基于机会约束规划的含多风电场动态经济调度

发布时间：2018-04-27 22:47

本文选题：风力发电 + 动态经济调度　；参考：《浙江大学》2017年硕士论文

【摘要】：随着能源需求增长与化石燃料的日趋枯竭,风力发电作为可再生能源受到各国的关注。但风力发电本质上具有波动性和随机性,大规模风电并网给电力系统的动态经济调度和安全运行带来新的挑战和要求。传统调度策略不再适用于包含大规模风电并网的系统,需要寻求新的调度决策方案。本文建立了基于机会约束规划的含多风电场动态经济调度模型,计入负荷和风电出力的不确定性,考虑机组爬坡率、线路安全、旋转备用等约束,以机会约束的形式保证正、负旋转备用满足负荷和风电实际出力的波动,优化常规机组、风电计划出力及预留的备用容量,确保系统失负荷和弃风的风险低于预定槛值。相比于现有的含风电场动态经济调度,能更好地处理多风电场接入的情形。针对机会约束的求解,采用两种方法,将其转化为确定性模型和利用基于随机模拟的粒子群算法求解。求解机会约束规划的传统方法是将机会约束其转化为确定性约束,重点和难点在于联合变量的累积分布函数及其反函数的快速求解。本文提出采用FFT快速计算卷积获得联合变量的概率分布,以将机会约束转化为确定性约束,使得模型转化为确定性模型,避免了复杂的卷积运算,用CPLEX求解得到的确定性模型,该方法大大减少了算法的运行时间。以修改的IEEE39节点系统为算例验证了所提调度模型的正确性及化简方法的有效性。但当机会约束的各变量之间不相互独立时,很难将机会约束转化为确定性约束,上述方法将不再可行;而基于随机模拟的智能优化算法不需要将机会约束转化为确定性约束,通过生成大量试验来模拟机会约束成立的概率。本文提出基于随机模拟的改进粒子群算法求解包含机会约束的含多风电场电力系统动态经济调度问题,适用范围广,能应用于求解非凸的调度优化问题。在优化过程中粒子群算法的学习因子设定为自适应变化,惯性因子非线性变化以平衡全局优化与局部优化,避免陷入局部最优;同时,加入可行化调整策略及变异调整策略,以增强粒子群算法的优化能力。本文采用机会约束规划处理风力发电带来的不确定性,较为详细地提出了模型的化简、求解方法,对含大规模风电场并网的电力系统优化问题具有一定的借鉴意义。
[Abstract]:With the increase of energy demand and the depletion of fossil fuels, wind power as a renewable energy has attracted much attention. However, wind power generation is inherently volatile and stochastic, and large-scale wind power grid connection brings new challenges and requirements to dynamic economic dispatch and safe operation of power system. The traditional scheduling strategy is no longer suitable for large-scale wind power grid connected systems, so it is necessary to seek a new scheduling decision scheme. In this paper, a dynamic economic scheduling model with multiple wind farms based on chance constrained programming is established. The uncertainty of load and wind power output is taken into account, and the constraints such as slope climbing rate, line safety and rotation reserve are considered, and the positive is guaranteed in the form of chance constraints. The negative rotation reserve satisfies the fluctuation of load and actual output of wind power, optimizes the conventional unit, the planned output capacity of wind power and the reserved reserve capacity, so as to ensure that the risk of system losing load and abandoning wind is lower than the predetermined threshold value. Compared with the existing dynamic economic dispatching of wind farm, it can better deal with the situation of multiple wind farm access. For the solution of chance constraint, two methods are used to transform it into deterministic model and particle swarm optimization algorithm based on stochastic simulation. The traditional method to solve the chance-constrained programming is to transform the opportunistic constraints into deterministic constraints. The emphasis and difficulty lies in the quick solution of the cumulative distribution function and its inverse function of the joint variables. In this paper, the probability distribution of joint variables is obtained by using FFT to calculate convolution quickly, so that the chance constraints can be transformed into deterministic constraints, so that the model can be transformed into deterministic models, thus avoiding the complicated convolution operation and solving the deterministic model by CPLEX. This method greatly reduces the running time of the algorithm. The correctness of the proposed scheduling model and the effectiveness of the simplified method are verified by an example of the modified IEEE39 node system. However, when the variables of opportunity constraints are not independent from each other, it is difficult to transform the opportunity constraints into deterministic constraints, so the above methods will not be feasible, and the intelligent optimization algorithm based on stochastic simulation does not need to transform the opportunity constraints into deterministic constraints. The probability of chance constraint is simulated by generating a large number of experiments. In this paper, an improved particle swarm optimization algorithm based on stochastic simulation is proposed to solve the dynamic economic scheduling problem of multi-wind farm power systems with chance constraints. It has a wide range of applications and can be used to solve non-convex scheduling optimization problems. In the process of optimization, the learning factor of PSO is set as adaptive change, the nonlinear variation of inertial factor is used to balance global optimization and local optimization to avoid falling into local optimum, at the same time, feasible adjustment strategy and mutation adjustment strategy are added. In order to enhance the optimization ability of particle swarm optimization algorithm. In this paper, the opportunity-constrained programming is used to deal with the uncertainty caused by wind power generation, and the simplification and solution of the model are presented in detail, which can be used for reference in the optimization of power system with large-scale wind farms connected to the grid.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【相似文献】