CoroCAM FAQ s Answered

Transcription

1 CoroCAM FAQ s Answered by

2 The most FAQ s by customers are: Why do I need a corona camera? From how far away can a discharge be detected? In other words, what is the discharge detection distance? Once I ve found a discharge how do I define it s severity/maintenance priority? What about corona counting? How often should inspections be performed? Why should I buy a CoroCAM? Under Construction Why the different fields of view?

3 Why do I need a corona camera? by

4 Why do I need a corona camera? External discharge activity is indicative of: Damaged components Contaminated surfaces Bad design Discharge activity causes: Power loss Damage to surrounding materials Annoyance through audible noise and/or radio interference

5 Why do I need a corona camera? Discharge activity has signatures which can be detected by the methods below: Signature: Audible & Ultrasonic noise Radio interference Heat Visible light UV light Detection Equipment: Acoustic detector Antenna Thermal camera Normal SLR Long exposure camera UV Camera Constraints: Not spatially accurate, many noise sources, no visual record. Not spatially accurate, no visual record Too little heat to detect from corona, sparking & micro-arcing. Large arcs are possible to see, but rare. Only after sufficient damage has occurred can the effect of a discharge be seen. Time intensive, night only, line of sight only Line of sight only

6 Why do I need a corona camera? Audible & Ultrasonic noise, Radio interference and Visible light detection equipment may be the cheapest to purchase, but they don t deliver quality information. Thermal cameras can only detect the heat from large arcs. Micro arcing can create radio interference, sound and UVc, but they are too small to heat metal parts, causing no heat signature. Only UV light is: spatially accurate enough to show where the fault location is. common & intense enough from both types of external electrical discharge to be detectable

7 Why do I need a corona camera? A corona camera is the only way to precisely visualize the presence and location of external electrical discharges (corona, sparking and arcing).

8 From how far away can a discharge be detected? by

9 From how far away can a discharge be detected? In other words, what is the discharge detection distance? Anyone who gives an answer to this question, is not telling the whole truth. It s a lot more complicated than just saying 200m. The question above has 3 parts: discharge an electrical discharge results in a quanta of light generated detection the light falling on the detector causes a response. distance indicating a distance is involved.

10 Discharge Detection Distance The discharge generates a quanta of light which is related to the number of electrons moving in the air. The intensity of the discharge is not related to the fault severity and is influenced by a number of factors, such as: The gradient of the electric field The ambient air density (air The geometry of the discharge point temperature and pressure) The observers point of view The ambient humidity The load applied The wind speed The supply frequency Ambient airborne particles The temperature of the discharge Ambient air ionization activity point Other discharges Cosmic ray activity The material of the discharge point

11 Discharge Detection Distance Both Corona and Arcing have signatures in the UV band. UVa & UVb is present in sunlight, while UVc is blocked by the Ozone layer.

12 Discharge Detection Distance Intensity Only in the UVc band is the corona signature brighter than sunlight, this is called the solar blind band.

13 Discharge Detection Distance The camera features a lens, a visible camera, a solar blind filter, an image intensifier and a B/W camera (together called a detector) Inbound light Video camera Lens UV Solar Blind Filter UV A + UV B blocked UV Image Intensifier Detector The lens collects inbound light, the UV light is split off and passed through a filter onto the detector. UV C B/W Camera

14 Discharge Detection Distance The solar blind filter blocks all the light from the sun. Light frequencies blocked by Solar Blind filter Intensity The remaining light

15 Discharge Detection Distance Since there is very little light left, an image intensifier is used to amplify the light to levels where the camera can sense it. Camera pixel array

16 Discharge Detection Distance The detected UV image is overlaid onto a visible image to show where the discharge is occurring. UV Image Detected UV Image After Binarize Visible Image Recorded Because the discharge event A white blob can is sporadic a sequence (video) be missed against of images are required to a cloudy sky. effectively show the discharge Therefore color is activity for diagnosis. applied.

17 Discharge Detection Distance

18 Discharge Detection Distance It can be understood that generating a stable discharge to calibrate cameras against is almost impossible, given all the factors that affects it.

19 Discharge Detection Distance It is known that a blackbody emits UV and other light irradiation in known volumes when the air inside is heated to above a certain temperature. The blackbody works consistently irrespective of many of the factors which influence the UV generation by a discharge. Blackbodies are routinely calibrated and can also be finely controlled. The irradiance in the UVc band can easily be calculated, it s a fundamental part of physics. Blackbody and temperature controller.

20 Discharge Detection Distance At 940 C the air in the Blackbody starts to emit UV radiation in sufficient quantities to be detectable. This correlates to 2.05x10-18 W/cm 2, in the solar blind band in which the camera is sensitive, falling on the detector and causing a positive response.

21 Discharge Detection Distance As the distance between the camera and the discharge is increased the amount of light reaching the camera from the discharge is decreased. Physics describes this as the inverse square law of light.

22 Discharge Detection Distance The CoroCAM has minimum detection limit of 2.05x10-18 w/cm 2 therefore: For a discharge to be visible at 16m it will have to create a light segment with an intensity of 1.37x10-16 w/cm 2 100x brighter than the minimum detected level. For a discharge to be visible at 50m it will have to create a light segment with an intensity of 1.26x10-15 w/cm 2 620x brighter than the minimum detected level Therefore a large discharge can be seen from further away than a small discharge.

23 Conclusion Detection of a small discharge at a long distance is possible if the camera is sensitive enough. Sensitivity is a variable for which there is no standard measurement. CoroCAM uses the black body measurement to determine sensitivity, this is a physics based measurement using a conventional setup which is immune to influence from most of the factors which would affect discharges. This test has been done at a number of labs, achieving similar results. Sadly this technique is foreign to most electrical engineers.

24 Conclusion One of our competitors expresses their product sensitivity in pc discharged from or RIV generated by a discharge which can be seen by the camera. Electrical engineers like this as it is in values they can understand. However, there is a significant problem with this as the exact setup is not defined by a standard, the way to measure the pc or RIV is defined by a standard, but not the discharge point setup. The organization who did the test for this competitor is also the only organization who can achieve the test results. With their assistance we ve tried to replicate the test at other labs, but have failed. The first requirement for a standard method is that it must be reproducible at any lab, which their sensitivity test is not.

25 Maintenance Priority by

26 Maintenance Priority The discharge is just a location indicator. The size of the discharge does not dictate it s Maintenance Priority. The inspector needs know if a discharge at that point can cause damage which would compromise the system and has to visually inspect the discharge point for damage. As the conditions at the discharge point changes the discharge could extinct and not be detectable at the next inspection.

27 Maintenance Priority The US EPRI teaches a protocol to assign maintenance priority. The protocol is based on outage avoidance. Discharges are assesses based on the likelihood of them causing an outage in the short term or damage to hardware which will result in an outage in the long term.

28 Maintenance Priority I.e. A discharge from an abandoned conductor may be massively intense, but it would not cause an outage. Conversely a discharge from a polymer insulator may be very small, but it indicates a physical fault which can quickly destroy the insulator, resulting in an outage. The EPRI protocol has proven the most useful.

29 Maintenance Priority No Can the discharge cause an RI/AN complaint? Priority A No action required Address RI/AN issue Discharge located Can the discharge cause damage to the component it is on or nearby components? Yes Is damage visible? No Priority B Monitor Yes Priority C Repair at next outage Will the damage cause an outage before the next maintenance shutdown? Yes Priority D Schedule outage 29

30 Maintenance Priority The EPRI maintenance priority relies heavily on the analyst being able to see damage and asses it s severity based on the visual image. None of the current corona cameras, available today, can record imagery at the required resolution. Therefore the inspector needs to carry a long zoom digital camera with which to take stills showing the visual condition. The following case study shows the application of a CoroCAM 6D camera to see the discharge location and a Canon SX50 Compact DSL with an ultra zoom lens to see the physical damage at the location.

31 A typical case 3 types of faults in one image. 33kV power line outside Port Elizabeth. Recently installed polymer insulators GPS ,

32 A typical case SLR photo Priority A No action required Should have used correctly sized bolts

33 A typical case Priority A No action required Should have folded split pins correctly

34 A typical case Priority D Immediate action required. Pin hole in polymer sheath, failure in <2yrs

35 What about corona counting? by

36 Corona Counting Initially corona camera users realised that the intensity of the discharge was related to the power lost. Power loss was seen as very important. Therefore they devised ways of quantifying the intensity. This is the corona count. Later it was shown that power losses observed on faults in the wild were in the nanowatt region and when the cost of each discharges was calculated it came to an insignificant value. Power loss due to bad design is important as the discharge will happen at each instance the design was used and the accumulated cost could be significant.

37 Corona Counting With all the counting going on users forgot to consider the aspects which affect the count and that it was not meant to show fault severity or maintenance priority.

38 Factors Affecting the Corona Count The counting method used: Blob counting Intensity Summing The focus The distance between the camera and the discharge. Camera hardware effects Camera settings

39 Corona Blob Counting Some cameras use blob counting to count the spots created by photon impacts. These spots can cover multiple pixels (average 6x6 pixels). Defocused UVc Image

40 Corona Blob Counting The first step in blob counting is converting the image into a binary image (B/W). Then the blobs are grouped based on connectedness.

41 Corona Blob Counting Blobs areas of less than 4 pixels were ignored, resulting in 11 blobs counted. Touching blobs were counted as 1 in these 3 instances

42 Corona Blob Counting High intensity blob indicating multiple photon impacts was counted as 1 Clearly this is not going to give accurate readings.

43 Corona Blob Counting Counting touching blobs as 1 results in a Counting vs Signal strength bell curve graph as shown below: MAX Count At high intensity a count of 1 results!! How can this be correct? 0 Intensity Focusing the image will cause more blobs to merge to a point and be counted as one. This method is not radiometrically correct. MAX

44 Corona Intensity Summation A better method is to sum the intensity of the pixels in the image.

45 Corona Intensity Summation The image pixel intensities are summed to give a value of When focused all these blobs are merged at the centre and their intensity can still be summed.

46 Corona Intensity Summation The summed intensity is then convolved using an algorithm which takes into account the camera Gain, quantum efficiency, etc. The result is that measurements now grow linearly until pixel saturation.

47 Distance Distance from the source has the primary impact on the apparent count. Doubling the distance will ¼ the intensity of light reaching the detector and result in a lower apparent intensity & vice versa.

48 Distance The insulator in the left image is half as far away as the insulator in the right hand image. Therefore the right hand count must be multiplied by 4 = Note that the gain on the right has been reduced, resulting in a lower count

49 Camera hardware effects

50 The effect of the Gain setting The camera detectors gain has an influence as too low a gain will result in low levels of light reaching it not being amplified enough to result in a detectable signal and too much gain will result in noise and saturation. Blob counting Decreasing the gain can cause blobs to extinct due to too little signal amplification to reach above the threshold. Increasing the gain will cause them to merge together, due to blooming. Therefore there s only a small useful range of gain.

51 Conclusion The count of discharges can only be compared if they are recorded under the same conditions by the same camera. Therefore the count value is of no use when it comes to severity assessment of the discharge.

52 How often should inspections be performed? by

53 Inspection Frequency Installation qualification (IQ) inspections should be done when the construction of a new line or substation has been completed before transfer to the operations group. This is to prove that the hardware is corona free. Thereafter a 1 year in operation inspection should be performed to ensure no latent faults have made their appearance. After the 1 year in operation inspection the inspection frequency depends on a number of factors.

54 Inspection Frequency This varies from situation to situation depending on: HV hardware used (Glass/Porcelain insulators vs Polymer) The weather conditions (Dry vs Rainy) The age of the infrastructure (New vs Old) Manpower and equipment available (Inspectors vs Analysts) Accessibility of the location Frequency of outages

55 Why should I buy a CoroCAM? by

56 Why should I buy a CoroCAM? CoroCAM (designed and delivered by the CSIR in 1991/1992) was the first camera designed specifically to visualise corona discharges on overhead power lines. In 2008 the CoroCAM team was spun out to form UViRCO. We have a lot of experience. This experience has allowed us to design the cameras to be user centric. To this end we are usually first to implement features which are advantageous to the user.

57 Why should I buy a CoroCAM? One handed operation UV colour selection UV transparency UV overlay and alignment on all optical zoom levels Motorised solar blind filter Combined UV, Visible and Thermal camera Full resolution recording Adaptive integration and noise reduction Icon based interface Features first introduced by UViRCO ISN (Image Series Numbering) Calibrated corona count

58 Why should I buy a CoroCAM? We design our cameras to be rugged. Of the 800 CoroCAM s built less than 5% have been returned for in or out of warranty repair. We are ISO rated

59 Why the different fields of views? by

60 Field Of View The field of view is the solid angle through which a detector is sensitive to light. FOV = Field Of View DFOV = Diagonal FOV HFOV = Horizontal FOV VFOV = Vertical FOV.

61 Zoom When a camera is zoomed in the FOV reduces. The black frame is the initial FOV The green frame shows the zoomed in FOV.

62 Resolution Resolution makes it possible for the operator to see the small details which indicates the cause of the discharge. Resolution is expressed as the number of pixels per degree of the FOV. Ofil Superb and Luminar recorded video at 320 x 240 pixels. LuminarHD records at 1280 x 720. CoroCAM 6D and later recorded video at 736 x 576 pixels.

63 CoroCAM FOV The CoroCAM design gives the UV camera a fixed 8 HFOV. The visible camera can zoom out to 16 HFOV, at this zoom level (0.5x) the UV only covers the central 8 portion. Top View 8 16

64 Ofil FOV The Ofil Superb (with wide FOV accessory lens) and Luminar duel FOV designs allows Ofil cameras to have FOV s equal to 5 or 10. The UV overlay disappears when any optical zoom is applied. Top View 5 10

65 Comparing FOV & Resolution Comparing FOV s only, only gives part of the answer the recorded video resolution must also be taken into consideration. Just the horizontal resolution is taken into consideration. CoroCAM Ofil: 320 (H)/ 5 = 64 pixels per degree. 320 (H)/ 10 = 32 pixels per degree. LuminarHD 1080(H)/ 5 = 216 pixels per degree CoroCAM: 768 (H)/ 8 = 96 pixels per degree. Ofil Top View

66 Comparing FOV & Resolution CoroCAM can optically zoom the visible channel and maintain the UV channel due to digital enlargement of the UV channel. The CoroCAM 7 & 8 visible channel can optically zoom to 2.1 (light blue FOV) CoroCAM 8 & 2.1 FOV (blue shades) CoroCAM: 768 (H)/ 2.1 = 365 pixels per degree LuminarHD: 1080(H)/ 5 = 216 pixels per degree Ofil 10 & 5 FOV (orange shades) Top View

67 Why the different FOV s? The CoroCAM Solar blind optics design started in Working with ESKOM (the SA national power utility) it was determined that the camera would be used in substation and for overhead line inspections, each requiring different FOV s. An optimal midway of 8 was selected as the best, given the cameras zoom with UV overlay and recordable resolution. Ofil s optical design requirements were set by US EPRI. EPRI s key view was that overhead line inspection was a more important part of inspections and resolvable detail, to base the assessment on, was the key requirement. The lower video recording resolution of the camera electronics required a smaller starting FOV to get suitable visual resolution, which is why Ofil settled on 5. After market pressure due to UViRCO sales efforts Ofil was forced to get a wider FOV, this was achieved with a 0.5x zoom attachment or drop in lens resulting in a 10 FOV.

68 Conclusion CoroCAM gives far more visible resolution with UV overlay than either Ofil Superb or Luminar at any FOV. When Ofil Luminar and Superb are using their wide FOV optic they can scan a slightly larger area, but at a 1/3 rd of the resolution of the CoroCAM. Ofil Wide 10 FOV - Area of View At 10m the Ofil area of view is 35cm wider. At 5m the Ofil area of view is 17.4cm wider. The larger Ofil area of view is insignificant when considering the distances between substation hardware. CoroCAM 8 FOV- Area of View

69 Thank you