Research on resistance detection robot for UHV tension insulator string

Insulator inspection robots can realize automatic and rapid inspection of insulators of ultra-high voltage transmission lines, which is an important guarantee equipment for the continuous safe operation of the national power grid. Aiming at the needs of ultra-high voltage insulator string inspection, an insulator robot is designed to realize efficient and stable insulator string motion and insulator detection, and the main work includes proposing a new insulator string motion robot configuration and realizing its string motion and inspection functions in the form of a physical prototype. It has been proved that the robot has a large motion load and strong motion performance, and can be adapted to the complex working conditions consisting of a variety of insulators within a certain size range.


Introduction
The insulator is an important part of overhead high-voltage transmission lines.As a key stress carrier and insulation device in the transmission line, the operation reliability of the insulator is closely related to the safety of the transmission system [1].Due to the harsh environment, the surface pollution of the insulator makes it easy to reduce the pollution resistance strength of the insulator itself in fog, rain, ice, and other weather, resulting in transmission line pollution flash and other faults [2][3][4].With the development of industrial automation, testing robots are becoming more and more popular, which improves the efficiency of production operations.
At present, there are many insulator-testing robots researched at home and abroad.Korea Electric Power Research Institute developed an insulator testing robot [5], which is stable and reliable in motion, but its control is complex; Northeastern University developed a drape insulator testing robot [6] whose principle is simple and the volume is lightweight, but it is insufficient in the movement of stability, data processing, and carrying capacity.The crawler robot [7] developed by the Chinese Academy of Sciences Research Institute can detect efficiency, but only up to 15°sloped movement.In addition, most of the existing insulator robot is basically closed by gravity force, which limits their carrying capacity.For the existing insulator climbing robot exists deficiencies in precision, stability, flexibility and operation, and other issues, we made further study in the field of high-voltage transmission line insulator climbing robot stability, movement accuracy and flexibility, etc, which can effectively improve the robot efficiency to promote the industrialization of insulator climbing robot.Our climbing robot is equipped with a resistance detection mechanism that can realize the detection of zero-value insulators to ensure the safe operation of power lines.

Function design
2.1.1.Insulator clamping end.In ultra-high voltage transmission lines, overhanging insulators usually exist through double and four rows.Insulators are divided into two types according to their umbrella structure, bell-type and double-layer umbrella type as shown in Figure 1, which are characterized by the outer dimensions and step distance as the key dimensions.As shown in Figure 2, the adaptive locking clamping jaw adopts a linkage structure driven by a linear actuator.Since the jaws need to fulfill the function of clamping insulators and detecting insulators, the requirements for the jaws are good positioning accuracy and good clamping stiffness.The framework of the clamping jaw is made of aluminum profiles, through the double-layer threedimensional structure to ensure the stiffness of the clamping jaw and the strength of the connection with the horizontal joints, while the connection is made of carbon fiber plate board, to increase the stiffness and at the same time to ensure the quality of lightweight.The contact area is made of A-B silicone material to increase the contact friction to prevent slipping.The drive adopts a linear actuator driven by a BLDC, which is capable of force control and position locking.

Five-degree-of-freedom manipulator arm.
Due to the complex arrangement of insulator strings on high-voltage towers, to ensure that the inspection robot can operate more efficiently on insulator strings and improve the efficiency and flexibility of insulator inspection, it must be able to realize both single-insulator string inspection and switching inspection between insulator strings in the same direction on the same tower.Based on this, we designed an insulator inspection robot containing a five-degree-of-freedom arm, whose overall configuration is shown in Figure 3.The robot is divided into two parts: a high-precision five-degree-of-freedom arm and an adaptive clamping gripper.The overall dimensions of the robot are about 950×690×480 (unit: mm).To ensure that the inspection robot can complete all motion gaits during operation, the workspace design of the five joints is shown in Table 1, where the rotational axes of joint 2, joint 3, and joint 4 are parallel to each other, the rotational axes of joint 1 and joint 5 are parallel to each other and perpendicular to the rotational axes of joint 2, joint 3, and joint 4, and the workspace of joint 3 is greatly increased by eccentric design.

Resistance measurement module
Zero-value insulator detection is divided into the insulation resistance method, spark gap method, voltage distribution method, and infrared thermal imaging method.Among them, the insulation resistance method is the power failure detection method, and the other three are the live detection method [8][9][10].The resistance measurement module is mainly used for the zero-value detection of porcelain insulators.Insulators with resistance lower than 500 MΩ are considered to be low-value insulators and insulators with resistance lower than 300 MΩ are considered to be zero-value insulators.Therefore, by measuring the effective resistance value of insulators, the effectiveness detection of insulators can be realized.The resistance measurement module consists of a central processing module, a power management module, a DC boosting module, a detection probe module, and an insulator detection module.Through the comparison with megohmmeter detection, it is confirmed that this measurement module can realize resistance detection in the range of 100-1200 MΩ, which can meet the insulator detection requirements.Three-way voltage acquisition is used in the detection module, which samples the DC boost power supply and the two sampling resistors on both sides of the insulator, respectively.Due to the unstable voltage of the DC boost power supply, which has a large impact on the detection results, the power supply voltage and the voltage of the detection resistor must be sampled at the same time.Through the differential circuit, the voltage value is reduced to the voltage range (0.5-4 V) that can be detected more accurately, and the detection resistor value can be derived from the voltage value.The detection module and the functional block diagram are shown in Figure 4.

Localization on insulator string
The common features of lidar are line segment features, break point features, corner point features, etc. Considering the many types of insulator strings involved and the lack of uniform size standards, it is difficult to describe them mathematically and accurately.Therefore, it is relatively difficult to estimate the position and correlate the data based on the line segment features.Comparatively speaking, breakpoints and corner points are simple in mathematical description, complete in detection, easy in feature matching, and can be used to realize insulator string feature detection.
Currently, the more commonly used segmentation algorithms can be divided into two categories: sequential habit algorithm and recursive algorithm [11].In this project, the point distance-based method (PDBM) is used for data point segmentation, i.e., by comparing the Euclidean distance of two neighboring points with the threshold value, to determine whether the two points can be divided into the same group.After obtaining the point cloud data of the target insulator, the edge features can be used to back-calculate the hand claw relative to the robot's position.The point cloud data of the target insulator is shown in Figure 5.

Inchworm gait simulation
In this paper, we designed a five-degree-of-freedom robotic arm for UHV ceramic insulator detection, and the multi-degree-of-freedom design allows the robot to have a very flexible movement [12], which can cope with the walking environment of the high-altitude insulators of the UHV power towers.Inspired by the movement patterns of insects in nature, we designed an inchworm movement gait [13] for the robot to crawl on the high-altitude insulator strings.Based on the robot's motion environment, we conducted simulation experiments of insulator string crawling with an inchworm gait at first, to verify that the robot can adapt to the unstructured surface of insulators and crawl stably over them.
A 3D model of the insulator string was built according to the actual dimensions and imported into Webots, the designed insulator string robot model was imported into Webots and the simulation physical parameters were set.In the simulation environment, the robot moves on a single insulator string and the inspection process is as follows in Figure 6: The process is repeated until all insulators on the insulator string are detected.The simulation results show that the insulator string inspection robot can realize the stable movement and detection of inchworm gait on a single insulator string.

Results
To verify the actual working effect of the robot, the experimental environment of a single insulator string is set up offline, and the actual test of geometrical movement and resistance detection is carried out.After analyzing the specific job content of insulator string detection, we use Bessel curves to realize the trajectory generation of the robot.The generalized formula of the Bessel curve is shown in Figure 7: Considering the motion requirements of this robot, six control points and fifth-order Bessel curves are used for curve trajectory generation, four control points, and third-order Bessel curves are used for linear trajectory generation, and the generated trajectory points can meet the requirements of trajectory smoothing [14], velocity, and acceleration continuity.Its actual movement is shown in Figure 8.   8, the motion of the robot prototype process can complete the robot in the insulator string with the inchworm movement gait.The experiment successfully showed that the robot can move on the insulator string surface with the inchworm gait.
After the experiment, the functional parameters of the robot are shown in Table 2.

Conclusion
In this paper, a multi-degree-of-freedom insulator detection robot adapted to the detection of tension insulator strings is designed, and the insulator robot's motion performance experiments under the building conditions are carried out.The conclusions are as follows: 1.The structure of this robot can meet the requirements of climbing movement and detection accuracy.The clamping claw structure of the robot has large stiffness and small deformation, which does not affect the detection movement and the operation of the robot.Its strength can meet the use of the insulator detection robot to achieve the optimization of the structure of the insulator detection robot's lightweight.
2. Through kinematic simulation analysis, the dangerous parts of the insulator robot and the links that should be focused on are obtained, which lays a certain foundation for further structural design and optimization.
3. The robot moves smoothly with a ruler-wig gait, and the hand claw clamping joint motor has no current impact, which shows the ability of the insulator inspection robot to realize the efficient movement on the insulator strings of the ultra-high-voltage transmission lines.

Figure 1 .
Figure 1.The robot repairs and cleans the insulator string.

Figure 3 .
Figure 3. Five degrees of freedom insulator detection robot.

Figure 6 .
Figure 6.The robot simulates the string.

Figure 7 .
Figure 7. Locus curve of the claw.

Figure 8 .
Figure 8.The robot runs on the string.

Table 1 .
Working space for each joint (unit: degrees).

Table 2 .
Working space for each joint (unit: degrees).