Artificial neural network modelling of cyanide transport in a soil contaminated with cassava wastewater
Department of Chemical Engineering, University of Uyo, Uyo, Nigeria.
Research Article
World Journal of Advanced Research and Reviews, 2024, 23(01), 550–562
Publication history:
Received on 18 May 2024; revised on 03 July 2024; accepted on 06 July 2024
Abstract:
Understanding the transport of cyanide present in cassava wastewater is very crucial especially when choosing the right remediation technique to employ to restore the soil’s fertility. In this study, the concentrations of cyanide for both dry and wet seasons were determined on two sites of area (2.5 m by 2.5 m) respectively. Physico-chemical characterizations were carried out on the soil samples collected. Using the artificial neural network (ANN) approach for this study, a total of 91 samples were used for each season which were divided into 63 samples for training, 15 samples for validation, and 15 samples for testing. Three input parameters consisting of (depth, time, and measured concentration) were used to obtain one output parameter (predicted concentration). The performance of the ANN model was evaluated using metrics such as MSE, RMSE, R, and R2. The results showed that ANN model was able to predict the concentration of cyanide for both seasons as evident in their MSE and RMSE values close to zero and the R2 values close to unity. However, the data set for dry season had better ANN predictability than that of the wet season as shown by the R2 values for the training, validation and testing of the dry season to be 0.9954, 0.9978, and 0.9942, while the wet season is 0.9944, 0.7887 and 0.7793 respectively. Therefore, ANN model could be used to predict the concentration of cyanide at varying depths and times.
Keywords:
Cyanide; Cassava Wastewater; Artificial Neural Network (ANN); Prediction; Transport
Full text article in PDF:
Copyright information:
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0