wherein the power generation unit comprises a photovoltaic power generation system, a wind power generation system, and a tidal power generation system, and the tidal power generation system comprises a tidal reservoir and a tidal power generator set, wherein a water storage reservoir is provided at bottom of the tidal reservoir, and a water storage reservoir turbine, a water storage reservoir generator, and a water storage reservoir pump are provided between the tidal reservoir and the water storage reservoir; the water storage reservoir turbine is equipped with a valve, and the tidal power generation system, upon the closure of a water gate of the tidal reservoir during a rising tide, starts a water pump of the tidal reservoir to pump seawater from the rising tide into the tidal reservoir;

the energy storage unit comprises hydrogen storage and a battery pack, and hydrogen produced by the hydrogen storage serves as a feedstock for the fuel cell to generate electricity; power of the energy storage unit is optimized through an upper-layer model and a lower-layer model, wherein the upper-layer model decomposes remaining power of the power generation unit using a discrete Fourier transform on the basis that the power generation unit satisfies the load unit, obtaining storable power with a highest clean energy utilization rate; and the lower-layer model obtains cost of the energy storage unit through different power allocation modes of the hydrogen storage and the battery pack, and selects appropriate power for the hydrogen storage and the battery pack; the remaining power refers to the power that remains after the power generation unit has satisfied the load unit;

the load unit comprises a system load and a hydrogen production load for the hydrogen storage; and

the energy management system connects the power generation unit, the load unit, and the energy storage unit, and allocates the remaining power to the hydrogen storage and the battery pack;

the upper-layer model obtains a first frequency offset point for storable power and a battery, and the lower-layer model obtains a second frequency offset point for the battery pack and the hydrogen storage;

steps for determining the first frequency offset point and the second frequency offset point are as follows:

S1. performing the discrete Fourier transform on the remaining power of the power generation unit to determine the storable power with the highest clean energy utilization rate;

S2. determining a range for the first frequency offset point meeting a fluctuation requirement of the storable power, and selecting a first frequency offset point with the highest clean energy utilization rate;

S3. determining a range for the second frequency offset point based on the selected first frequency offset point;

S4. solving the upper-layer model and the lower-layer model of the energy storage unit using a genetic algorithm; and

S5. selecting a second frequency offset point with a lowest energy storage unit cost to determine power allocation for the hydrogen storage and the battery pack.