| 引用本文: | 张 野,李凤婷,张高航,等.考虑风电爬坡备用需求的风电高渗透电力系统优化调度方法[J].电力系统保护与控制,2024,52(23):95-106.[点击复制] |
| ZHANG Ye,LI Fengting,ZHANG Gaohang,et al.Optimization and scheduling methods for wind power high-penetration power systemsconsidering wind power ramping reserve requirements[J].Power System Protection and Control,2024,52(23):95-106[点击复制] |
|
| 摘要: |
| 准确描述风电出力预测误差特性有助于合理配置系统备用容量，优化日前调度计划，提出了一种考虑风电爬坡备用需求的电力系统日前优化调度方法。首先，基于风电爬坡段提取爬坡特征，建立爬坡幅值-预测功率二维区间，采用自适应核密度估计方法拟合风功率预测误差概率分布。其次，基于风电预测误差分布确定系统备用需求，以含备用成本和风险成本的系统综合运行成本最小为目标，建立连续时间日前优化调度模型。然后，采用Bernstein多项式插值解空间变换进行模型转换求解，优化备用容量、机组组合及出力计划。最后，通过算例验证了建立的风电预测误差分布模型能更准确地描述风电随机波动特性，提出的日前调度方法可以合理配置系统备用容量，保证运行安全性和经济性。 |
| 关键词: 预测误差分布 爬坡特征 自适应核密度估计 连续时间模型 风险成本 |
| DOI：10.19783/j.cnki.pspc.240373 |
| 投稿时间：2024-03-31修订日期：2024-08-25 |
| 基金项目:新疆维吾尔自治区重点研发项目资助(2022B01016) |
|
| Optimization and scheduling methods for wind power high-penetration power systemsconsidering wind power ramping reserve requirements |
| ZHANG Ye1,LI Fengting1,ZHANG Gaohang1,XIAO Zhongjie2 |
| (1. College of Electrical Engineering, Xinjiang University, Urumqi 830017, China;
2. State Grid Xinjiang Electric Power Co., Ltd., Urumqi 830036, China) |
| Abstract: |
| Accurately describing the error characteristics of wind power output prediction is helpful for the rational allocation of system reserve capacity and the optimization of day-ahead scheduling plans. This paper proposes a day-ahead optimization scheduling method for power systems considering wind power ramping reserve requirements. First, based on the wind power ramping segment, the ramping characteristics are extracted, and a two-dimensional interval of ramping amplitude-predicted power is established. The adaptive kernel density estimation method is used to fit the probability distribution of wind power prediction errors. Then, based on the distribution of wind power prediction errors, the system reserve requirements are determined. Looking for minimal comprehensive operating costs of reserve costs and risk costs, a continuous-time day-ahead optimization scheduling model is established. Next, the Bernstein polynomial interpolation solution space transform is adopted to complete model conversion, thereby optimizing reserve capacity, unit combination, and output plans. Finally, a case study verifies that the established wind power prediction error distribution model can accurately describe the stochastic characteristics of wind power. The proposed day-ahead scheduling method can effectively allocate system reserve capacity, ensuring operational safety and economic efficiency. |
| Key words: forecast error distribution ramping feature adaptive kernel density estimation continuous-time model risk cost |
