Multi-objective Bi-level Research on Decision-making of Power Retailers Considering Dual Uncertainty and User Disutility
With the liberalization of market access for electricity retail,different kinds of power companies have entered the electricity market.The trading mechanism of electricity retail market has been a focus.A multi-objective bi-level stochastic optimization modelwith nonlinear constraints is developed to study the decision of electricity retailer considering the influence of disutility of users under demand response.The uncertainty of day-ahead pricing and renewable energy generation are also taken into account.The retailer'sprofit maximization and the users'cost minimization are defined as the upper-lay erand lower-layer problem of developed bi-level model,respectively.Then an improved GA algorithmis used to solve this model.The electricity sales of electricity retailers in each market are defined as individuals in the GA population,and the expected revenues of electricity retailers are used as the fitness functions.The Monte Carlo method is used to generate different scenarios,and simulation analysis is carried out under these different scenarios.Finally,the developed model is verified with real data from the US PJM market.It can be seen from the simulation results that after the implementation of DR,the decision model constructed in our paper can effectively balance the power load of users at various times of one day,and the use of the negative utility mechanism can effectively avoid peak-valley transfers and reduce the electricity cost of users.Energy storage equipment and renewable energy power generation equipment have significantly improved the revenue of electricity retailer,while the impact of uncertain factors such as price fluctuations in the spot marketand the randomness of renewable energy output can be alleviated.Specially,the users'negative utility cost function in the renewable energy power generation scenario is considered,which provides a new idea for studying users'electricity consumption behavior.
NETL
NSTL
万方数据
