Abstract
Power loss is a bottleneck in every power system and it has been in focus of majority of the researchers and industry. This paper proposes a new method for determining the power loss in wind-solar power system based on deep learning. The main idea of the proposed scheme is to freeze the feature extraction layer of the deep Boltzmann network and deploy deep learning training model as the source model. The sample data with closer distribution with the data under consideration is selected by defining the maximum mean discrepancy contribution coefficient. The power loss calculation model is developed by configuring the deep neural network through the sample data. The deep learning model is deployed to simulate the non-linear mapping relationship between the load data, power supply data, bus voltage data and the grid loss rate during power grid operation. The proposed algorithm is applied to an actual power grid to evaluate its effectiveness. Simulation results show that the proposed algorithm effectively improved the system performance in terms of accuracy, fault tolerance, nonlinear fitting and timeliness as compared with existing schemes.
Highlights
In recent years, the distributed power generation-based PV systems have achieved rapid development
The calculation accuracy of the deep belief network based deep neural network (DBN-deep neural network (DNN)) model is 4.95% higher than that of the back propagation (BP) neural network, which shows that the deep belief network (DBN)-DNN model has more advantages in processing high-dimensional input data of the power system
The calculation effect of T-BP is better than BP neural network, and the calculation accuracy is increased by 5.74%, which shows that in the case of wind power photovoltaic power supply, Migration learning improves the calculation performance of BP neural network and deep learning model by migrating data from the source data that is closer to the distribution of the data to be calculated
Summary
The distributed power generation-based PV systems have achieved rapid development. The network loss rate calculation methods mainly include traditional power flow calculation methods, intelligent algorithms, and big data processing technologies based on cloud computing platforms. The authors in [16] proposes a joint network loss rate calculation method for vertical-lateral error matching, which improves the calculation accuracy, but there are insufficient calculation speeds when processing high-dimensional data. The power transmission caused by wind power and photovoltaic grid connection and the impact on the power system flow make the change trend of the grid loss rate and the output of wind power photovoltaic It is closely related, and the data collection system of the new energy power generation system is imperfect, making the data missing situation more serious. The current shallow learning methods have limited processing capabilities for high-dimensional data input features, and cannot well achieve the problem of calculating the network loss rate when the power system is high-dimensional.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
