……….………………………………….
01
Background and Significance of the Results
Transmission lines, as an important infrastructure in the power industry, are a vital component of the power grid. Their safe and stable operation is related to the reliability of the power system and the sustainable development of the national economy. If there are issues in the construction process of key components in transmission lines, it could jeopardize the stability of the entire power grid. Therefore, researching the appearance recognition methods for construction processes and developing automatic recognition tools are particularly important.
…………………………………………..
…………………………………………..
Currently, the identification of key components in transmission lines mainly relies on inspection personnel climbing towers for line checks and X-ray imaging methods. The workload for inspection personnel climbing poles is large and inefficient; while X-ray imaging is intuitive for detection and defects are easy to qualify, this method has high detection costs, and the lines must be powered off, making it difficult to meet the needs for large-scale quality inspections.
To address the above issues, this project follows the “Hydraulic Crimping Process Specification for Overhead Conductors (below 800mm2) and Ground Wire” (DL/T 5285-2018), and through on-site comparisons and crimping test studies, has developed an automatic recognition method and software based on the proportional relationship of the crimping zone and non-crimping zone of strain clamps. This method can quickly and automatically determine whether the crimping process complies with crimping specifications through drone inspection photography, thus saving a significant amount of time and labor for quality control checks of strain clamps.
…………………………………………..
…………………………………………..
02
Principle of Results and Effectiveness Verification
2.1 Principle of Results
The junction between the crimping zone and the non-crimping zone serves as a control area for determining whether it is qualified. The start and end of components and the centerline of components are indicated by red lines, and the qualified range of each area (such as Area A in the figure, front end of the groove, rear end of the groove) is represented by rectangular boxes.
Detection Method
Software
▼
Detection Tool
…………………………………………..
2.2 Effectiveness Verification
To determine the effectiveness of this tool, this project selected six different models of hydraulic strain clamps for experimental research, producing one standard sample for each model of hydraulic strain clamp, as well as ten test samples with different defects, and used the created templates to determine whether the ten defective hydraulic strain clamps were qualified. Finally, X-ray detection was used to check if the strain clamps were qualified. The effectiveness of this tool was verified through a comparative analysis of the results from both tests.
(b) Defective Sample
(c) Standard Template
From the comparison of the detection results of the two tools, it can be seen that the detection and identification results of the two tools are consistent, proving the effectiveness of the tool developed in this project.
After a comparative analysis and verification of all samples, the recognition accuracy of this tool is 95%, greatly saving costs and improving work efficiency.
Due to the limited number of test samples in this project, the detection accuracy is relatively low, and over 90% of the test models for strain clamps are those used in the Shantou Bureau’s operating lines. In future research, the number of test samples will be increased, and various types of strain clamps will be supplemented.
03
On-Site Application
▲ Figure: Detection of Important Cross-Over Strain Clamps on the 500kV Lu-Shan First Line
(Click to enlarge)
04
Economic Benefit Analysis
Direct Benefits:
For the 500kV Lu-Shan First Line, there are 10 important cross-over strain towers, with four split conductors, totaling 240 strain pipes.
Indirect Benefits:
05
Application Status and Promotion Prospects
-
Promotion Prospects
-
Innovative Verification
This result, after project acceptance, has been further explored in application with experts from the Guangdong Power Grid Co., Ltd. Machine Patrol Management Center, combining automatic driving, image AI recognition, and front-end recognition feasibility, resulting in a new generation of results based on automatic driving and image front-end recognition for the appearance ratio recognition method of transmission line strain clamps and the automatic driving system, with three invention patents applied for (accepted) and two papers published (accepted).
-
Effectiveness Verification
-
Promotion Prospects
This project developed tool is low-cost and does not require power outages for cooperation, making it applicable for all strain tower inspections.
This project developed tool has a fast detection speed and can be reused, allowing real-time tracking of the crimping quality of construction units.
This result has been applied in multiple power enterprises and has received high application evaluations, meeting the conditions for nationwide promotion.
This article is edited and organized based on the PPT of the same name from the Shantou Power Supply Bureau of China Southern Power Grid. The copyright of the images and texts belongs to the original author. If there is any infringement, please contact us, and we will delete it in a timely manner. We welcome more friends to share their work experiences, work insights, and work ideas. This article reflects the personal views of the original author and does not represent the views of this public account nor bear responsibility for its authenticity.