Prediction of Carbon Emissions in Zhejiang Province Based on ATT-CNN-LSTM Model

Abstract

For the improvement of energy efficiency and reduce carbon emissions in Zhejiang province, this paper proposed a spatial carbon emissions prediction model based on industrial distribution and energy consumption levels in Zhejiang province. First, the grey correlation method is employed to analyze the impacts of economic level, population size, environmental growth and energy on carbon emissions. Secondly, the convolutional neural network (CNN) combined with the extended attention mechanism (ATT) is presented to capture the fine-grained features in the processed data, and then the long-short-term memory (LSTM) neural network is casted to evaluate the coarse-grained time series features hidden in the fine-grained data, thus establishing a spatial prediction model of power carbon emissions at the time and space scale to predict the power carbon emissions of Zhejiang province from 2022 to 2030 base on the Zhejiang development plan. The prediction results show that the proportion of clean energy should be increased, the energy structure should be optimized, and the industrial output value should be maintained. Moderate growth is an effective path for Zhejiang province to help achieve the overall goal of carbon peak by 2030.