In view of the prediction of NO_x emissions from coal-fired power plants

the NO_x prediction model is proposed based on particle swarm optimization (PSO) algorithm and bidirectional gate recurrent unit (Bi-GRU).Principal component analysis (PCA) is used to reduce the dimension and eliminate coupling of operating parameters which affecting the NO_x emission.The nonlinear characteristics and time sequence information of NO_x emission are achieved through the Bi-GRU neural network

which is helpful to improve the ability of feature extraction.The PSO algorithm is applied to optimize the parameters of the Bi-GRU network model

so that the network structure can match the feature of NO_x data to the greatest extent

and the problem of low prediction accuracy caused by selecting parameters according to experience or manual adjustment can be overcome.Finally

based on the operation parameters of a 660 MW coal-fired boiler

the NO_x emission prediction model is established and compared with the single direction GRU model and Bi-GRU model.The simulation results show that

the PSO-Bi-GRU model has the best prediction performance

and the mean absolute error

mean absolute percentage error and root mean square error are the minimum

which verifies the effectiveness of the proposed algorithm.