| In recent years,with the prosperity and stable development of social economy,people have put forward higher requirements for living standards,and the demand in all aspects has also risen,especially in energy.As the world’s largest energy consumer,energy security issues in China are crucial to the lifeline of national economic and social development.From the implementation of energy-saving,emission reduction,and energy structure transformation plans to the establishment of the "dual carbon" policy,it is sufficient to highlight the importance China attaches to energy.In the power industry,which is most closely related to energy,the output power distribution of each unit in the power generation side is not coordinated,and the lack of reasonable scheduling strategy leads to the increase of resource consumption and affects economic benefits.At the same time,the user side load demand is growing rapidly,which can no longer be satisfied only by the optimal scheduling of the generation side.Therefore,it is necessary to strengthen the power market regulation ability,take flexible load as the breakthrough point,reduce unnecessary load consumption,so as to ensure the balance between supply and demand of the system.In view of the above problems,this paper takes maintaining the stable operation of the power system as the fundamental,reducing the waste of resources and improving the efficiency of energy utilization and economic benefits as the goal,and achieves the desired effect through in-depth optimization of intelligent algorithm.The research content of this paper is as follows:Firstly,a dynamic economic dispatching model of power system is proposed to solve the problem of resource waste caused by large power output consumption of generating units.This model aims to improve the operation efficiency by optimizing the power size of each unit under the premise of stable operation of the system,so as to achieve the target of user demand at the lowest economic cost.Based on this,this paper innovates the traditional particle swarm optimization algorithm and proposes an improved parameter algorithm based on deep Q learning.By establishing a neural network,the algorithm can learn the relationship between the particle and the global optimal position online,dynamically adjust the weight of the parameter in the particle velocity updating formula,and indicate the direction for the particle to move to the global optimal position.At the same time,a new information sharing method is proposed to accelerate the speed of learning and information exchange between particles and avoid falling into local optimal.The efficiency of the proposed algorithm is verified by test function and model simulation.Secondly,in view of the huge challenges to the balance of supply and demand due to the sharp increase in the user side load demand,the smart grid technology and the power market price regulation mechanism are considered to adjust the flexible load of power users,optimize the power consumption strategy of users,save unnecessary load consumption,and alleviate the power shortage.On this basis,the dynamic demand response model of the electricity market is established,and the mechanism of setting retail electricity price is optimized to consider the profit of the power company and the cost of the power user,so as to form a win-win situation for both sides.For this model,this paper puts forward the deep deterministic strategy gradient algorithm to optimize the retail price formulation strategy.By taking advantage of its deep neural network of Actor-Critic mechanism and its advantages in dealing with continuous action and high dimension problems,it learns the retail price through interaction with the electricity market environment and works out the optimal dynamic retail price strategy.Compared with Q-learning algorithm,the proposed algorithm has outstanding performance in realizing load consumption reduction and ensuring stable operation of the system. |
PDF Full Text Request