A method of scaling distortion assessment based on dynamic time warping. (November 2022)

• Record Type: Journal Article
• Authors: Liu, Zihan; Bai, Tianze; Peng, Changhong; Bao, Hui; Lu, Jianchao
• Is Part Of: Annals of nuclear energy. Volume 177(2022)
• Publication Date: 2022-11
• Subjects: Dynamical system scaling -- Scaling distortion -- Dynamic time warping -- Accumulated distortion; Nuclear energy -- Periodicals; Nuclear engineering -- Periodicals; 621.4805
• DOI: 10.1016/j.anucene.2022.109304
• Languages: English
• ISSNs: 0306-4549

Abstract

Highlights: Dynamic time warping algorithms are first introduced to assess the scaling distortion. The process curve mapping results of four dynamic time warping algorithms are compared. The total distortion factor and accumulated distortion factor of a parameter are obtained. The method to assess scaling distortion of a system is proposed. Abstract: The scaling distortion assessment needs to be improved to meet the expectations of international nuclear community. In other research fields, dynamic time warping (DTW) has been widely used in similarity measurement of time series. In this paper, DTW is introduced into the scaling distortion assessment. By comparing the mapping effect of process curves, the t ∗, β ∗, ω ∗ scheme in the dynamical system scaling method and the corresponding time-value-derivative DTW can get a good match. The total distortion factor and accumulated distortion factor of a single parameter are defined based on the DTW distance. In addition, a system distortion assessment method based on system parameter weights and the DTW distance is proposed. It is applied to the reduced enthalpy difference models and can provide the system distortion and the contribution of each parameter distortion to the system distortion. This method is of great significance to identify the distortion source term and optimize the model design.