中国电力 ›› 2025, Vol. 58 ›› Issue (7): 197-206.DOI: 10.11930/j.issn.1004-9649.202411008

• 技术经济 • 上一篇    下一篇

基于机器学习算法的新型电力系统中电网投资成效评价及投资推演

田鑫1(), 靳晓凌1(), 韩新阳1(), 杨军伟2, 张新圣1   

  1. 1. 国网能源研究院有限公司,北京 102209
    2. 国网滁州供电公司,安徽 滁州 239099
  • 收稿日期:2024-11-01 发布日期:2025-07-30 出版日期:2025-07-28
  • 作者简介:
    田鑫(1985),男,通信作者,硕士,高级工程师,从事电网投资效益评价、电力系统技术经济等研究,E-mail:tianxin@sgeri.sgcc.com.cn
    靳晓凌(1979),女,博士,高级工程师(教授级),从事电网规划、电网技术经济等研究,E-mail:jinxiaoling@sgeri.sgcc.com.cn
    韩新阳(1972),男,硕士,高级工程师(教授级),从事电网规划、城市电网、电网技术经济等研究,E-mail:hanxinyang@sgeri.sgcc.com.cn
  • 基金资助:
    国家电网有限公司科技项目(新型电力系统下源网荷储运行管理优化关键技术与实证应用研究,5700-202257454A-2-0-ZN)。

Evaluation of Grid Investment Effectiveness and Investment Simulation for New-Type Power Systems Based on Machine Learning Algorithm

TIAN Xin1(), JIN Xiaoling1(), HAN Xinyang1(), YANG Junwei2, ZHANG Xinsheng1   

  1. 1. State Grid Energy Research Institute Co., Ltd., Beijing 102209, China
    2. State Grid Chuzhou Electric Power Supply Company, Chuzhou 239099, China
  • Received:2024-11-01 Online:2025-07-30 Published:2025-07-28
  • Supported by:
    This work is supported by Science and Technology Project of SGCC (Key Technologies and Empirical Application Research on Optimization of Source Grid Load Storage Operation Management under New Power System, No.5700-202257454A-2-0-ZN).

摘要:

分布式电源、储能、微电网等新要素的接入给电力系统运行特性带来较大影响,电网作为新型电力系统的重要组成部分,其规划和投资决策需要充分考虑新要素对电网投资成效的影响,确保电网投资规模、结构与新型电力系统构建的目标相一致。当前电网投资成效评价大多关注成本投入和经济效益,新型电力系统构建要求电网投资成效以整体效益为指引,提取关键影响因素,为电网投资推演提供方向性指引。首先,建立了一种基于机器学习算法的新型电力系统电网投资成效评价和投资推演方法,基于最小二乘支持向量机(least square support vector regression,LSSVM)的机器学习算法构建了电网投资成效评价模型,采用粒子群优化算法(particle swarm optimization,PSO)进行参数寻优,并以分布式电源和储能建设场景为例进行算例分析。然后,基于新型电力系统下电网物理指标、电网投资指标与电网投资成效指标之间的量化映射关系,建立电网投资推演方法和模型,采用差异化场景对电网投资推演方法进行案例分析,验证方法的可行性,为新型电力系统构建背景下电网投资决策提供理论和技术支撑。

关键词: 电网投资, 机器学习, 最小二乘支持向量机, 粒子群优化, 典型场景, 投资成效评价, 策略推演

Abstract:

The access of emerging elecments such as distributed power generation, energy storage, and microgrids has a great impact on the operation characteristics of the power system. Power grid, as a critical component of new-type power systems, requires that its planning and investment decisions fully account for the emerging elements' impact on investment effectiveness, so as to ensure that the grid investment scale and allocation align with the construction objectives of new-type power systems. At present, evaluation of the power grid investment effectiveness mostly focuses on cost input and economic benefits. The construction of new-type power systems, however, requires that the power grid investment effectiveness be guided by the overall benefits, and the key influencing factors be extracted to provide directional guidance for the power grid investment simulation. This paper proposes a power grid investment effectiveness evaluation and investment simulation method for new-type power systems based on machine learning algorithm, and constructs a power grid investment effectiveness evaluation model based on the machine learning algorithm of the least squares support vector machine (LSSVM), and uses the particle swarm optimization algorithm (PSO) to optimize the parameters of LSSVM. The distributed power generation and energy storage scenarios are used for case study. Based on the quantitative mapping relationship between the physical indicators of the power grid, the power grid investment indicators and the power grid investment effectiveness indicators under the new-type power systems, the power grid investment simulation method and model are established, and case study is carried out using differentiated scenarios to verify the feasibility of the proposed power grid investment simulation method. This study can provide a theoretical and technical support for the power grid investment decision-making for the new-type power systems.

Key words: grid investment, machine learning, least squares support vector machines, particle swarm optimization, typical scenarios, investment effectiveness evaluation, strategy simulation


AI


AI小编
您好!我是《中国电力》AI小编,有什么可以帮您的吗?