TimesNet: A algorithm for day-ahead forecast of dynamic carbon emission factors in power grids

Carbon emission factor is the core parameter of demand-side carbon reduction, which can be used as a carbon reduction signal to guide the user to achieve the goal of reducing carbon emissions by improving electricity consumption behavior. Because there are different periods of carbon emission factors in time dimension, this paper uses TimesNet to build a prediction model. TimesNet is a time series prediction algorithm based on multi-period, which expands the time series from one-dimensional space to two-dimensional space, can integrate the characteristics of different time periods. Finally, based on the data set of New York State in the United States in 2016(New York Set), we use carbon emission factors of 168 hours advance to predict that of the next 24 hours, and achieve MAE 0.00527, 0.00756, 0.01121 of node 5, node7, node9 in IEEE 9-bus system.