Time synchronization(TS)is crucial for ensuring the secure and reliable functioning of the distribution power Internet of Things(IoT).Multi-clock source time synchronization(MTS)has significant advantages of high reli...Time synchronization(TS)is crucial for ensuring the secure and reliable functioning of the distribution power Internet of Things(IoT).Multi-clock source time synchronization(MTS)has significant advantages of high reliability and accuracy but still faces challenges such as optimization of the multi-clock source selection and the clock source weight calculation at different timescales,and the coupling of synchronization latency jitter and pulse phase difference.In this paper,the multi-timescale MTS model is conducted,and the reinforcement learning(RL)and analytic hierarchy process(AHP)-based multi-timescale MTS algorithm is designed to improve the weighted summation of synchronization latency jitter standard deviation and average pulse phase difference.Specifically,the multi-clock source selection is optimized based on Softmax in the large timescale,and the clock source weight calculation is optimized based on lower confidence bound-assisted AHP in the small timescale.Simulation shows that the proposed algorithm can effectively reduce time synchronization delay standard deviation and average pulse phase difference.展开更多
The BeiDou-3 navigation satellite system(BDS-3)provides a full-domain high-precision positioning service for the power system to ensure safe and stable operation.However,BDS-3 power system positioning faces certain ch...The BeiDou-3 navigation satellite system(BDS-3)provides a full-domain high-precision positioning service for the power system to ensure safe and stable operation.However,BDS-3 power system positioning faces certain challenges,such as complex electromagnetic interference and incomplete error elimination.Herein,a back propagation neural network-improved least mean square(BP-LMS)adaptive filtering method is proposed for the BDS-3 full-domain and high-precision power system positioning,which utilizes the loss function to update the weight of the BP hidden layer,computes the pseudo compensation range,and eliminates the impact of electromagnetic interference to enhance the accuracy of power system positioning.Simulation results confirm the superior performance of BP-LMS in positioning accuracy and error elimination.Compared with LMS and normalized least mean square(NLMS),the filtering error of the proposed BP-LMS adaptive filtering method is decreased by 57.14%and 51.38%,respectively.展开更多
基金supported by Science and Technology Project of China Southern Power Grid Company Limited under Grant Number 036000KK52200058(GDKJXM20202001).
文摘Time synchronization(TS)is crucial for ensuring the secure and reliable functioning of the distribution power Internet of Things(IoT).Multi-clock source time synchronization(MTS)has significant advantages of high reliability and accuracy but still faces challenges such as optimization of the multi-clock source selection and the clock source weight calculation at different timescales,and the coupling of synchronization latency jitter and pulse phase difference.In this paper,the multi-timescale MTS model is conducted,and the reinforcement learning(RL)and analytic hierarchy process(AHP)-based multi-timescale MTS algorithm is designed to improve the weighted summation of synchronization latency jitter standard deviation and average pulse phase difference.Specifically,the multi-clock source selection is optimized based on Softmax in the large timescale,and the clock source weight calculation is optimized based on lower confidence bound-assisted AHP in the small timescale.Simulation shows that the proposed algorithm can effectively reduce time synchronization delay standard deviation and average pulse phase difference.
基金supported by the Science and Technology Project of the China Southern Power Grid Company Limited under Grant Number(GDKJXM20202001).
文摘The BeiDou-3 navigation satellite system(BDS-3)provides a full-domain high-precision positioning service for the power system to ensure safe and stable operation.However,BDS-3 power system positioning faces certain challenges,such as complex electromagnetic interference and incomplete error elimination.Herein,a back propagation neural network-improved least mean square(BP-LMS)adaptive filtering method is proposed for the BDS-3 full-domain and high-precision power system positioning,which utilizes the loss function to update the weight of the BP hidden layer,computes the pseudo compensation range,and eliminates the impact of electromagnetic interference to enhance the accuracy of power system positioning.Simulation results confirm the superior performance of BP-LMS in positioning accuracy and error elimination.Compared with LMS and normalized least mean square(NLMS),the filtering error of the proposed BP-LMS adaptive filtering method is decreased by 57.14%and 51.38%,respectively.
