综合能源一体化采集系统的多任务自适应实时调度方法

Multitask and adaptive real-time scheduling method of energy integration collection system

随着能源互联网和综合能源服务的建设推进,促进水、气、热、电的远程自动集采集抄建设已尤为关键。文中以用电信息采集系统为基础,设计了电、水、气、热能源计量一体化的采集技术架构体系,明确能源计量系统的技术路线。通过对系统采集任务类别解析,提出基于Spark集群的多任务自适应的实时调度方法,解决了采集系统基于Spark集群现有调度机制存在任务分配不均、总体任务完成时间较长和优先级高的任务超时响应造成用户体验差等问题。实验结果表明,把总数为200万采集任务划分六类,采用文中和Spark默认调度方法进行对比,该方法针对时效性和优先级要求高的任务完成时间均早于默认方法,获得较好的响应速度,极大改善系统用户体验。

With the development of energy Internet and integrated energy services,it is very important to promote the construction of remote automatic collection and transcription of water,gas,heat and electricity.Based on the electrical information collection system,this paper designs a collection technology framework for the integration of electric,water,gas and thermal energy measurement,and defines the technical route of the energy measurement system.Then,by analyzing the collection task category of the system,a multitask and adaptive real-time scheduling method based on Spark cluster is proposed in this paper,which solves the problem that the collection system is based on the unequal assignment of task allocation in the existing scheduling mechanism of the Spark cluster,the longer task completion time and the higher priority of the task,resulting in the poor user experience.Finally,the experimental results show that the total number of 2 million collection tasks is divided into six classes,and the comparison between this method and the Spark default scheduling method is carried out.This method has shorter executing time and better response speed,which can greatly improve the user experience of the system.