Power flow calculation method based on power flow embedding technology

Abstract

The invention discloses a power flow calculation method based on the power flow embedding technology. The power flow calculation method comprises the following steps: determining a maximum node numberN, and constructing a training set K, a verification set V and a test set T; for the training data in the step 1), constructing a corresponding positive sample K+ and a corresponding negative sampleK; based on the training sample K, the positive sample K+ and the negative sample K in the step 2), using a triple-based twin neural network, and performing sufficient training to obtain a power flowembedded layer P. The calculation method disclosed by the invention is a direct method, and during final use, a final power flow value can be obtained only by arranging known parameters according to rules and taking the parameters as input of the deep neural network through multiplication of a plurality of matrixes and nonlinear operation of neurons, so the method is relatively simple, relativelyhigh in calculation speed and relatively low in calculation cost; the method can be used for online load flow calculation, and has no convergence problem.

Description

technical field [0001] The present invention relates to the technical field of power systems, in particular to a power flow calculation method based on power flow embedding technology. Background technique [0002] With the continuous development of new energy technologies, the Energy Internet, a new energy utilization system, has emerged. The role of the Energy Internet is to integrate a series of power grid operation data with the support of artificial intelligence technologies such as big data and machine learning to predict various situations, and ultimately enable all machines, equipment, and systems to be dynamically adjusted in real time to improve power grid performance. overall operating efficiency. [0003] With the advent of the era of big data and the improvement of computer technology, the performance of neural networks, especially deep learning, in the field of artificial intelligence has greatly surpassed that of other machine learning models, and has been ob...

AI Technical Summary

Problems solved by technology

[0005] Applying deep learning technology to power flow calculation is a supplement to the existing traditional power flow calculation method. Under the new situation of energy Internet, the power grid structure involved in power flow calculation is becoming more and more complex, which affects the speed and convergence of the algorithm. The Gauss-Seidel iterative method has a simple principle and takes up less computer memory, but when it is applied to a large-scale power system, the number of iterations will increase and it will not converge; the Newton-Raphson method has fast convergence and high accuracy. High, but the Jacobian matrix needs to be recalculated in each iteration process, which causes problems such as taking up too much computer memory and slow calculation speed; the fast decoupling method improves the calculation speed and memory usage, but when Some ill-conditioned conditions may lead to non-convergence

Images