Since around 2012, multi-rotor drones have entered the field of power line inspection, becoming an extension of the visual and motor skills of inspection personnel and a handy tool. Multi-rotor drones can reach places that are inaccessible or inconvenient for inspection personnel, efficiently and safely completing the originally time-consuming and labor-intensive cyclical work, gradually freeing inspection personnel from high-intensity physical labor.
In recent years, in response to China's power Internet of Things strategy, the process of daily power operation and maintenance has been further refined into several stages: data collection, data processing, data analysis, information extraction, and conclusion reporting.
Currently, the development of drone-based power line inspection is still in a stage of "human inspection as the main method and drone inspection as a supplement." This is because drone inspection is still primarily manual, and inspection workers need to undergo certain training before they can use it. With the gradual increase in power operation and maintenance requirements, simplifying the inspection process and improving inspection efficiency have become important issues for the development of drones in the field of inspection. Automation and intelligence will undoubtedly be important directions for the development of drones.
Challenges and Vision
The current working mode of power line inspection drones mainly relies on manual operation and data processing, and there is still considerable room for improvement in inspection efficiency and the time spent on manual intervention. The current inspection operation mode and data processing mode face challenges in three main aspects.
They are:
1. Requires a significant amount of manpower (including initial training and on-site operation).
2. Data is not closed-loop (the data collected by the drone cannot be synchronized to the data processing unit in real time).
3. Information discontinuity (the information obtained from data processing cannot be transmitted back to the production system in a timely manner to guide production).
To address these issues, industry experts have proposed three approaches:
1. Unmanned aerial vehicles (UAVs) employ automated inspection to address the issue of high manual labor requirements during the data collection phase.
2. Introduce intelligent recognition technology to process data at the front end (drone)/back end (server) to obtain key information.
3. Introduce a drone operation management platform to close the information loop of the entire drone operation process.
Automating Inspection Processes: Teaching Robots to Fly, or Letting Them Fly on Their Own?
Traditional outdoor power line inspections relied on manual walking and climbing, which was labor-intensive and inefficient. The advent of drones can greatly solve this problem. Current automated inspections primarily use a flight path mode, supplemented by a closed-loop process to ensure inspection accuracy.
There are two main types of drone inspection route planning: teaching planning and offline planning.
Drones are actually a type of robot. In short, drone flight teaching planning involves programming the drone with the steps it needs to perform.
Another approach is offline planning, which is based on offline programming. This involves using software to reconstruct a 3D virtual environment of the entire work scene on a computer. The software can then perform custom actions based on task requirements. In conjunction with the operator's actions, it automatically generates the robot's motion trajectory, i.e., control commands. The trajectory is then simulated and adjusted in the software, and finally, control code is generated and transmitted to the drone.
Both planning methods have their advantages and disadvantages. As practical applications continue to emerge, the automation of drone inspection processes will become more intelligent.
Data processing automation: Deep learning for images will be key
During routine inspections, the key is to quickly identify detailed issues with equipment such as power lines and towers. However, during on-site operations, external forces often negatively impact the accuracy of drone inspections. These external forces include strong winds, bright sunlight, and signal interference. Therefore, introducing image recognition technology into the inspection process allows the drone to accurately identify the objects being photographed and automatically correct the shooting angle and adjust camera parameters, thereby improving system robustness.
A common application is to utilize the high-performance computing platform onboard the drone to analyze captured images in real time, process lightweight defects, and report them to the production system via the network. For example, DJI's high-performance computing platform, Manifold2, can be used with the Matrice M200V2 series to achieve customized intelligent recognition and analysis.
In addition, platform-level computing power can be used to perform detailed and in-depth analysis of image data, further uncovering detailed defects and ensuring operational security.
Automated Information Feedback: Central Hangars Are Coming
A crucial aspect of the inspection process is effectively managing all the data collected on-site and extracting it into useful information.
We know that data is not the same as information. Data may be massive or large in size, but information must be condensed and generalizable. Under current technological conditions, real-time transmission of large-size data samples via wireless links is not yet mature. Therefore, establishing a comprehensive UAV data management platform covering aspects such as flight missions, data accumulation, resource management, and integration is a fundamental task to effectively ensure the safety of power transmission and distribution lines and their equipment.
Flight mission management solves the problem of seamless information flow from the end to the front end. Through data accumulation, flight log data is stored on the platform for statistical analysis and query tracking. Through resource management, the correspondence between aircraft and organizational structure and personnel is realized, and the permissions of different roles are clearly defined. Through integration and fusion, UAV operations are used and combined with existing information systems to upload system data to the upper-level platform and leverage network synergy.
Similar platforms are expected to become the "central hangar" for drone power maintenance in the future.
Prospects for Full Automation of Power Operation and Maintenance
The automated flight paths of drones save significant manpower, freeing inspection personnel from heavy manual labor and allowing them to focus on higher-level data analysis and processing. This also optimizes the operational models and personnel structure of maintenance teams, improving the reliability of the power grid.
The drone is equipped with a high-performance computing platform for lightweight image recognition and processing, which can effectively improve the efficiency of problem detection and real-time processing during on-site operations; and output the problem analysis and upload it to the backend.
The addition of an online management platform has comprehensively improved the automation level of drone-based power maintenance. The combination of a drone-based integrated data processing platform and automated precision inspection is also an environmentally friendly and efficient inspection method, effectively promoting the process of environmentally friendly and green power inspection.
Currently, there are already relatively mature drone models in the industry used for automated inspection of power transmission line towers, and power companies and similar enterprises will continue to invest and deepen their expertise in this area.
With the continuous development of automated inspection processes, automated data processing, and automated information feedback, we can already see a positive outline of the future: perhaps, around 2020, we will be able to achieve full automation of the power operation and maintenance process.
Disclaimer: This article is a reprint. If it involves copyright issues, please contact us promptly for deletion (QQ: 2737591964). We apologize for any inconvenience.
