A Power Load Clustering Method Based on Limited DTW Algorithm

Gao, Mingyuan; Gong, Taorong; Lin, Rongheng; Zou, Hua

doi:10.1109/itnec.2019.8729249

Cited by 5 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the most common one is setting a window for warping, restricting the path to the diagonal region. Different structures of warping windows are proposed in [17] and dynamic time window strategy is presented in [18].…”

Section: Missing Value Treatmentmentioning

confidence: 99%

A density-based matrix transformation clustering method for electrical load

Xian

Dong

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

Feature extraction of electrical load plays a vital role in providing a reliable basis and guidance for power companies. In this paper, we propose a novel clustering algorithm named the Density-based Matrix Transformation (DBMT) Clustering method to extract features (peaks, valleys and trends) of electrical load curves. The main objective of the algorithm is to reorder the data items until the data items belonging to the same cluster are organized together; that is, the adjacent matrix is rearranged to the type of block diagonal. This method adaptively determines the number of clusters and filters out noise without input global parameters. Moreover, for the specific characteristics of raw electrical load data, we propose a variant of Dynamic Time Warp (DTW) distance, dsDTW, which aligns the peaks, valleys and trends of load curves meanwhile dealing with missing values in different situations. After feeding the dsDTW adjacent matrix to DBMT, the results indicate that our proposal can accurately extract the feature of the load curves compared to different clustering methods.

show abstract

Section: Missing Value Treatmentmentioning

confidence: 99%

A density-based matrix transformation clustering method for electrical load

Xian

Dong

et al. 2022

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Existing studies have effectively improved DTW and suppressed pathological alignments. The LimitDTW algorithm was proposed in [17] to restrict the search area to the regions close to the diagonal area of the distance matrix. However, limiting the number of connection points may cause incorrect alignment owing to its rigidity.…”

Section: Introductionmentioning

confidence: 99%

High-resolution Load Profile Clustering Approach Based on Dynamic Largest Triangle Three Buckets and Multiscale Dynamic Warping Path Under Limited Warping Path Length

Wen,

Ma,

Zhang

et al. 2023

Journal of Modern Power Systems and Clean Energy

View full text Add to dashboard Cite

With the popularity of smart meters and the growing availability of high-resolution load data, the research on the dynamics of electricity consumption at finely resolved timescales has become increasingly popular. Many existing algorithms underperform when clustering load profiles contain a large number of feature points. In addition, it is difficult to accurately describe the similarity of profile shapes when load sequences have large fluctuations, leading to inaccurate clustering results. To this end, this paper proposes a high-resolution load profile clustering approach based on dynamic largest triangle three buckets (LTTBs) and multiscale dynamic time warping under limited warping path length (LDTW). Dynamic LTTB is a novel dimensionality reduction algorithm based on LTTB. New sequences are constructed by dynamically dividing the intervals of significant feature points. The extraction of fluctuation characteristics is optimized. New curves with more concentrated features will be applied to the subsequent clustering. The proposed multiscale LDTW is used to generate a similarity matrix for spectral clustering, providing a more comprehensive and flexible matching method to characterize the similarity of load profiles. Thus, the clustering effect of a high-resolution load profile is improved. The proposed approach has been applied to multiple datasets. Experiment results demonstrate that the proposed approach significantly improves the Davies-Bouldin indicator (DBI) and validity index (VI). Therefore, better similarity and accuracy can be achieved using high-resolution load profile clustering.

show abstract

“…There are a lot of methods have been designed and proposed [7][8][9][10][11][12][13][14] to increase the quality of similarity measurement. Euclidean distance (ED) [7] and dynamic time warping distance (DTW) [8] are two typical examples.…”

Section: Introductionmentioning

confidence: 99%

A New Similarity Measurement Method for the Power Load Curves Analysis

Xin¹,

Zhu²,

Deng³

et al. 2021

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering

View full text Add to dashboard Cite

In order to improve the quality of the power load curves similarity measurement, a new similarity measurement method based on Euclidean distance is proposed in this paper . Among the commonly used similarity measurement methods, Euclidean distance is not sensitive to the fluctuation of the load curves, which results in the lack of shape measurement capability. For the numerical distribution on the timeline is not concerned, the dynamic time warping (DTW) distance is not accord with the requirement of the power system load analysis. Focus on those issues, the proposed method introduced a correction factor that contains the dynamic characteristics of the numerical difference between two power load curves without compromising time warping. The advantages and performance of the proposed method are evaluated by similarity computing and clustering analysis. As shown in the experimental results of similarity computing, the proposed method performs as same as ED and DTW, but the calculating time is less than DTW. In the clustering analysis, it also decreases the calculating time from 3.9 s to 0.595 s compared with DTW and shows better clustering effect that make the Davies-Bouldin index from 0.438 for ED and 0.325 for DTW to 0.249.

show abstract

A Power Load Clustering Method Based on Limited DTW Algorithm

Cited by 5 publications

References 15 publications

A density-based matrix transformation clustering method for electrical load

A density-based matrix transformation clustering method for electrical load

High-resolution Load Profile Clustering Approach Based on Dynamic Largest Triangle Three Buckets and Multiscale Dynamic Warping Path Under Limited Warping Path Length

A New Similarity Measurement Method for the Power Load Curves Analysis

Contact Info

Product

Resources

About