Journal
SOFT COMPUTING
Volume 25, Issue 8, Pages 6401-6413Publisher
SPRINGER
DOI: 10.1007/s00500-021-05632-5
Keywords
Electricity load; Prediction; Boost clustering; Time series; k-means; HAC
Categories
Funding
- National Natural Science Foundation of China [61672439]
- Fundamental Research Funds for the Central Universities [20720181004]
Ask authors/readers for more resources
Power load prediction is crucial for energy management in power systems. In this study, a boost clustering-based approach is proposed to enhance the traditional k-means algorithm and predict power load by clustering users and summing up the predicted results. Experimental results show the effectiveness of this approach over direct prediction methods.
Power load prediction which helps make the optimal decision for energy management is of great significance to the safe, reliable, and economical operation of the power system. It is also a challenging task; however, if every large customer of a special transformer is modeled and forecasted for power load, a huge amount of calculation work is needed and it is not practical. Therefore, in this study, we propose a boost clustering-based approach for the prediction of power load. The traditional k-means algorithm is enhanced, and the initial cluster centers are determined in advance instead of random selection. Then, the enhanced k-means paired with the HAC algorithm are used for the clustering of power consumption users. Next, the power load of each group is predicted after the users are clustered into the different groups, and the predicted results of each group are finally summed to obtain the prediction value of the power load. Experimental findings demonstrate the validity of the proposed procedure, and the boost clustering-based approach significantly outperforms the direct prediction approach in the empirical analysis.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available