Substance Flow Analysis Combined with Neural Networks to Solve Potential Pollution Problems in Secondary Lead Industry

Abstract

The amount of waste lead acid battery (LAB) has increased to new levels. Therefore, the disposal of waste LAB will inevitably lead to more secondary pollutants (such as lead dust, lead slag and wastewater) polluting soil and underground water. The total emission of pollutants is affected by various key processes, including crushing, separation, pre-desulfurization, crude lead smelting, refining and slag making, etc. Therefore, only by determining and effectively monitoring the process with the largest pollution emission can pollution be fundamentally controlled. This article uses the substance flow method to analyze the whole material flow of waste LAB treatment, and find out the main pollution emission links that may occur in the process; then identify the characteristics of main pollutants; Finally, the artificial neural network (ANN) model optimized by genetic algorithm (GA) is introduced, and MATLAB software is used to realize the simulation of pollutant monitoring in key pollution processes. The results show that waste slag discharge is the main way for lead pollutants to enter the environment; the waste slag obtained by formal enterprises has a stable vitreous structure, and the lead leaching toxicity (0.3 mg/L) is significantly lower than the standard limit (5 mg/L), which also meets the preconditions for further reuse; the established GA-ANN model has high prediction accuracy (MSE=0.0003), which enables enterprises to dynamically monitor the lead content in lead slag in real time under the condition of fixed resource input, and timely adjust key input parameters to minimize pollution emissions.

Keywords: Secondary lead, Substance flow analysis, Artificial neural network, Genetic algorithm