The Study Of An Adaptive Backoff Mechanism In Smart Grid Online Monitoring

With the rapid development of power industry in China,more and more objects are need to be monitored in smart grid,traditional supervisory control and data acquisition(SCADA)system can not meet the information requirement of smart grid.The research on wireless sensor network becomes a hot topic of wireless network filed,because it is the key technology in sensor layer and the main means of collecting information.Wireless sensor network has a good research and application prospect being used for electricity information collection system of smart grid.It effectively guarantees smart grid's construction specification development.The increasing of information and the presence of unexpected events,it will inevitably cause network congestion.Congestion brings about network latency,packet loss,waste of bandwidth.So it is necessary to research on congestion control algorithms of wireless sensor network.This paper Wireless sensor network is used in the SCADA system and proposes a new backoff strategy using predicting the network load and least squares linear fit(PNL_LS)method to cope with the impact on network performance of the constantly changing data transmission.This paper firstly discuses the requirement of the electricity information collection system,then the paper designs a structure for the electricity information collection system.Secondly,we introduce the backoff strategy of PNL_LS.When carrier sense multiple access with collision avoidance(CSMA/CA)begins,the next time network load is predicted based on the past network load and data packet sending status,and the value of backoff exponent is dynamically adjusted by LS.Thirdly,compared with the IEEE802.15.4 standard protocol,the simulation results of NS2 show that the PNL_LS algorithm not only improves the throughput of the network,but also reduces the packet loss rate and energy consumption,significantly improves the network performance.Finally,we carry out the system using ARM9,GPRS and ZigBee and port the PNL_LS to Z-stack by Contiki.Hardware experiments also show that the algorithm can effectively reduce the packet loss rate.