Predicting Cyanide Degradability and Destruction Using Artificial Neural Networks: A Case Study in West Azerbaijan, Iran”

The presence of cyanide compounds in soil can lead to the formation and evaporation of hydrogen cyanide (HCN), while some of these compounds can undergo conversion by microorganisms present in the soil. However, high concentrations of cyanide can be toxic to soil microorganisms. During the gold extraction of mines, sodium cyanide (NaCN) solution is commonly used which results in a significant amount of cyanide waste that is regarded as an environmental pollutant. Previous studies have proposed physical, chemical, and biological methods to eliminate cyanide, but they have not achieved optimal efficiency. Unfortunately, the potential of artificial intelligence (AI) in this domain has been overlooked. Artificial Neural Network (ANN) models, specifically the multilayer perceptron (MLP) and simple linear regression, can significantly enhance and expedite the cyanide remediation process due to their remarkable predictive capabilities. In this study, an MLP and simple linear regression were employed to predict the biodegradation procedures of cyanide in soil. The study focused on cyanide waste generated by gold extraction factories and utilized environmental factors such as initial waste cyanide concentration, pH, soil and environmental temperature, humidity, chloride concentration, alkaline, electrical conductivity, precipitation, evaporation intensity, cyanide concentration, and initial pH as input data for the artificial neural network. The results of the MLP model revealed that electrical conductivity is the most influential factor in predicting the cyanide rate and initial pH of the waste soil. Conversely, the results of the simple linear regression indicated that the variables with the greatest impact on cyanide concentration are electrical conductivity (0.881), time (0.862), and alkalinity (0.724). By leveraging AI techniques such as ANN, this study demonstrates the potential for improved cyanide remediation. The integration of environmental factors and predictive models can contribute to more effective strategies for addressing cyanide pollution in soil.