Research On Economic Dispatch Of Power System For Improving Renewable Energy Utilization

In recent years, the energy crisis and environmental problems have become the important factors that restrict the sustainable development of economy and society in the world. Renewable energy sources, such as wind energy, are advantageous over traditional thermal generation both economically and environmentally, and it is an inevitable trend that power supplies need to rely more and more on renewables. However, wind energy is stochastic and fluctuating in nature, which needs corresponding peaking units to stabilize wind power fluctuations, and at the present power structure of our country is in lack of peaking units with strong load changing ability, which makes the grid cannot be accommodating large amount of wind power. The system wind utilization is quite low and much of the wind power is curtailed. China is rich in coal, but in lack of oil, gas and the water resources. This situation gives us no choice but to relay on the coal-fired generator with fast and deep ramping capabilities. Through the advanced control means to make use of the heat in thermal power units, is an effective way to improve the fast and deep ramping capabilities of the thermal units. Then the system wind power utilization can be increased. How to make full use of the units with heat storage in power system, and use these units to improve wind power utilization are important and valuable research topics.While the above heat storage generators are helpful in system dispatch to improve wind power utilization, using heat storage requires additional coal consumption in controlling the cooling water pump. Therefore, generator costs in fast mode become higher than regular, and are dependent on the actual ramp rate of units. Also, ramp rate limits of fast mode generators are not constant, but changing with the output levels. In existing economic dispatch models, generator costs are fuel costs depending only on unit output levels, and unit ramp rate limits are usually constants. It is thus not applicable to use existing models for systems containing fast mode generators.To cope with the above problems, this paper carried out three aspects of research work. Firstly, we analyze the impacts of thermal generation ramping on the maximum level of wind penetration in the security constrained economic dispatch problem. A stochastic formulation incorporating both thermal and wind generators is presented, and network security constraints of power balance and transmission limits are considered. Analyses are made for different values of ramping rates and output ranges of thermal generators, and under different decision horizons and intervals by using practical wind power data.Secondly, a new formulation of dynamic economic dispatch for wind-thermal power systems is presented, and ramping capabilities and costs of generators by using their heat storage are payed attention to. In this model, the objective is to minimize a two-variable quadratic generator cost function depending on both output levels and ramp rates, and generator ramp rate limits change with output levels. The formulation has linear constraints and is solved by using existing quadratic programming methods. In numerical examples, parameter fitting for the two-variable cost function and ramp rate limits of a 600MW thermal unit is presented, and this model is tested on the IEEE 30-bus system by using CPLEX. Results show that compared with regular units, heat storage units lead to substantial savings of curtailed wind energy and lower generator costs in system dispatch.Last but not least, the "memory" feature of heat storage units is taken into account, which means the greater ramping ability cannot be used continuously. A mixed integer linear formulation for the dispatch of wind-thermal power system containing fast ramping units with heat storage is presented. The fast ramping from heat storage are not always available, and as a simplified model, it is assumed that heat storage cannot be used in two consecutive decision periods. Such constraints involve multiplications of 0-1 binary variables and continuous variables and are thus non-linear. In order to facilitate the solution procedure, a big enough constant M is used to linearize the constraints of ramping capacities and a rolling-window method is proposed to accelerate the solving procedure.