A Discrete Gradient Stepwise Optimization Algorithm for Medium- and Long-Term Dispatching of Reservoir Groups Considering HVDC Constraints

Abstract

The invention relates to the field of hydropower dispatching operation, relates to a discrete gradient stepwise optimization algorithm for medium- and long-term dispatching of reservoir groups considering the constraints of direct current transmission, and is a new optimization algorithm for solving the dimension disaster problem in reservoir dispatching. On the basis of the traditional gradient and its corresponding gradient descent method, the present invention proposes a discrete gradient calculation method and a corresponding discrete gradient descent method, combines it with the POA two-stage sub-optimization problem, and proposes a discrete gradient step-by-step optimization algorithm DGPOA. This method can use the information of the discrete gradient to determine the optimal search direction under a given discrete step without direct derivation, and use the line search iterative equation for iterative solution, which can significantly reduce the calculation scale and quickly obtain the optimization result. This rule has important guiding significance for power generation dispatching of cascade reservoir groups.

Description

technical field [0001] The invention relates to the field of hydropower dispatching operation, relates to a discrete gradient stepwise optimization algorithm for medium- and long-term dispatching of reservoir groups considering the constraints of direct current transmission, and is a new optimization algorithm for solving the dimension disaster problem in reservoir dispatching. Background technique [0002] By the end of 2016, my country's hydropower installed capacity and annual power generation had exceeded 300 million kW and 1 trillion kWh, accounting for 20.9% and 19.4% of the total energy respectively. Hydropower has become the largest renewable energy in China. Making full use of hydropower resources and optimal dispatching of reservoir groups is an important measure to build a clean, low-carbon, safe and efficient modern energy system. The optimal dispatching of cascade hydropower station groups operating in modern AC / DC large power grids is a typical high-dimensional...

AI Technical Summary

Problems solved by technology

When solving two-stage sub-problems, POA uses a grid search method with a given discrete step length to ensure global convergence in solving sub-problems, but with the increase in scale, there is still a serious "curse of dimensionality" problem

Images