• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

Electric Power Construction ›› 2019, Vol. 40 ›› Issue (11): 1-7.doi: 10.3969/j.issn.1000-7229.2019.11.001

Previous Articles     Next Articles

Communication Bandwidth Prediction Based on Hybrid Queue Model for Condition Monitoring Service in Distribution Network

WANG Zhiwei1, LI Jianqi2, HUANG Biyao2   

  1. 1.School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China; 2.Global Energy Interconnection Research Institute Co., Ltd., Beijing 102209, China
  • Online:2019-11-01
  • Supported by:
    This work is supported by State Grid Corporation of China Research Program( No. SGGR0000XTJS1800957).

Abstract: The construction of condition monitoring system in distribution network is an important part of constructing ubiquitous power internet of things. Accurate estimation of communication bandwidth is of great significance to identity business requirements and to optimize resource allocation. In view of current communication bandwidth estimation method based on queuing theory results in the limitation of bandwidth computation complexity when hybrid services coexist, a method based on G/M/1/N queue model is proposed. On the basis of the architecture of condition monitoring system and the characteristics of data service, this paper analyzes the statistical description and basic transmission rate of the inter-arrival time of mixed data service, and gives the optimal solution method of bandwidth to meet the requirements of QoS. Taking the typical condition monitoring application as an example, the quantitative evaluation process among service bandwidth, QoS and bandwidth utilization is discussed. The “small data” characteristics of the information flow in the condition monitoring network of distribution network are verified, which provides a theoretical support for the application of low-power wide-area IoT communication technology in condition monitoring.

Key words: status monitoring in distribution network, hybrid queue, G/M/1/N model, bandwidth prediction

CLC Number: