Distribution Network State Monitoring using Compressive Sampling

The application of renewable energy sources complicates the distribution network structure, that raises requirements on a large number of sensors for state monitoring of the network, which causes big challenge on data transmission. Thus a compressive sampling method is proposed in this paper for decreasing the sampling and transmission data of current of the power line. In this method, the discrete cosine transform was first used as orthogonal basis for signal decomposition, then the random Gaussian matrix was applied as the measurement matrix for observation. Finally the signal was reconstructed based on the convex optimization method with L1 parametrization. Simulation results show that the number of sampling points of current at a single node using proposed compressed sampling method could achieved 91.9472% less than the number using Nyquist sampling method. Furthermore, The compressed signal can be reconstructed at the distribution network sub-station, and the RMSE is only 0.5185, which greatly reduces the data required to be transmitted for grid line monitoring and reduces the communication network load to a certain extent.