2019 Vol.2 Dec. No.6

• Title: Prediction algorithm for distributed photovoltaic mass access to power grid based on machine learning
• Name: Hagiwara Michiko
• Company: Miyazaki Sangyo-keiei University
• Abstract:

  The short-term power network output prediction period of traditional algorithms is generally one to several days. If the prediction result is too high, the system operation efficiency will be too low. As a result, resources are wasted. To this end, a large number of distributed photovoltaics based on machine learning are connected to the power grid output prediction algorithm. First, a power output prediction model is established to limit the system line loss and transformer loss. Secondly, based on the distributed photovoltaic grid output prediction model, the vector moment method and the information method are used to reduce the search space. The voltage output prediction calculation formula of the distribution network node with distributed photovoltaic is deduced, and the power output prediction algorithm is realized. Finally, experiments have proved that a large number of distributed photovoltaic power output prediction algorithms can effectively improve the system's operating efficiency.

• Keyword: Operating Efficiency; Photovoltaic Capacity; Radial Structure; Power System;
• DOI: 10.12250/jpciams2019060648
• Citation form: Hagiwara Michiko.Prediction algorithm for distributed photovoltaic mass access to power grid based on machine learning[J]. Computer Informatization and Mechanical System, 2019, vol. 2, pp. 91-97.

[1] Hao Zhen. Research on short-term photovoltaic power prediction based on cloud computing and machine learning [D]. North China Electric Power University, 2017. 12(5): 1118-1122.

[2] Cheng Mengzeng, Dai Xiaoyu, Wang Shanshan, etc. Study on Voltage Characteristics of Distributed Photovoltaic Access to Active Distribution Network [J]. Northeast Power Technology, 2017, 38 (5): 2118-2221.

[3] Wang Hongkun, Ge Leijiao, Li Hongwei, et al. Summary of characteristics analysis and prediction methods of distributed photovoltaic power generation [J]. Electric power construction, 2017, 38 (7): 2321-2329.

[4] Fu Yang, Tian Can, Li Zhenkun, et al. Distributed photovoltaic spatial-temporal distribution prediction considering space capacity saturation [J]. Power grid technology, 2017 .23(12): 3917-3925.

[5] Li Lei, Wang Junxi, He Yi, et al. Multi-point PV-DG day-ahead allocation plan based on K-means clustering particle swarm optimization [J]. High Voltage Technology, 2017.44 (4): 1263-1270.

[6] Huang Wei, Gao Yifang, Zhang Yongjun, et al. Limit Capacity Assessment of Distributed Photovoltaic Access Distribution Network Considering Uncertainty [J]. Power System Protection and Control, 2018.11 (14).1453-1654.

[7] Jin Zhuoshen, Xiang Tieyuan, Chen Hongkun, et al. Distributed photovoltaic access planning method considering power quality problems [J]. Power system protection and control, 2017, 45 (9): 1231-1238.

[8] Cui Yang, Pan Yu, Wang Zeyang, et al. Output optimization and capacity allocation method of distributed photovoltaic-energy storage system considering power loss [J]. Renewable Energy, 2017, 35 (2): 245-251.

[9] Zhang Wenqi, Wang Qi, Xu Qichen, etc. Study on location and capacity of distributed photovoltaic access distribution network considering time series characteristics [J]. Journal of Nanjing Normal University (Engineering Edition), 2017, 17 (3): 1122-1128.

[10] Xiong Youming. Research on the impact of distributed photovoltaic access in typical functional areas on the construction and transformation of distribution network [J]. Communication World, 2017 .12(10): 190-191.

[11] Li Sheng, Jiang Chengcheng, Zhao Zhiyu, et al. Transient Voltage Stability of Distributed Photovoltaic Power Station Connected to Low Voltage Distribution Network [J]. Power System Protection and Control, 2017, 45 (8):167-172.

[12] Cheng Mengzeng, Zhang Mingli, Liang Yi, et al. Characteristics of distributed photovoltaic access to low-voltage distribution network under voltage constraints [J]. Electrical applications, 2018 .22(1): 128-135.