Real-time Pricing Strategies For Smart Grid Based On Demand Response

Smart grid is widely regarded as the next electricity system,which incorporates advanced modern information and communication technology and smart metering infrastructure.It provides higher efficiency,reliability and environmental protection in power generation,transmission,distribution,consumption and management.Demand side management(DSM)is one of the most important features in smart grid.By taking effective measures,it can mobilize users to participate in power operation,guide them to optimize consumption and improve power efficiency.Demand response is an important part of DSM.Through direct compensation or adjustment of electricity price,it guides users to transfer or reduce load actively,so as to achieve the purpose of saving and efficient consumption and promote the safety and stable operation of power grid.Price response is the core of demand response,and it is an effective means to guide users to transfer or reduce demand autonomously.Real-time pricing(RTP)is the most ideal pricing mechanism for smart grid in the future.The research on RTP strategy based on demand response can not only enrich the power distribution and consumption optimization and control theory for smart grid,but also provide a strong basis for the formulation of electricity price and promote the implementation of RTP mechanism in practice.Although the research on RTP strategy has been widely carried out and achieved a series of valuable results,there are still some problems.For example,the research in the complex power systems such as multiple suppliers,multiple users and renewable energy supply is not comprehensive.Therefore,it is of great significance to further strengthen the RTP research.This dissertation studies the RTP problem in situations such as single supplier,multi suppliers and multi-energy generation.Combined with various theories,such as optimization theory,smoothing method,bi-level programming,Markov decision process(MDP)and reinforcement learning,RTP models reflecting the actual situation are established,and operable distributed algorithms with the characteristics of real-time information interaction are given.The main work of the dissertation is as follows:(1)The power generation from renewable energy on demand side is considered,and appliances are classified.In the case of uncertain power consumption of users,the RTP strategy is studied.For the purpose of maximizing the total utility of users and minimizing the cost of energy supplier,an expectation maximization model of social welfare in multitime slots is established using the relationship between power consumptions in adjacent time intervals.As solving the model,firstly,the objective function of uncertainty problem is transformed into a deterministic non-smooth optimization.Then,after smoothing the constraints,dual method is used to solve the problem,and the real-time price is obtained.The proposed model can provide RTP strategy effectively,guide users to actively and reasonably participate in the operation of power grid,and well realize resource allocation,peak shaving and valley filling.(2)Based on the social welfare maximization model,the RTP strategy is studied in the case of multiple suppliers.According to the working characteristics,appliances are divided into three categories.The operation of elastic and semi-elastic appliances are analyzed.It is pointed out that the power consumption of the two kinds of appliances has the coupling property about time,and a multi-time slots model is established.This model is decomposed into a set of single-time slot optimization problems by the relaxation method.Based on the theory of duality,a distributed algorithm is proposed.The real-time electricity price is obtained.This algorithm not only has the advantage of solving the single-time slot optimization problem,but also takes into account the global nature of the multi-time slots model.It provides an effective method for solving multi-time slots social welfare maximization models.(3)Considering the interests of both supply and demand,the RTP strategy different from the social welfare maximization method is studied from the global perspective.A holistic model focusing on the interactions between users and the power supplier is proposed in the framework of MDP.The MDP presentation of appliances’ operational processes well embodies their characteristics and the energy correlation of adjacent time slots.A novel distributed online algorithm based on a reinforcement learning approach is proposed to solve the MDP model without acquisition of the transition probabilities.Realtime electricity price is decided adaptively.The proposed model and algorithm balance energy supply and demand well,and have a good performance in peak shaving and valley filling.The approach not only guarantees that both users and the supplier all benefit,but also improves the social welfare of the whole grid.(4)Considering the random generation of renewable energy on the power supply side,the RTP strategy is studied in a multi-energy generation system.Focusing on the interaction among the power plants,the power market scheduling center and users,a bilevel stochastic model for RTP in the framework of MDP is formulated.Without loss of generality,carbon emission trading,classification of appliances,small-sale distributed energy generation,and power storage systems are also considered in the model.To solve the MDP model without acquisition of the transition probabilities,a reinforcement learning algorithm is exploited and a distributed online multi-agent learning algorithm is proposed.With the aid of the interactions between the upper and lower layers,this algorithm adaptively provides the optimal real-time prices,the assignment plan for every plant,the power scheduling and distributed energy production scheme for each user.The proposed model and distributed algorithm take the stochastic generation on the supply side into consideration in the RTP study,which fills the research gap.Through the above research,the results show that an effective RTP mechanism can well mobilize users to participate in demand response,improve power efficiency,realize peak shaving and valley filling,and adjust the balance of power supply and demand.It is expected that the dissertation can provide a reference for the follow-up research of the RTP strategy based on demand response in smart grid.