Aiming to solve the problem of low utilization and poor economic efficiency of coal tar, this study proposes a new method to convert coal tar into syngas that can be used for coal-to-oil production through chemical looping gasification reactions, thus reducing the cost of coal-to-oil production. This proposal converts tar to syngas for Fischer-Tropsch reaction through chemical looping gasification cleaning process, which is realized by process simulation and artificial intelligence prediction to ensure the stable products quality under feed jump abnormal conditions. The proposed process model is first built and the optimal operating parameters are determined by sensitivity analysis after the optimization sequence is determined from correlation analysis. A dynamic model based on the optimal steady-state model is second generated to design an optimal control scheme to efficiently deal with the unusual conditions of the feed step. The data-driven deep learning model LSTM is used to intelligently screen the optimal proportional control scheme with the advantage of low time and cost consumption. Finally, the optimal control scheme is obtained by combining the prediction results of the LSTM model and the controller additions of other important parameters, and the control effect is tested by simulating the tar feed leap conditions proves that the control6 scheme is highly resistant to disturbances. The method achieves 84.1 % high carbon conversion rate of medium-low temperature tar and 88.14 % excellent selectivity of syngas. At the same time, the adjusting of feed step condition with long-term and short-term memory network promotes the development of process design with the low time and cost consumption.
Read full abstract