Abstract: In the cases that the phasor measurement units (PMUs) cannot be deployed with all buses in the power system， the traditional nonlinear state estimation needs to be carried out by using the hybrid measurement of PMU and supervisory control and data acquisition (SCADA). SCADA data are low accurate and have bad data， and the iterative solution process of mixed data will lead to low computational efficiency and truncation error. In order to solve this problem， a high fault-tolerant fast state-estimation method for power systems on the basis of stack denoising auto-encoder (SDAE) and extreme learning machine (ELM) pseudo-measurement modeling is proposed in this paper. The SCADA measurement data with bad measurement are used as the input of the SDAE-ELM pseudo-measurement model， and the real and imaginary parts of the bus voltage phasors are used as outputs， the pseudo-measurement value and the pseudo-measurement error model are obtained by training the historical data. The pseudo measurement with high precision is obtained after the training and then the pseudo-measurement and PMU measurement are used for fast linear state estimation. The simulation results show that the proposed method improves the computational efficiency and verify the effectiveness of the proposed method on the basis of ensuring the estimation accuracy.

Key words: high fault tolerance, fast, state estimation, stacked denoise auto encoder, extreme learning machine