Reactive Power Flow Convergence Adjustment Based on Deep Reinforcement Learning

Power flow calculation is the basis of power grid planning and many system analysis tasks require convergent power flow conditions.To address the unsolvable power flow problem caused by the reactive power imbalance,a method for adjusting reactive power flow convergence based on deep reinforcement learning is proposed.The deep reinforcement learning method takes switching parallel reactive compensation as the action space and sets the reward value based on the power flow convergence and reactive power adjustment.For the non-convergence power flow,the 500 kV nodes with reactive power compensation devices on the low-voltage side are converted into PV nodes by node type switching.And the quantified reactive power non-convergence index is acquired.Then,the action space and reward value of deep reinforcement learning are reasonably designed and the adjustment strategy is obtained by taking the reactive power non-convergence index as the algorithm state space.Finally,the effectiveness of the power flow convergence adjustment algorithm is verified by an actual power grid system in a province.

Effect of Left Ventricular Outflow on Flow ConvergenceRegion on the Left Septal Surface in Ventricular SeptalDefect

The corrected shunt flow rate (Fc ) and corrected defect orifice area (Ac) were calculated by modified equation F= 2πR2 ×(NL-VL.voT× Sinθ) in 23 patients with single membranous ventricular septal defect, in order to correct the ef fect of left ventricular outflow on flow convergence region on the left septa1 sur-face. The results indicated that F. was closely correlated with Qp -Q5. and Qp/Q5measured by pulsed wave Doppler (r = 0. 95 and r = 0. 81 respectively, P < 0.001 ). And the correlation between A. and the diameter of defect (Dd) rneasureddirectly in two-dimensional views was better than that between uncorrected defectorifice area (A ) and the Dd (r = O- 98 and O- 69, respectively, P< O. Ool ). Theshunt flow rate calculated by ideal equation F= 2ffR2 X NL overestimated the actu-al shunt flow rate in ventricular septal defect, especialIy in mernbranous type.Our study concluded that F. can be used for a more accurate eva1uation of theshunt severity of ventricular septal defect.

Automated Two-stage Conversion of Hourly Optimal DC Flow Solution to AC Power Flow

Conversion of hourly dispatch cases derived using DC optimal power flow(DCOPF)to AC power flow(ACPF)case is often challenging and requires arduous human analysis and intervention.This paper proposes an automated two-stage approach to solve ACPF formulated from DCOPF dispatch cases.The first stage involved the use of the conventional Newton Raphson method to solve the ACPF from flat start,then ACPF cases that are unsolvable in the first stage are subjected to a hotstarting incremental method,based on homotopy continuation,in the second stage.Critical tasks such as the addition of reactive power compensation and tuning of voltage setpoints that typically require human intervention were automated using a criteriabased selection method and optimal power flow respectively.Two datasets with hourly dispatches for the 243-bus reduced WECC system were used to test the proposed method.The algorithm was able to convert 100%of the first set of dispatch cases to solved ACPF cases.In the second dataset with suspect dispatch cases to represent an extreme conversion scenario,the algorithm created solved ACPF cases that satisfied a defined success criterion for 77.8%of the dispatch cases.The average run time for the hotstarting algorithm to create a solved ACPF case for a dispatch was less than 1 minute for the reduced WECC system.