State Grid Beijing Urban District Power Supply Company, Beijing, 100032, China.
Hubei Key Laboratory of Construction and Management in Hydropower Engineering, China Three Gorges University, Yichang, 443002, Hubei, China.
Sci Rep. 2023 Apr 5;13(1):5568. doi: 10.1038/s41598-023-31022-y.
Based on the counted power system emission factors of North China Power Grid, a community carbon emissions sample database is constructed. The support vector regression (SVR) model is trained to forecast the power carbon emissions, which is optimized by genetic algorithm (GA). A community carbon emission warning system is designed according the results. The dynamic emission coefficient curve of the power system is obtained by fitting the annual carbon emission coefficients. The time series SVR carbon emission prediction model is constructed, while the GA is improved to optimize its parameters. Taking Beijing Caochang Community as an example, a carbon emission sample database is generated based on the electricity consumption and emission coefficient curve to train and test the SVR model. The results show that the GA-SVR model fits well with the training set and the testing set, and the prediction accuracy of the testing set reaches 86%. In view of the training model in this paper, the carbon emission trend of community electricity consumption in the next month is predicted. The carbon emission warning system of the community is designed, and the specific strategy of community carbon emission reduction is proposed.
基于华北电网分机组碳排放因子统计数据，构建了社区碳排放样本库。利用遗传算法（GA）对支持向量回归（SVR）模型进行优化，设计了社区碳排放预警系统。通过拟合年度碳排放系数，得到了电力系统的动态排放系数曲线。构建了时间序列 SVR 碳排放预测模型，并对其参数进行了改进优化。以北京草厂社区为例，基于电量和排放系数曲线生成碳排放样本库，对 SVR 模型进行训练和测试。结果表明，GA-SVR 模型对训练集和测试集拟合效果较好，测试集的预测精度达到 86%。针对本文所建模型，对社区下月的电量消费碳排放趋势进行预测，设计了社区碳排放预警系统，并提出了社区碳减排的具体策略。
