https://orcid.org/0009-0006-4598-7306
![Sample our Engineering & Technology journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5933/TOC-Subject-Sample-2021-Engineering-Tech.jpg"})
ABSTRACT
To evaluate the health state or failure of the converter valve in the early stage and realize the goal of condition-based maintenance rather than regular maintenance, in this work, we are the first to propose a state evaluation network for UHVDC converter(SE-Net). Taking full advantage of the working principle and data characteristics of the converter valve, in the proposed SE-Net, we customize the network architecture with emphasis on the one-dimensional operation. The multi-dimensional features extracted from the converter valve are fused according to the channel weight to improve the accuracy of health state evaluation. This paper takes the converter valve of a ± 800kV converter station as the research object and builds a fine particle size simulation platform. The platform collects the data of 19 measuring points such as the voltage and current of the converter valve. Based on the measuring data, the SE-Net can extract effective features, and identify the health state of the converter valve. The experimental results show that SE-Net accurately and effectively identifies eight health state types of the converter valve, with an average accuracy of 95.3%.
CO EDITOR-IN-CHIEF:
ASSOCIATE EDITOR:
Nomenclature
| 1D | = | One dimension |
| AC | = | Alternating current |
| CNN | = | Convolution neural networks |
| DC | = | Direct current |
| EMD | = | Empirical mode decomposition |
| GPU | = | Graphics processing unit |
| HMM | = | Hidden Markov models |
| KNN | = | K-nearest neighbor |
| L | = | Sample length |
| SE-Net | = | State evaluation network |
| SVM | = | Support vector machines |
| TPU | = | Tensor processing unit |
| UHVDC | = | Ultra-high voltage direct current |
Disclosure statement
No potential conflict of interest was reported by the author(s).