基于改进 LSTM 的袋式除尘器进口烟气温度预测

With the increasing awareness of environmental protection, high-temperature flue gas emissions have received extensive attention, and bag dust collectors have been used in an increasing number of industrial scenarios. In order to ensure the temperature environment of the bag dust collector, a modeling method of inlet flue gas temperature prediction based on improved particle swarm optimization (IPSO) and attention mechanism (AM) combined with a short term memory network (LSTM) was proposed. First, the field conditions were analyzed, key data variables were collected, and variables were normalized. Then, the AM-LSTM inlet flue gas temperature prediction model was established to improve the prediction accuracy of a single LSTM neural network model. Moreover, IPSO was used to optimize the AM-LSTM model's hyperparameters to improve the performance of the neural network. The simulation results show that compared with other models, the modeling method proposed in this paper has higher prediction accuracy and smaller error, and is more suitable for the actual inlet flue gas temperature change, which can lay a good foundation for the inlet flue gas temperature control of the dust collector and effectively extend the service life of the cloth bag.