Fault Location in Active Distribution Network Under Hybrid Classical Quantum Computing Architecture

Abstract

At present, research on the combination of quantum computing and artificial intelligence algorithms applied to the fault location of distribution networks is very sufficient. Because this type of algorithm usually uses the switching function in logical form, it leads to numerical instability and is difficult to expand to the problem of a larger scale distribution network. Therefore, in this study, we propose a hybrid classical quantum computing architecture model for the fusion of quantum annealing and fault location in distribution networks. Firstly, an algebraic objective function suitable for D-wave quantum computer was constructed according to the distributed idea, and the feasibility of the model was verified by using quantum platform. Then, a decoupling optimization model was proposed to improve the poor accuracy of the model in the case of multiple faults, and the decoupling optimization operation was completed by classical calculation and verified experimentally on the quantum platform. Then the original model and decoupled optimization model were applied to fault scenarios of 20-node single power distribution network, 33-node single power distribution network, 9-node distribution network with distributed generation source(DG), 16-node distribution network with DGs, and 33-node distribution network with DGs. The performance of the two models was compared in five aspects: the accuracy of fault location, quantum bit resource consumption, model parameter range, running time and fault tolerance performance. Finally, the quantum classical hybrid decoupling optimization model was compared with the simulated annealing algorithm, the quantum annealing algorithm and the improved quantum annealing algorithm in three aspects of accuracy, running time and quantum bit consumption resources. Both the comparison between the two models constructed in this paper and the comparison between the hybrid model and the classical model can reflect that the decoupled optimization model is committed to realizing the fault location of a larger scale distribution network with fewer quantum bits, and can play the advantages of the classical quantum hybrid architecture model.

Keywords

• Quadratic unconstrained binary optimization problem
• hybrid classical quantum
• distributed idea
• decoupling optimization
• distributed generation source