A three-phase state estimation in active distribution networks

ANN-based correlation of measurements in micro-grid state estimation

“…For FF ANN‐based model, the internal active power represents the output layer from FF ANN, while the internal reactive power is determined by MGSE itself (depending on the applied DG connection to the MG (synchronous or induction generator, electronically‐coupled DG units, etc.) ). The output three‐phase apparent power in FF ANN‐based model was calculated as described in Section 3.3.…”

“…Using the theory for calculation of PV- [18,19,[26][27][28][29][30] and WT-based [4,18,19,24,28,31,32] internal (from primary energy resource) and output active powers, we identified the following weather parameters in PV-and WT-based input matrix (1), respectively: ; ;…”

“…Typically, applied approach uses different mathematical models that determine the relation between weather input data and internal active power of the DG unit [4,[28][29][30][31][32].…”

“…For FF ANNs, the initial IG's slips (s) are added to the input layer (see architecture of ANNs in Figure 2). The initial internal active powers (used at the output layer of FF ANNs in the Training phase, shown in Figure 2a) are calculated based on mathematical models of different DG units and loads [4,18,19,24,[27][28][29][30][31][32].…”

“…x state vector composed with components of the complex voltage, as well as additional state variables, depending on DG unit type (see [32] for details) z measurement vector (real-time and pseudo measurements) h(x)measurements-state variables model R covariance matrix (defined for real-time measurements in Section 3.4 (Step 2b) or for pseudo measurements in Section 3.4 (Step 2a) and Appendix A).…”