Solar power forecasting using robust kernel extreme learning machine and decomposition methods

Majumder, Irani; Bisoi, Ranjeeta; Nayak, Niranjan; Hannoon, Naeem

doi:10.1504/ijpec.2020.10027381

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The operation status monitoring of urban smart grids is a key link in smart power systems. As an efficient learning algorithm, ELM has shown superior performance in processing speed and generalization ability, and is therefore widely studied and applied in power grid state assessment [7]. Scholars have conducted various studies on ELM and achieved significant results, especially in the rapid processing of power grid big data and accurate prediction in complex environments.…”

Section: Related Workmentioning

confidence: 99%

“…, including the encoding bias b , decoding bias d , and corresponding weight matrices w and w′ . The data stream is first converted into low-dimensional encoding through an encoder, and then reconstructed by a decoder to complete the data conversion from high-dimensional to low-dimensional and then to high-dimensional in equation (7).…”

Section: Visible Layer Neuronmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation and Analysis of Urban Power Grid Operation Status Based on Online Sequence Extreme Learning Machine and Self-Coding Network

Lin,

Zhu,

Shi

2024

IEEE Access

View full text Add to dashboard Cite

In terms of processing high-dimensional data for evaluating the operational status of urban power grids, this study aims to improve the accuracy and processing speed of the evaluation. Firstly, the combination of Analytic Hierarchy Process and self-coding network was used to reduce data dimensionality and simplify the complexity of model input. And improvements were made to the online sequence extreme learning machine algorithm to enhance the model's adaptability to dynamic data streams and real-time learning capabilities. These results confirmed that the reduced dimensionality of the self-coding network could effectively capture the core information of the 85 dimensional raw data. The prediction accuracy of the weighted online sequence extreme learning machine algorithm was significant. Its predicted average value of 0.0994 was almost the same as the actual value of 0.0991, which was better than the comparison between the standard online sequence extreme learning machine's predicted value of 0.0946 and the actual value of 0.0986. In addition, the weighted online sequence extreme learning machine converged faster and had higher prediction accuracy than traditional support vector machine after 450 iterations. These results confirmed that the combination of dimensionality reduction technology and dynamic update learning mechanism could effectively improve the accuracy of urban power grid state assessment. This study provides a new technical solution for real-time monitoring and fault prediction of urban power grid operation status, which is very important for ensuring the safe operation of the power grid.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Visible Layer Neuronmentioning

confidence: 99%