Study On The Influencing Factors And Scenario Prediction Of Industrial Carbon Emissiom In Nanchang City Based On AOA-LSTM Model

As a major emitter of carbon dioxide,we hope that China can make more contribution to reducing emissions in the future.By studying the influencing factors of industrial carbon emissions in Nanchang and the direction and degree of each influencing factor,and exploring the future carbon emission trend of industrial carmissions under different scenarios,the government and enterprises can develop more effective carbon emission reduction policies and make contributions to the realization of the goal of reaching the peak before 2030.In this thesis,based on previous studies,five influencing factors of carbon dioxide emissions are selected,and LMDI method is adopted to decompose the influencing factors of carbon emissions of the whole industrial industry and subdivided industries.In order to make up for the inadequacy of LMDI method in measuring elasticity,STIRPAT and PLS were used to investigate the relationship between industrial carbon emission and its influencing factors in Nanchang.In the part of prediction model construction,since the traditional prediction method is difficult to accurately describe the nonlinear relationship in the carbon emission system,LSTM can solve the nonlinear problem well.However,the selection of superparameters is highly subjective,which requires repeated fitting to find the appropriate superparameters.In order to improve the efficiency of parameter optimization and the accuracy of model prediction,this thesis combines the parameter optimization of AOA with the nonlinear information extraction ability of LSTM model,and proposes an AOA-LSTM model to capture the relationship between the time series of influencing factors of energy consumption carbon emissions,so as to achieve carbon emission prediction.Finally,the AAO-LSTM model is combined with scenario analysis to predict the predicted carbon emission and the time of carbon peak under three scenarios during 2021-2035.The results show that:(1)From the analysis of influencing factors,energy structure,labor productivity,investment scale and industrial working population scale can promote the growth of industrial carbon emissions in Nanchang City,while energy intensity,research and development intensity and industrial structure factors play an inhibitory role in industrial carbon emissions in Nanchang City.(2)In the prediction of carbon emissions,MAE,RMSE and MAPE of the constructed AOA-LSTM model are 20.14,14.98 and 0.96%,respectively,which has higher prediction accuracy than LSTM,SSA-LSTM,SOA-LSTM and PLS.In the scenario analysis and prediction analysis,the total industrial carbon emissions of Nanchang City showed a gradual upward trend under the high-carbon scenario,while the baseline scenario and lowcarbon scenario showed an inverted U-shaped growth.The base scenario will peak in 2029,while the low-carbon scenario will peak three years earlier,in 2026.The carbon peak in the low-carbon scenario is 4.3 million tonnes lower than in the base scenario,meaning that the lowcarbon scenario will reduce carbon emissions by more in the same amount of time.In view of this,efforts should be made to improve energy efficiency and promote research and development of advanced and low-carbon technologies.These efforts will help reduce carbon emissions and contribute to achieving the carbon peak target.