Multi-parameter co-optimization for NOx emissions control from waste incinerators based on data-driven model and improved particle swarm optimization

The waste incineration process is complex and variable, posing a great challenge for precise NOx emissions control based on selective non-catalytic reduction (SNCR). Complex changes in the operating conditions of waste incinerators have led to poor economics and low stability of NOx emissions control. Regarding SNCR denitrification technology, the combustion and denitrification processes are deeply coupled and interact with each other. The precise and economical control of NOx emissions could be achieved by optimizing combustion and denitrification parameters. Therefore, we propose a combustion (air flow) and denitrification (ammonia flow) parameters co-optimization method to achieve safe, economic, and environmentally friendly control of NOx emissions from waste incineration processes. This method encompasses the following three parts: (1) a sparse autoencoding bidirectional long and short-term memory neural network (SAE-(Bi-LSTM)) model is applied to predict combustion (main steam flow and average temperature) and emissions (NOx) status. (2) Opposition- based learning and population decision-making based on Metropolis acceptance criterion strategies are introduced into the particle swarm optimization (PSO) algorithm to enhance its global optimization-seeking capability under complex operating conditions. (3) Combined the SAE-(Bi-LSTM) models with an improved PSO algorithm to investigate the co-optimization of air flow and ammonia flow. In the co-optimization case with a control target set at 120 mg/m3, 3 , the average NOx emissions increased from 116.4 mg/m3 3 to 119.81 mg/m3. 3 . Simultaneously, there is a significant reduction in the standard deviation of fluctuations, decreasing from 9.39 % to 1.35 %, and ammonia consumption decreased by 4.43 %. In sum, the proposed co-optimization method can control NOx near the target value while saving ammonia consumption.