Industrial electricity consumption prediction method based on secondary decomposition

The invention belongs to the technical field of industrial electricity consumption prediction, and particularly relates to an industrial electricity consumption prediction method based on secondary decomposition. Comprising the following steps: preprocessing original industrial power consumption data to obtain an industrial power consumption time sequence; vMD is adopted to decompose the industrial electricity consumption time sequence, and a KLD value is used to determine the number of modes decomposed by the VMD; calculating a sample entropy of each IMF component obtained by VMD decomposition, and dividing the IMF components into a low-complexity sequence and a high-complexity sequence; summing the high-complexity sequences; performing decomposition on the summed sequence by adopting CEEMDAN to obtain a plurality of subsequences; screening the plurality of subsequences by adopting F test and MI values, and respectively inputting the screened subsequences and low-complexity sequences into an industrial elect