Robotic Monitoring of Power Systems


Economically effective maintenance and monitoring of power systems to ensure high quality and reliability of electric power supplied to customers is becoming one of the most significant tasks of today's power industry. This is highly important because in case of unexpected failures, both the utilities as well as the consumers will have to face several losses. The ideal power network can be approached through minimizing maintenance cost and maximizing the service life and reliability of existing power networks. But both goals cannot be achieved simultaneously. Timely preventive maintenance can dramatically reduce system failures. Currently, there are three maintenance methods employed by utilities: corrective maintenance, scheduled maintenance and condition-based maintenance. The following block diagram shows the important features of the various maintenance methods.

Corrective maintenance dominates in today's power industry. This method is passive, i.e. no action is taken until a failure occurs. Scheduled maintenance on the other hand refers to periodic maintenance carried out at pre-determined time intervals. Condition-based maintenance is defined as planned maintenance based on continuous monitoring of equipment status. Condition-based maintenance is very attractive since the maintenance action is only taken when required by the power system components. The only drawback of condition-based maintenance is monitoring cost. Expensive monitoring devices and extra technicians are needed to implement condition-based maintenance. Mobile monitoring solves this problem.

Mobile monitoring involves the development of a robotic platform carrying a sensor array. This continuously patrols the power cable network, locates incipient failures and estimates the aging status of electrical insulation. Monitoring of electric power systems in real time for reliability, aging status and presence of incipient faults requires distributed and centralized processing of large amounts of data from distributed sensor networks. To solve this task, cohesive multidisciplinary efforts are needed from such fields as sensing, signal processing, control, communications and robotics.

As with any preventive maintenance technology, the efforts spent on the status monitoring are justified by the reduction in the fault occurrence and elimination of consequent losses due to disruption of electric power and damage to equipment. Moreover, it is a well recognized fact in surveillance and monitoring fields that measurement of parameters of a distributed system has higher accuracy when it is when it is accomplished using sensing techniques. In addition to sensitivity improvement and

subsequent reliability enhancement, the use of robotic platforms for power system maintenance has many other advantages like replacing man workers for dangerous and highly specialized operations such as live line maintenance.


Generally speaking, the mobile monitoring of power systems involves the following issues:
SENSOR FUSION: The aging of power cables begins long before the cable actually fails. There are several external phenomena indicating ongoing aging problems including partial discharges, hot spots, mechanical cracks and changes of insulation dielectric properties. These phenomena can be used to locate the position of the deteriorating cables and estimate the remaining lifetime of these cables. If incipient failures can be detected, or the aging process can be predicted accurately, possible outages and following economical losses can be avoided.

In the robotic platform, non-destructive miniature sensors capable of determining the status of power cable systems are developed and integrated into a monitoring system including a video sensor for visual inspection, an infrared thermal sensor for detection of hot spots, an acoustic sensor for identifying partial discharge activities and a fringing electric field sensor for determining aging status of electrical insulation. Among failure phenomena, the most important one is the partial discharge activity