The Study Of Load Dispatch For Smart Distribution Network Based On Flexible Load

With the development of renewable energy and the imbalance of regional economic, balance problem of active power is getting worse. Power balance were relaying on conventional thermal power units. Haze problems and shortages of fossil energy reduce economic and social benefits of that approach. The development of battery technology and networking technology make it possible to find the potential to provide grid ancillary services from the demand-side, and electric vehicle and the air conditioning load are representatives of these technologies.According on the research, large-scale EVs and air conditionings will start to run at peak time if they are not under control. This situation will increase the peak-valley difference and the pressure of frequency regulation. Moreover the starting-time law is similar between the EV and air conditioning, and the wind power has the character of inverse regulation. So the imbalance of power is worse. To this end, we should make the charging of EVs and the load of air conditioning under control and minimize the imbalance of power created by the loads and wind power and also we should ensure user comfort at the same time. Specific work is as follows:First, this paper analyzes the working principle of a single load and load characteristics of EV charging and air conditioning load. Then according to the law of using this kind of load, the paper models the statistical power using Monte Carlo random sampling and analyzes the effect of this two kinds of load on the conventional loadSecondly, this paper models the dispatching process of flexible load and renewable energy. The definition of the flexible load is given and then the fluctuation of renewable energy analyzed. This paper use the method of singular value decomposition (SVD) to separate renewable energy generation into two parts according to the size of the fluctuations. The large part is made as the target curve of flexible loads, and the task of this model is to minimize the tracking error according to the constraints. Then this paper describes the particle swarm algorithm, and a measure to improve the shortcoming of the algorithm is raised. A simulation is made to prove the effectiveness of this model and the idea.Finally, a model is made to illustrate the process of ordered charging. The control object is single EV. The object is to minimize the error between the summation of the charging loads and the plan. The solution is the CPLEX algorithm. By analyzing the results, it is proved that this model is feasible to complete the plan under the premise that the comfort of users is met. The fluctuation of the renewable energy is stabilized indirectly.