Digital twin early warning method study for overload risk of distribution network with a high proportion of photovoltaic access
DOI：DOI: 10.19783/j.cnki.pspc.211422
Key Words:distribution network  distributed photovoltaic  load safety  digital twins  Markov chain  Gibbs sampling
Author NameAffiliation
DU Xiaodong 1. State Grid Hebei Electric Power Co., Ltd. Research Institute, Shijiazhuang 050021, China
2. Power Distribution Technology Center, China Electric Power Research Institute Co., Ltd., Beijing 100192, China 
ZHAO Jianli 1. State Grid Hebei Electric Power Co., Ltd. Research Institute, Shijiazhuang 050021, China
2. Power Distribution Technology Center, China Electric Power Research Institute Co., Ltd., Beijing 100192, China 
LIU Keyan 1. State Grid Hebei Electric Power Co., Ltd. Research Institute, Shijiazhuang 050021, China
2. Power Distribution Technology Center, China Electric Power Research Institute Co., Ltd., Beijing 100192, China 
ZHAN Huiyu 1. State Grid Hebei Electric Power Co., Ltd. Research Institute, Shijiazhuang 050021, China
2. Power Distribution Technology Center, China Electric Power Research Institute Co., Ltd., Beijing 100192, China 
Hits: 3784
Download times: 814
Abstract:With reference to building a new power system with new energy as the main body, distributed photovoltaic (PV) will show a trend of extremely high levels of access. The randomness of PV output given its high proportion will aggravate the uncertain change of power flow in the distribution network. It brings serious overload risks to the distribution equipment, and the problem of load safety is prominent. Based on this, a digital twin based load security analysis and warning method of the distribution network with a high proportion of PV is proposed. The realization mechanism and function scheme of overload risk warning in twin distribution network are presented. The overload tolerance of equipment is analyzed according to relevant standards, and the load safety early warning levels in uncertain scenarios are established. Based on the historical operational data, Markov models of the random behavior of sources and loads in the distribution network are established. The Gibbs sampling algorithm is employed for random Monte Carlo sampling to the Markov models. Based on the current states, by means of the ultra-real-time computing capacity of the digital twins to quickly simulate the afterward random scenes of the distribution network, an overload risk index is calculated and evaluated. Numerical cases verify the effectiveness and rationality of the method. The case study demonstrates that the proposed method can analyze the load security situation of a distribution network with a high proportion of photovoltaic access in the analysis period, providing a certain reference for power system dispatchers. This work is supported by the Key Research and Development Program of Hebei Province (No. 21312102D).
View Full Text  View/Add Comment  Download reader