Ultraviolet (UV) discharges and thermal hotspot activity are indicators of defects on high voltage components. Given this, being able to efficiently visualize their source has long been a goal of power engineers. Thermal imaging using infrared (IR) detectors has been used for years and enables location of high temperature problem areas or ‘hotspots’. But past methods of corona detection, including audible noise, was not able to pinpoint the source or location of discharge. The advent of modern corona imaging systems changed that since knowing that there is corona discharge present and finding its location and cause are different challenges. Often, the sources of hotspots and UV discharges are also different and this requires different technologies to visualize them. At the same time, it is easier to assess the severity of a problem if it can be seen and compared against visual results of the same defect. The following overview is based on a 2010 contribution to INMR by experts at Uvirco Techologies, Eskom and CSIR in South Africa.
Principles of UV Discharge Generation & Detection
During stormy weather, sailors noticed flame-like reddish or bluish lights on the tops of masts and on the end of yardarms. They associated this with benign protection and named it after their patron saint, St Elmo. Because the rounded head of the mast wore this light like a crown, it was later referred to using the Latin word, corona. Years later, as sources of high voltage electricity were developed, the same light-like phenomena were observed in the laboratory and also referred to as corona. The term ‘corona’ is now commonly used to describe electrical phenomena that occur either internally or externally in or on network components, resulting in UV discharges.
Understanding the technology to visualize corona requires knowledge of the physical process of UV discharge generation as well as of the electro-optic principles for detection. UV discharges are generated by high electric field that surrounds HV equipment and ionizes nitrogen molecules in air. During the ionization and de-ionization process, photons are emitted to the surrounding air at a wavelength characteristic of the spectroscopic properties of nitrogen. The nitrogen molecules emit photons mainly in the ultraviolet region, with dominant peaks at 298 nm, 347 nm and 358 nm. Emissions at those short wavelengths are barely visible since receptors in the eye are sensitive only from 400 to 700 nm. Hence, shorter or longer wavelengths go undetected. Most commercial grade lenses in binoculars and camera lenses are poorly transparent for ultraviolet radiation and suppress any UV discharges that might be seen with such equipment.
Acoustic detection of electrical discharges was the earliest technology to link lightning flashes and the associated sound of thunder. With the dawn of the electrical age, it became apparent that high voltage sparks in air produce acoustic emissions equivalent to lightning during a storm but on far reduced scale. As increasingly higher voltages came to be used, corona at highly stressed regions on electrical equipment and conductors could be seen by eye, in the dark, and their associated acoustic emissions heard. In the early days, the human ear served as a valuable discharge detector and locator of surprising sensitivity. Engineers are familiar with this phenomenon since the sound of corona can often be heard at HV substations and from transmission lines, particularly on wet or foggy days.
The primary method of problem detection through UV discharge visualization used to be visual, assisted only by an optical device such as binoculars. Unfortunately, this had to be done at night to eliminate the overwhelming UV effect of the sun. Even then, only significant discharges on high voltage systems would be visible. Introduction of the first corona cameras began in the early 1990s, initially only for night-time, low light and indoor use. These allowed corona discharges to be seen in the UVa (320 to 400 nm) and UVb (280 to 320 nm) wavebands at its point of inception. Such systems provided a real-time image of the phenomenon. Progress was later made in this technology through more complex lens systems that yield greater light collection and sensitivity. These are still in use where manpower availability dictates against daylight surveys or where the survey is indoors, away from overpowering natural sunlight.
User need soon progressed to requiring cameras that could be used during daylight hours for greater efficiency and safety as well as improved background imaging via the visible channel. Development of the daylight solar-blind camera, although constrained by low levels of UVc corona emissions below 280 nm at which solar energy ceases, allowed inspection during the day as well as relative quantification of the number of UV photon events taking place. The next generation of daylight UV detection cameras offered a removable solar blind filter on the UV channel and near-IR filter on the visible channel. This allowed the same system to be used with greater sensitivity in indoor situations or under low light conditions.
Due to the need to split the image entering the camera to allow the solar-blind filter to block solar radiation, these daylight systems are digital. The UV image is a series of pixels superimposed on the visible daylight image. This has the advantage that it allows ample processing and gives the user options in displaying the UV image to best advantage. These include image coloration to enable the UV to contrast with the background, frame integration to remove background noise, photon counting for UV quantification and high definition zoom to see the discharge location more clearly. Generally, the lower voltage limit for daylight detection of HV faults has been around 11 kV, depending on distance from the object and problem severity. This is reduced to around 3 kV when the solar-blind filter is removed.
Thermal Hotspot Visualization
Infrared detection technology has been used much longer than UV visualization and is an accepted method to detect mechanical faults that could lead to failures on HV systems. A number of sophisticated and expensive systems are available. Unfortunately, IR tells only part of the story since it does not pick up surface discharges where heat is not generated but which relate to voltage differences, thereby indicating areas with potential for flashover and outage. The opposite applies to UV detection, which does not pick up hidden or current related faults from which UV photons cannot escape but which may produce hotspots detectable by the IR imaging system. Both phenomena provide valuable information on the integrity of a power line and its associated hardware. As such, both visualization technologies are required together to see the whole picture. If this can be done using a single instrument, with the addition of a visible spectrum image as well, the inspection process become much more efficient. This also helps to ensure that the picture shows both phenomena under exactly the same climatic conditions and instrument settings thus eliminating many of the variables to be accounted for in comparing survey results from separate instruments, often taken at different times.
Now that the visualization technologies are available there are several ways of deploying those technologies depending on the application. For small area surveys such as in substations then surveys on foot using portable systems are suitable. For transmission and distribution line surveys, mounting of the detection/visualization system into a motor vehicle for cross-country transmission lines, a train for railway lines or a helicopter for long line or inhospitable terrain becomes necessary. Each system has its advantages and drawbacks and careful selection of the appropriate method is required for efficient use to be made of equipment, manpower and time. The helicopter option has further options in that the visualization system(s) may be handheld by an operator, mounted in some form of suspension system within the helicopter or, in the ultimate usage, mounted into a gimbal under the helicopter and controlled from within the cockpit. Of course for all these options it is preferable to have some form of automatic location systems such as a GPS and essential to have a decent recording system so that the faults detected during the survey may be analyzed and reported on later. A recent refinement on the helicopter theme, for lower cost aerial inspections of short runs, is the development of Unmanned Aerial Vehicles (UAVs). Currently these are radio controlled by a ground based operator but it will not be long before autonomous flight systems will fly pre-programmed patterns along sections of line as far as they can go on the fuel load available.
A further usage of the visualization technology is in 24/7 surveying of fixed locations such as particularly vulnerable areas within a substation. This may be done with either a fixed Field of View (FOV) camera or, more likely, one mounted on a pan and tilt mechanism allowing an operator to survey a larger area. This system can then be used with image or pattern recognition software to alert operators to unscheduled changes between successive sweeps indicating the early onset of a potential problem.
Utilizing Visualization Findings
Of course visualizing the problem is still only part of the solution in that one needs to be able to interpret and understand the effect of the results of the visualization process.
The first prerequisite for understanding the results is to have those results in pictorial form to allow easy interpretation. This means that the visualization systems must be capable of recording preferably both still images and video clips, from which further still images can be captured, for use in reports. The low-light cameras generally only output video streams which can be recorded on separate video recorders. The daylight systems have on-board still image storage as well as the video output streams allowing more effective use of the survey time. It is important to understand the following when considering the interpretation of results from an IR or UV survey:
• Not all UV discharges or hotspots are dangerous or damaging to equipment but are an indication of some developing abnormality;
• Any UV discharge or hotspot indicates a loss of transmission/ distribution power, at the least;
• Individual UV and IR detection camera readings are not always conclusive and should be used in comparison with one another within each technology;
• There are many HV equipment, environmental and instrument variables which affect UV photon readings or IR temperature readings;
• Any source/cause of UV discharge or hotspot must be understood before a course of action can be decided upon.
Although daylight corona and Infrared cameras have the function of counting and displaying the number of photons or the temperature of a hotspot respectively, it is important to realise that this is not an absolute measurement of the severity of a discharge or hotspot. Nor can any specific conclusion be taken from a single reading without considering the many variables including distance, atmospheric conditions, camera settings and HV equipment design parameters. The cameras should rather be used as data collection systems, either:
• to compare two equivalent pieces of HV equipment and highlight a significant difference in UV discharge activity or hotspot temperature with the same camera settings;
• to build up a picture over time of the change in behaviour and UV discharge activity or hotspot temperature reading from specific pieces of equipment.
In order to determine what actions should be taken with the results from UV or IR surveys of equipment, it is necessary to:
• Establish a list of the critical equipment to be monitored
• Establish the design criteria/safety factors for equipment and thus the potential failure modes
• Develop a matrix for UV discharge/hotspot causes versus effects with actions for various situations
• Build a set of criteria for visualisation camera settings for taking consistent equipment monitoring readings
• Build a database of UV discharge and/or IR hotspot values on critical equipment under the above criteria, including the physical environment factors.
Continually update the matrix and criteria as more data becomes available (e.g. photon/temperature readings vs physical inspection of equipment)
The final step in utilizing the results of the visualization surveys is to create reports that can be used to highlight the issues to management or to maintenance departments for action to be taken to rectify the problem areas. There are a number of standard software packages available to create such reports and most utilities would have a company format for such reports. If necessary it is extremely simple to create a one-page document, with MS Word for example, summarizing the key aspects and including the images, either taken as still images by the cameras (usually only applicable to daylight cameras) or as still images grabbed from a video recording.
Future of Visualization Technology
The next steps in the evolution of the visualization technology include the application and incorporation of new sensor technologies from research programs in the Space, Military, Medical and Nano-technology fields. These will inevitably lead to more sensitive (earlier detection of smaller faults on lower voltage systems), more compact and lighter systems with more features, more powerful image processing capabilities and more user adaptable features. As previously mentioned the challenge in the aerial inspection field will be the autonomous flight UAV systems as a way of reducing survey costs. Another challenge will be the fulfillment of user demands for instant analysis of the results of a visualization process into a severity report and set of recommendations for actions to be taken.
Of course, over and above the development of the visualization techniques will be the development of new technologies for non-destructive testing of the equipment on which faults are found and visualized.
It can be seen that the technology and instruments available to visualize fault indicators on electrical systems have come a very long way from the naked eye through to the latest all-inclusive multi-channel UV/IR/Visible systems that can be hand-held or mounted in a variety of ways to facilitate efficient surveys of equipment. Such systems have made the life of the maintenance engineer much easier but at the same time have introduced new challenges in the interpretation of the outputs of the systems. There is therefore still much development left in the evolution of the visualization of HV faults through the use of UV and IR imaging.