Distributed Accelerated Optimization Algorithm And Application In Economic Dispatch And Resource Allocation Problem Of The Smart Grid

With the rapid development of renewable energy market economy in China,traditional power grid has gradually been replaced by the smart grid.The smart grid can improve the reliability,economy,and environmental friendliness of the power system,and achieve efficient use and sustainable development of energy.In practical applications,the smart grid needs to obtain the optimal generation and energy distribution plan in a shorter time to ensure the efficient and stable operation of the power system,which requires more efficient optimization algorithms.Compared with traditional optimization algorithms,accelerated optimization algorithms can converge to the optimal solution in a shorter time,and better meet the requirements of smart grid operation efficiency.Therefore,this paper focuses on the application of accelerated optimization algorithms in the economic dispatch and load allocation problems of the smart grid.(1)For the economic dispatch problem of the smart grid,this paper proposes a distributed optimization scheduling model.This model fully considers the interconnection of multi energy system.The aim of this model is minimizing the overall operating cost of the smart grid and maximizing social welfare.For this optimization problem,a finite-time distributed algorithm is designed based on distributed theory and sliding mode control theory.Compared with centralized methods,distributed methods have great advantages in privacy protection and data storage.Then,the algorithm is proven to converge to the optimal value through finite-time stability theory.Finally,the superiority of the proposed algorithm is verified through comparative experiments.Compared with the existing projection distributed algorithm,the proposed algorithm can converge to the optimal solution in a shorter time and effectively reduce the jitter.(2)For resource allocation in the smart grid,this paper proposes a fully distributed optimization and scheduling model.This model considers the application of distributed energy storage systems.Unlike typical load allocation problems,the cost function coefficients of the energy storage system are modeled as time-varying parameters.Then,a distributed time-varying algorithm based on predictive-correction method and finite-time consistency theory is designed.To further accelerate the convergence of the time-varying algorithm,the hardware structure of the algorithm is optimized based on parallel method using the xilinx vivado development platform and implemented in FPGA.Finally,the effectiveness of the proposed algorithm is verified through an energy storage system load allocation case.Then,a hardware circuit is built using model composer and simulink.Compared with the simulation results in matlab,the circuit designed by the circuit framework proposed in this paper can effectively track the optimal trajectory,and has a shorter convergence time compared to the CPU.