Day-Ahead And Intraday Coordinated Optimal Dispatch For Power System Considering Stochasticity Of Source And Load

As the concern with environmental issues,energy crisis and climate change increases,wind power,a kind of green and environment-friendly renewable energy source with high development maturity,keeps growing penetration in the current power system.Gridconnected large-scale wind farms deliver much clean energy,but they further increase the uncertainty of power system in source side due to the strong intermittent and volatility.At the same time,the continuous growth of peak load demand and rapid development of flexible load such as electric vehicles further expand the randomness of power system in load side.The randomness of both wind power and load affects the dynamic balance and stability of the power system,which increases the difficulty of dispatch.Considering that the forecast accuracy of wind power and load increases with time scale refining,based on the idea of “gradual refinement and multi-level coordination”,the day-ahead and intraday coordinated optimal dispatch method for power system with source-load uncertainty is proposed in this dissertation.Firstly,a chance constrained programming model for day-ahead unit commitment and economic dispatch considering the source-load randomness is established.The interruptible flexible load is incorporated into the day-ahead dispatch plan to reduce the peak-to-valley difference of load through source-load interaction.The competitive particle swarm optimization algorithm is used to optimize the solution of the day-ahead unit commitment and generation schedule.A case study is carried out on the IEEE 30-node system,and the simulation results verify the effectiveness of the day-ahead dispatch strategy.Secondly,taking account of the stochasticity of source and load,dispatch properties of units and adjustment ability of AGC units,a ultra-short-term dynamic optimal dispatch model is established with the goal to minimize running cost of power system,which includes coal consumption cost of generation,output power adjustment cost of ultra-short-term plan dispatchable units and AGC units.Considering that the current dispatch scheme has an impact on the system operation of following periods,the ultrashort-term dynamic optimal dispatch problem is modeled as a discrete-time Markov decision process.The deep Q learning method is adopted to achieve an optimal dispatch strategy,which can dynamically optimize the output of the planned units for lower running cost based on the ultra-short-term forecast data of wind power and load.The simulation results show that the ultra-short-term optimal dispatch strategy can track the intraday fluctuation of source and load,relieve operating pressure of AGC units and effectively maintain dynamic balance of system power.