21, rue d’Artois, F-75008 PARIS B3-114 CIGRE 2016

http : //www.cigre.org

An Autonomous Intelligent Robot for Electronic Equipment Inspection Used in

Substation

L. LI, G.B. WU, S.Y. MU, M.C. FU, J. L. ZHAO, R. GUO, J.X. LI
State Grid Shandong Electric Power Research Institute
State Grid Shandong Electric Power Company
Shandong Luneng Intelligence Technology Co., Ltd.,

SUMMARY

Substation equipment inspection is an important daily work for the power industry. With the
development of the robotics and the sensor technology, more and more intelligent robotics have been
used in the power industry. The inspection robot is an autonomous robotic system for inspecting
substation equipment to help or replace workers, and which we called “SmartGuard”. The robot
system consists of a robot and a data center.
With a mobile platform, a control system with location and navigation module, a visible-light camera
and an infrared thermograph based inspection system, the robot can run autonomousily in the
substation inspect the equipment. SmartGuard has four wheels for the robot travelable. With an
independent steering mechanism, the robot could flexibly steer. Protective designing of the inspection
system and the whole shell make the robot adapting to environmental changes,and prevent the dust,
dirt and moisture. Based on the large capacity battery, the robot could continuous working 8~10h. In
order realize safe and reliable charging, special charger and charging room has been developed.
Magnetic Guidance is used to guide the robot, and the localization is implemented by using radio
frequency identification technology. With the development of new sensor technologies, now we use
laser and vision navigation system to sense its environment and deal with environment uncertainty.
The inspection system of the SmartGuard consists of an infrared thermal imager, a visible light camera,
a pan & tilt and an image processing module. The online infrared thermal imager is used for hot spot
monitoring in electric power substation. Based on the visible light image robot could recognize the
station of the equipment such as the switch, isolator in substation and the abnormal appearance of the
equipment.
The Data center mainly contains a human–computer interaction (HCI) interface, a comparison output
interface, an analysis interface, and a system and tasks configuration interface. The result of the
intelligent inspection system and the diagnosis report also could be obtained from the data center.
Wireless network is using between the robot and the data center in the communication system of the
robotics system.
Until 2015, there are more than 300 SmartGuard robots working in substation in China.
KEYWORDS

Substation, robot, automatic inspection, SmartGuard

[email protected].
 I. BACKGROUND
Traditional manual inspection and operation in substations requires plenty of man power and
has issues with efficiency and quality. It also has great potential safety problems under bad
weather conditions, such as rain, snow, fog or mist. With the rapid increase of the cost, the
difficulty are rising. With increasing demands for substation and unattended operation, it is
necessary to develop new technologies for inspection and operation in substations.
The rapid development of robotics technology makes it possible for its application in electric
power industry. The robot can also replace or assist human to complete equipment inspection,
analysis and diagnosis, remote control, supplementary maintenance and live working, etc.,
and effectively reduce inspection cost, work intensity and safety risk of the operation and
maintenance.
II. THE STATE OF THE ART
Different types of the robots have been developed and working in the power industry in many
countries. In Japan, the distribution live working robot has been developed, and in Korea, the
insulator detection robot has been applied to the overhead line. And in China, the unmanned
helicopter base robot has used for transmission line inspection etc.
The robotics working in substation are relatively mature. Four generations of robots have
been working in substation from 100kV~1000kV[1][2][3] from 2001 to now, in China. In other
countries, such as Japan developed a patrolling robot called “Big Mouse” for substation in
1989[4]. In 2008, São Paulo University in Brazil developed a mobile robot to monitor hot spot
of the equipment in power substation[5]. In 2012, field tests were performed in Hydro-Québec
substations using a mobile robot developed by IREQ[6].
The international conference CARPI has been held successfully for three sessions since 2010
to 2015, and more and more solution developers (universities, research centres and industry
suppliers) and power industry end-users (utilities, including power generation, transmission
and distribution) pay more attention to the robotics technology.
III. THE CHALLENGES OF IMPLEMENTING ROBOT WITHIN A SUBSTATION
Because of the substation constructed with High voltage equipment, so the robot developed
instead of works to operation maintenance in the substation should match some challenges as
below:
 First, the robot system should be operating autonomous, without too much human
intervention, so the system contains an independent power supply system and a battery
management system[7], which could supply stable and reliable power. And the ego state
diagnosis system also be need, to avoid the failure of the unit. The robot in the
substation could operation fit to the outdoor environment in different temperatures and
different weathers, and etc.
 Second, the robot running in the specific trace and area, live parts should be guarded or
enclosed. So the robot system should with a moving platform, and a location and
navigation system to control and guide the robot running in setting path.
 Third, the inspection and troubleshooting based on the robot should in accordance with
applicable industry standards. So the results of the robot system would be reliable.
 Fourth, communication of the robot system should be stable and reliable, and suitable
to the work environment, which with the basic safety requirement.
 Fifth, The GUI of the robot system should be nice, including the task setting and
inspection and diagnose result display.

1
IV. OVERVIEW OF THE ROBOT SYSTEM
From the Fig.1(a) the framework of the system, we could see the whole robotic system consist
of two parts: a robot and a data centre. The architecture of SmartGuard[8] is shown in anther
figure Fig.2(b). The robot runs in substation to inspect equipment with its inspection module.
In order to help or replace worker’s daily patrolling work, the robot captures an infrared
thermal imager through a visible light camera, which is installed on a pan & tilt. Based on the
visible-light images and infrared images, the functions of equipment recognition, abnormal
detection, image fusion and instrument automatic recognition are developed, which could do
the inspection intelligently replacing workers. The inspected information is sent to the data
center through the wireless communication module. In the data center, the graphic user
interface (GUI) is implemented by the presentation layer.

The chargeable room

The navigation path

master-control room

The robot
(a) The operating mode of the robot sytem (b) SmartGuard
Fig.1 The SmartGuard and the data center of the robotic system

L/W/H :870/550/1060mm Range of the Pan-Tilt:horizontal direction ±180°

Weight :-70kg pitch angle -10°~ +90°
Speed :0~1.5m/s Working ranging :equipment area in substation
Gradeability :≦15° working temperature range:-20℃~+50℃
IP :IP55 Maximum working rainfall : heavy rain
Continuous working period :8~10h Maximum working snow :snow thickness 50mm
PositionaI accuracy :±1cm Maximum working wind :28m/s
transmission distance :10km（unboarded） Maximum fording depth :120mm
Tab. 1 technical parameters of the SmartGurad 4S
V. KEY TECHNOLOGY OF THE ROBOTIC SYSTEM
A. The body design of the robot
There are several different platform structures, including rail-mounted, wheel, crawler, etc.
The SmartGuard has four wheels for the robot. The two front wheels are the driven wheels,
which are driven by two motors. The rear wheels are two follower wheels, which are
omnidirectional wheels. And we also research on the four driven wheels, which could be
more traversable. With an independent steering mechanism, the robot could flexibly steer.
Protective designing of the inspection system and the whole shell make the robot adapt to
environmental changes, and prevent the dust, dirt and moisture, with dust filter adapted in the
vent, which makes the system attend 55 grade in ingress protection system.
Based on the large capacity battery, the robot could continuously work for 8~10h. In order to
realize safe and reliable charging, special charger and charging room has been developed. The
energy management system is also necessary to predicting the remaining battery power, with
which the robot could track back to charger automatically.

2
B. Location and navigation control system
Magnetic Guidance is used to guide the robot, which is not affected by falling or
accumulation of rain or snow. And localization is implemented by using radio frequency
identification (RFID) technology, which could reach ±3cm. This system has been tested and
proven to be robust under a wide variety of operating conditions. With the development of the
new sensor technologies, now we use laser and vision navigation system to sense its
environment and deal with environment uncertainty. With the new navigation system, the
robot could walk any position in the substation, detect and avoid the obstacle automatically.
A laser navigation system based on map matching method has been developed[9] . An iterative
MLE(Maximum Likelihood Estimation) algorithm is used to generate global electronic map,
which is represented by grids. Firstly, use the mapping software to calculate the maximum
likelihood position of the robot by using the odometer data, and then match the current laser
rang data to the map which has already generated, finally, use the ICP method to search the
best match, once the best match is found; it will be merged with the previous map to expand
the area of the global map. After the map is constructed, an EKF (Extended Kalman Filter)
algorithm is employed to assist the positioning process by using the global electronic map
already created. Some experimental tests have been carried out in a 500kV substation is
shown in Fig.2[7].

(a) Substation (b) Verify the location precision (c) Verify the head angle precision
Fig.2 The scene and data of Experimental Test
A stereo vision system is used to capture more information of the surroundings. In the
substation, the road is regular and the robot is running along the road. The environmental
information was more important for the self-adaption robot. The stereo camera was installed
on the front of the robot to capture the images. In order to get the edge of the road, the Inverse
Projective Mapping (IPM) has been used to remove the perspective effect and then we use
canny edge detection method to get the edge of the IPM image [10]. Then based on the line
feature we detect the road edge and the road plane regain, which could navigate the robot
running. Simultaneously based on the line direction and the parallel of two boundaries, we
take the right road edge and remove the disturbance of the other edges from the road ground.
In order to make the obstacle detection result with more exactly and less time complexity, we
take the stereo IPM method to detection the obstacle above on the road plane.

(a)Sence image (b) IPM image (c) Canny edge detection (d) Hough line detection
Fig.3 The images of the original image inverse projection from the image plane to the road plane
and the processing flow.

3
C. The inspection system
The inspection system of the SmartGuard consists of an infrared thermal imager, a visible
light camera, a pan & tilt and an image processing module. For different equipment, the
camera uses different focal lengths to capture clear image. As follows, we give an
introduction of the visual serving control system and the Inspection & Diagnosis system
based on the sensor signal processing.
Because of the motion error of robot and the re-positioning errors of pan & tilt, the view filed
of the camera and imager could be shifted. Sometimes the target equipment would be out of
view, especially when the camera has a large focal length. We propose an image based visual
servo control system to precisely make the cameras with the pan & tilt to capture the
substation equipment to be inspected. The vision control system consists of an image
acquisition system, a pan & tilt and an embedded servo control system [11]. The proposed
approach firstly uses SURF local feature to take the offset of the new acquired image and the
template image, which is invariant to light, scale and rotation. Without the calibration of the
pan & tilt and the vision system, we calculate the Jacobian matrix based on the test movement.
Then we take the servo until the image feature reach to the target image. Using this image
based vision system, SmartGuard can avoid the effect of the motion error and the repeating
errors of pan & tilt system, and take the precise image of the power equipment for automatic
optical inspection system.
The bellows show what can the inspection system do and the method we developed.
 The online infrared thermal imager is used for hot spot monitoring in electric power
substation. The temperature of equipment ontology such as transformer, mutual inductor
and switch contact, busbar connection head, can be detected through an integral and
scanning temperature acquisition. Operating staff can set critical temperature measuring
points where defects may occur for each type of equipment in the background, and the
robot can perform an all-round, multi-angle fixed-point and accurate temperature
measurement.
 By the HD camera, based on the image processing, pattern recognition technologies,
robot could recognize the station of the equipment, just like the switch, isolator in
substation[12]. The abnormal appearance of equipment in substations was detected by
change detection method based on the normal image from the same equipment. We
present a fast elliptical target extraction method based on boundary that can effectively
solve the extraction of the circular target in the image, to help reading meters[13].
 A bicubic interpolation algorithm is used to fuse the infrared image correctly over the
visible image, so that people could get the temperature information and color information
of equipment simultaneously through one image. A diagnosis module is used to monitor
important equipment through historical data comparison.
 The sound of the equipment under operation can be gathered and abnormalities inside the
equipment can be identified by the robot through analyzing and comparing with the
equipment abnormal sound characteristic library.

a. Switch status b. Meter station c. Isolator status d. Foreign matter

Fig.4 Inspection results

4
D. Communication system
The communication system between the robot and the data center is a wireless network
system, which in the frequency band: 5.8GHz, with bandwidth: 50Mbps, and the max
transmission distance of 10km. Based on the communication system, the data and the status of
the robot could send to the data center, the control command could from the data center was
send to the robot. Based on the wireless bridge added on the robot, the data from intelligent
instruments also collected and send to control room, which could help the intelligent
instrument using in substation.
E. Data center
With the TCP/IP protocol, the images and the video from the camera, infrared thermal imager
and also the signal from the sound pick-up, were transformation to Data Center. All the data
processing, analysis and decision done by the Data Center, including creating historical curve
and diversity report to facilitate diagnosis and analysis for operation staffs based on the
database. The abnormal report would be generated. The task designing and path setting carry
out though human-computer interaction interface.
Based on the IEC-61850 standards, the information from the robot system Data Center
transferred to the Management Information Service (MIS) of the substation. Through the
robot, monitoring center could take the guardianship through remote monitoring and
remote control to the workers.

Fig.5 the GUI of the Data center

VI. THE APPLICATION OF THE ROBOT IN SUBSTATION
Because of the shortage of the human resources and the aging of the assets, the utilities and
asset owners have to reduce the man-hours required for the operation and maintenance in
order to reduce the costs. So the robot patrolling in substation as an alternative method is an
effective solution.
Let us see the application of the patrolling robots in China. The research began from 2006,
and in 2009, the SmartGurad robot was successfully applied in several substations to
demonstrate its performance. Different types of robots were developed with the intelligent
visual inspection system and automatic running system. Until 2015 the State Grid Corporation
of China has more than 300 robots working in the substations from 110 kV to 1000kV[14]. The
robots have finding more hundreds of defects or hidden dangers, which could avoid black out
and reduce the economic loss, just like a temperature abnormal detection: a lead wire T joint
the temperature was 120.32℃，has exceed the max temperature 70℃.
the robot can also replace or auxiliary artificial to complete the special inspection of
substation equipment in highland or cold geographical environment, under bad weather
conditions such as wind, fog, snow or thunderstorm, and effectively reduce labor intensity and
security risk of the operation and maintenance staffs.
5
VII. CONCLUSIONS
So, robotics is a viable, effective, and safe way working instead of people on the hard and
dangerous work. The International Conference on Applied Robotics for the Power Industry
have been hold for three times from 2010 to 2015, and more and more solution developers
(universities, research centers and industry suppliers) and power industry end-users (utilities,
including power generation, transmission and distribution) pay more attention to the robotics
technology. So in the future, it will be an important developing direction for the power
industry.
The new robot can be equipped with the modular working tools to conduct living works such
as cleaning the post insulators, mutual inductors, surge arresters[15]. And other equipment, and
other live work projects like cleaning filthy surface, bolt fastening, bird nests removal and
other foreign matters.

BIBLIOGRAPHY

[1] X. Li,S. Lu, H. Wang, R. Guan, G. Chu.Introduction of the Substation Intelligent Mobile Robot.
Shandong Electric Power, 2005,143:3-6.
[2] Lu S, Qian Q, Zhang B, et al. Development of a Mobile Robot for Substation Equipment
Inspection[J], Dianli Xitong Zidonghua, 2006,30(13):94-98.
[3] Guo R, Li B, Sun Y, et al. A patrol robot for electric power substation[C], Proc. of the
International Conference on Mechatronics and Automation (ICMA), 2009:55-59.
[4] Takahashi H. Development of Patrolling Robot for Substation [J]. Japan IERE Council, Special
Document R-8903, 1989:10-19.
[5] Pinto J K C, Masuda M, Magrini L C, et al. Mobile robot for hot spot monitoring in electric
power substation[C],Transmission and Distribution Conference and Exposition, 2008:1-5.
[6] Jean-Francois Allan, et al. Robotic Systems Applied to Power Substations – A State-of-the-Art
Survey. Applied Robotics for the Power Industry (CARPI), 2014 3rd International Conference
on IEEE, 2014:1-4.
[7] Binhai Wang, et al. The specific Requirements of Applied Robots for the Power Utility. Applied
Robotics for the Power Industry (CARPI), 2010 3st International Conference on IEEE, 2010:1-4
[8] B. Wang, R. Guo, B. Li, L. Han, Y. Sun. "SmartGuard: An Autonomous Robotic System for
Inspecting Substation Equipme Journal of Field Robotics, 2011, 29(1):123–137.
[9] Peng Xiao, et al. Design of a Laser navigation System for Substation Inspection Robot. IEEE
International Conference on Control and Automation(ICCA), 2013 10th International
Conference, 2013 :739-743.
[10] Li, Li, et al. Road edge and obstacle detection on the SmartGuard navigation system. Applied
Robotics for the Power Industry (CARPI), 2014 3rd International Conference on IEEE, 2014:1-
4.
[11] Li L, Wang B, Li B. The Application of Image Based Visual Servo System for SmartGuard[C],
Control and Automation (ICCA), 2013: 1342 – 1345.
[12] Wanguo Wang, Binhai Wang, Zhenli Wang, Li Li. Status Recognition of Isolator Based on
SmartGuard[C]. International Conference on Digital Image Processing, 2013: 82-89.
[13] Wanguo Wang, Zhengfei Xu, Shirong Wang, Zhenli Wang. Optimal Ellipse Fitting Method
Based on Boundary of Least Square Approach[C]. International Conference on Information
Technology and Software Engineering, 2012, vol(3): 951-959.
[14] Yang Xudong, Huang Yuzhu, Li Jigang，Li Li，Li Beidou.A State-of-the-art of robots applied
in substations for equipment inspection. Shandong electric power, 2015, 1(42):30-34.
[15] Jianjun Su. Dianli Jiqiren Jishu[M]. China Electric Power Press, 2015.