Abstract: The smart grid is not only the need of the development of grid technology itself, but also shines the light on the road of the economic and energy development for various countries. As a prominent representative of clean energy, the distributed generation system of renewable energy is the inevitable trend of smart grid development. However, the high uncertainty and intermittence of renewable energy due to weather and other uncertain factors pose great challenges to the stability of smart grid systems. In order to meet this challenge and give full play to the cost advantage of renewable energy, the decision-making problem of energy procurement in the many-to-many energy network supply chain with renewable energy supply is studied from the perspective of micro-operation, so as to obtain the optimal energy procurement amount for energy integrators, as well as the impacts of the energy shortage cost, power plant capacity and renewable energy on the optimal solution.
An energy supply chain network composed of distributed micro-grid is considered. The energy supply chain network consists of multiple power plants and smart grids. Each smart grid is equipped with renewable energy, such as wind turbines or solar panels. In the smart grid, energy integrators are responsible for collecting users' demands and deciding to purchase the total power of traditional energy from power plants. If the uncertainty of renewable energy output leads to the imbalance of energy supply and demand, it is necessary for integrators to temporarily meet the demand of surplus power through the auxiliary power market, but occurring very high penalty cost. With the goal of minimizing the total cost, the integrator makes the optimal electricity purchase decision based on the power plant price and capacity, user demand and the prediction of renewable energy distribution. Game theory in the proposed supply chain network is used. The decision process is as follows. First, the power plants decide the electricity prices. Second, the integrators decide procurement quantities from each power plants. Third, the renewable energy output is realized. Finally, the integrators use purchased traditional power and renewable energy to satisfy smart grid users' demand, penalty cost occurs if there are any unsatisfied demands.
Two research circumstances are considered. The first one is when the renewable energy can be perfectly controlled. The second one is when the renewable energy cannot be controlled, but only can be predicted based on its distribution function. It is found that the power generation capacity of the power plant has a direct impact on the strategy of energy integrators to use renewable energy. Moreover, energy integrators with larger renewable energy production capacity will increase their energy procurement amount even more to avoid the risk of uncertainty.
Numerical studies are conducted to show our results. In the perfect case, the distribution proportion of the power plants' capacity is only related to the penalty cost. It can be seen that if the smart grid is located in a remote area or has other factors that cause large energy waste, the penalty cost will be larger, and the larger the distribution proportion, that is, the higher the priority of energy access. However, in the uncertainty case, it is found that if integrators further invest in renewable energy capacity and improve the ability of energy self-sufficiency, the proportion of traditional energy from power plants should be reduced. However, the opposite conclusion is found. When the integrator's own renewable energy capacity increases, it will increase the traditional energy proportion from the power plants. The conclusions of this paper can help smart grid energy integrators to save electricity costs, and provide some reference for realizing economical, reliable and stable power grid system.

Key words: smart grid, energy supply chain, energy network, energy procurement