基于计算负荷预测的配用电终端算力供需动态均衡方法

Dynamic balancing method for supply and demand of power distribution and consumption terminal arithmetic based on computational load forecasting

配电网侧海量分布式对象的持续涌入令配用电终端基于确定边界配置转变为柔性资源配置,该转变是在终端有限资源的基础上适应更广泛对象接入的可行保障。为此,提出一种基于计算负荷预测的配用电终端算力供需动态均衡方法,在配用电终端柔性配置转变下最大限度地降低配用电业务与终端供需配合时的计算负担影响,提高终端针对计算负担的平衡能力与稳定能力,实现资源的充分挖掘利用,以适应未来更广泛的对象接入。根据配用电业务的时移特性与调整策略,建立计算负荷模型;基于配用电业务计算负荷曲线的“季节性”特征,构建结合三次指数平滑与迭代马尔可夫链修正的区段预测模型;根据算力资源的交互关系,并结合业务调整策略的划分特点与计算负荷曲线特点,建立配用电业务算力资源均衡模型。算例结果表明:所建预测模型所获区段标准差值满足预测精度要求,所提均衡方法有效降低了算力资源的不平衡度,表现出良好的计算负担削减能力。

The continuous influx of massive distributed objects in the distribution network has transformed the definite boundary-based configuration of power distribution and consumption terminals into a flexible resource configuration,which is a feasible guarantee to adapt to a wider range of object access based on the limited resources of the terminal.Therefore,a dynamic balancing method for supply and demand of power distribution and consumption terminal arithmetic based on computational load forecasting is proposed to minimize the impact of the calculation burden during the coordination of power distribution and consumption tasks and terminal demand and supply under the change of flexible configuration of power distribution terminals,improve the terminal ability to balance and stabilize the calculation burden,and realize the full exploitation and utilization of resources to adapt to a wider range of object access in the future.According to the time-shift characteristic and adjustment strategy of power distribution and consumption tasks,the model of computational load is established.Based on the seasonal characteristics of the computational load curves for power distribution and consumption tasks,a segment prediction model combining triple exponential smoothing and iterative Markov chain modification is constructed.According to the interactive relationship of arithmetic resources,combined with the division characteristics of task adjustment strategy and the characteristics of computational load curve,a balancing model of the power distribution and consumption task arithmetic resources is established.The results of the example show that,the standard deviation values of the segments obtained by the proposed prediction model meet the requirement of prediction accuracy,and the proposed balancing method can effectively reduce the imbalance of arithmetic resources,showing a good ability to reduce calculation burden.