Research On Optimal Dispatching Model Of Wind Power Integrated Power System And Its Solving Algorithms

With a large number of wind farms connected to the power grid,the wind power optimal dispatching is one of the most popular research fields in recent years.Due to the random and intermittent characteristics,wind power is less controllable than the conventional power sources.The optimal dispatching of wind power integrated power system becomes a complex constrained optimization problem,which not only needs to deal with the uncertainty of wind power,but also consider the benefits of other power sources and numerous physical constraints.Since the traditional power dispatching method is based on deterministic power sources,it cannot handle the wind power uncertainties.In order to use wind power rationally and ensure the safety and stability of the power system,it is urgent to study new dispatching theories and solving methods.The optimal dispatching including wind power should solve two major problems:how to establish the dispatching model and how to solve the model.In this dissertation,the two problems are discussed and studied in detail,and some research results with theoretical significance and practical value have been obtained.The main research work and innovations are as follows:(1)Two improved particle swarm optimization(PSO)algorithms are proposed in this dissertation,which are suitable for complex optimal dispatching problems.Firstly,based on graph theory and information theory,a new analytical method for PSO topology is developed.Through the binarization of information carried by the particles,the algorithm performance is analyzed from the perspective of information exchange,and the main factors that affect the performance of PSO are clarified.Based on the analysis,an improved algorithm with dynamic topology is proposed.By changing the topologies,the information transfer rate among particles can be controlled.Numerical experiments and dispatching examples show that the improved algorithm has good search ability and reasonable convergence rate.Secondly,several improved methods for multi-objective particle swarm optimization are proposed.During the initializing phase,the uniform initialization is used to enhance the uniformity of the initial solution in search space.Meanwhile,a new non-dominated solutions maintenance strategy based on the minimum spanning tree theory is proposed to improve the uniformity of the Pareto solution set.In addition,the selection method of guided particle is also improved.Compared with the other algorithms,the improved multi-objective optimization algorithm has moderate convergence rate,good searching ability and better uniformity of non-dominant solutions.(2)To describe the wind power forecasting error,the non-parametric kernel density estimation is introduced.Since the optimal window width of non-parametric estimation is difficult to detennine,the particle swarm algorithm is applied to optimize the window width.The error data analyzation shows that the non-parametric estimation has good adaptability and higher degree of fitting goodness than the traditional parameter estimation methods.Based on the nonparametric error statistical model,a dynamic economic dispatching model including wind power is established.The conditional error probability distribution under the different forecast values is estimated by the kernel density estimation method,and the corresponding error confidence interval of Gauss kernel function is derived.By using the lower and upper limit of the error confidence intervals,the stochastic optimization model can be converted to deterministic interval optimization problem.The numerical results show that the proposed model has a better performance in error data fitting,which can avoid overvalued or undervalued of the forecast error,and the dispatching plan is more economical.(3)Based on predictive control theory,a new rolling dispatch model for the wind power integrated power system is proposed.The power output from thermal units are treated as a series of system states,and the adjustment power output as the system inputs.The state space theory,by which the traditional obj ective function and constraints can be transform ed into matrix forms,is used to describe the relationship between the system states and inputs.Considering the multiple units and multiple periods,this dispatching model will make full use of the latest forecast inform ation.By simplifying the proposed model and deriving incremental matrix,the original model was converted into a standard quadratic programming problem,which can be easily solved by the interior point method.Since the appropriate initial points of the interior point method are difficult to set,the improved particle swarm optimization algorithm proposed in the previous chapters is used to filter the initial points.The numerical simulations show that the proposed model has less total running cost than the traditional single-time optimization,and the rolling schedules are more proactive.(4)A multi-objective optimal dispatching model including wind power,thermal units,hydropower and energy storage is established.In the dispatching model,the cost of power generation,environmental protection and clean energy consumption are considered.The third independent objective is aimed to reduce the abandoned wind and water,which can improve the consumption of the clean energy.Due to the large number of objectives and complicated constraints,the violation of constraints is optimized as an independent objective,which can enhance the search ability at the constrained boundaries.For the choice of non-dominated solutions,a filtering method based on minimum standard deviation of satisfaction is proposed to ensure the fairness between objectives.The simulation examples show that the proposed model realizes the joint optimization of different types of power sources.It fully embodies the regulatory role of hydropower and energy storage,effectively avoiding the waste of hydropower and wind power.Compared with the traditional method of environmental economic dispatch,the proposed model increases the consumption of the clean energy.