stills are conventional, stepped, pyramid, and tubular. Artificial Rabbits optimizer outperformed other optimizers. Abstract: In this study, a coupled multi-layer perceptrons (MLP) model with an artificial rabbits optimizer (ARO) is developed to predict the water productivity of different designs of solar stills (SSs). The investigated SSs are conventional, stepped, pyramid, and tubular SSs. The accuracy of the developed model was compared with three other MLP models optimized with conventional optimizers, namely gradient descent algorithm, particle swarm optimizer (PSO), and genetic algorithm (GA) optimizer. All models were trained and tested using the experimental data of the four SSs under the meteorological conditions of different cities. The performance of all models was evaluated using different statistical measures for all investigated cases. The MLP-ARO outperformed the other three models. The computed root mean square deviation values for all SSs' designs are 130.79, 57.07, 40.28, and 2.82 ml for MLP, MLP-GA, MLP-PSO, and MLP-ARO models, respectively.