A game-changing flashover prediction technology for power utilities

Disruptive flashover prediction technology

The MetrySense-4000 (MS4000) Flashover Prediction system is a unique patent-pending solution designed for High Voltage T&D power lines. As of today, this disruptive flashover prediction technology is the only proven and operational solution of its type deployed in live environments. MS4000 reliably and efficiently provides real-time analysis of contaminated insulators and predicts likely flashover-prone insulator strings. Each MS4000 sensor covers a two-kilometer power line segment, which includes all the insulators situated in its range, as opposed to  leakage current detection systems , whereby one sensor can only monitor a single insulator. This is the only available monitoring system featuring a rapid ROI of approximately one and a half years.

Existing solutions are not efficient

The issues relative to the determination of contamination levels of power transmission lines have been studied in numerous research centers and energy systems for over 50 years.

Due to concerns regarding environmental pollution, this issue benefits from increased attention today. Yet, existing solutions are not efficient.

There are three main methods used today for determining where and when to perform line washing operations:

  1. Periodical maintenance: Most of the power utilities are performing routine periodical washing of the contamination-prone lines based on the history of power interruptions in the grid. However, in many cases, contamination builds rapidly, and the line must be washed again soon after a previous washing operation in order to avoid flashovers and power interruptions. For example, agricultural pollution may occur immediately as a result of ambient dust stirred up during field work when humidity is high. The dust sticks to the surface of the insulators and often causes a flashover. Furthermore, in many cases periodic washing operations can be inefficient. From our experience, in some instances over 50% of the routine washing operations could have been avoided without introducing any flashover risk.
  2. Leakage currents measurement sensors: There are some systems that utilize a sensor attached to a designated insulator, which measures leakage current. The main limitation of this method is that the sensor covers only a single insulator string, compared to MetrySense-4000 sensors that monitor all the insulator strings in a 2 kilometer range. This substantial disadvantage is the main reason that leakage current systems are not widely deployed.
  3. Corona cameras: Corona cameras are usually installed on helicopters. The cameras are useful for detecting impairments in insulators or other elements, but are inefficient when detecting the insulator string’s contamination level. The reasons for this are (1) The high cost of making a survey of the lines under inspection using helicopters or from the ground; (2) Dry-band-arcing are visible to the cameras but under certain weather conditions can generate moisture on the surface of the insulator string (especially high relative humidity conditions, and also depending on temperature and wind speed). Therefore, when scrutinizing a high voltage tower with a camera, the team may get no visual evidence of dry-band-arcing if the ambient humidity was lower at the time of the inspection, (3) the visual observation is subjective and inaccurate. It is hard to conclude from a visual inspection of the dry-band-arcing (discharge) what is its actual physical level, and unreliable in terms of methodology, thus rendering decisions based on these reports as approximate at best. By comparison, the MetrySense-4000 provides objective physical measurements of the contamination level with absolute accuracy, (4) even when inspection videos using corona cameras are indeed kept, they do not produce a useful historical database that can be leveraged as a reference for future improvement or for analysis of past events and sources of contamination. For all these reasons, corona cameras cannot provide an effective solution for monitoring contamination levels.

The physical process of contamination, dry-band-arcing and flashovers

Main contamination sources of power line insulator strings are

  • Coastal pollution on overhead lines located near the sea (mainly salt contamination)
  • Industrial pollution from industrial zones.
  • Agricultural pollution may occur immediately as a result of ambient dust stirred up during field work when humidity is high. The dust sticks to the surface of the insulators and this often a cause for a flashover.
  • Pollution from anti-freeze substances, which is applied on highways, rise up due to car traffic in under some and whether conditions and sticks to the insulator strings
  • Desert pollution
  • Dust storms and other pollution

The contamination of insulators is not uniform: it also depends on the shape of the insulators. When insulators become wet due to fog, rain or dew, the salts in the contaminator are dissolved, and the conductivity of the insulator surface rises sharply. The magnitude of the conductivity depends on the level of contamination, as well as on the wetting of the insulator. A leakage current that passes along the surface of the wet and contaminated insulator can reach hundreds of mA in severe cases.

Since the insulator’s surface area is not even, neither is the current density along the string uniform. As a result, different areas of the insulator heat up unevenly. At locations where the current density is maximal, distinct patches around the circumference of the insulator dry up, and therefore the resistance of these patches increases considerably. The voltage drop on such a patch increases, up to the value at which a breakdown of the air gap occurs. Thus, an electrical arc (dry-band-arcing or discharge) is formed, shunting the dried insulator patch. Then the current passes along the surface of the insulators, drying the water drops, and the arc is extinguished. The dried surface becomes humid again and the process repeats itself. The resulting arc may develop on any area in the string, and it may cover one or several insulators, depending on the contamination level, the structure of the insulator string, and the amount of wetting (moisture/humidity).

The following videos were taken in IEC’s lab as a part of the MerySense-4000 R&D process, which included extensive tests on many types of insulators. In the following videos, high voltage was applied on contaminated insulators, and the voltage was gradually increased. First, dry-band-arcing appears on the surface of the contaminated insulator. Then, when the voltage reaches a critical level, a flashover occurs followed by a trip of the protection system, which in turn causes a power interruption.

Dry-band-arcing on a contaminated porcelain

Dry-band-arcing (partial discharge) and a flashover on a contaminated porcelain insulator string

Dry-band-arcing (partial discharge) and a flashover on a glass insulator string with salt contamination


Dry-band-acing picture that was taken using a stills camera with long exposure

Principle of operation of MetrySense-4000

The patent pending MetrySense-4000 system can detect electromagnetic signals, which are generated by dry-band-arcing over contaminated insulator strings. The range between the contaminated insulator string and the sensor should be in a range of  up to 1 km from the sensor.

A MetrySense-4000 sensor is mounted on the shield wire (ground wire) near the top of the tower, and is connected to a MetrySense-4000 data acquisition unit which is mounted on the tower. The signal received by the sensor is processed by the DAQ in frequency domain and in time domain, and the processed information  is kept in a local data base in the DAQ. This information is then sent via the IPv6 radio network that is automatically created by the DAQs and reaches the MetryView server for final analysis and presentation.

The MS4000 in action on a 400kV line

An example of operation of the MetrySense-4000 system in a 400kV line

The following example illustrates the operation of the MetrySense-4000 system before and after a washing operation. The system was installed on a 400kV line in Israel on 30.8.2012.


 (a) The sensor which is installed on the shield wire


 (b) The data acquisition unit (DAQ) which is installed on the tower and connected to the sensor by a coaxial cable

 On 19.9.2012, about three weeks after the installation of the system, the transmission line was washed by a helicopter, and the washing operation was captured by the following video:

The following graph shows a MetryView graph of dry-band-arcing which existed before and after the washing operation. This graph is a result of processing of hundreds of sensor samples.


Dry and arcing levels recorded by the system before and after the washing operation

It can be seen from the above graph substantial changes in the readings before and after the washing operation. The changes in discharge levels ofduring the days preceeding washing are attributed to fluctuations in relative humidity. The system is measuring hundreds of samples per day every day, and looks for the worst-case dry-band-arcing levels in the last two weeks in order to determine the status of the line. The discharges before the washing operation were at the “Yellow” level, which means that washing operation was not urgent. The graph also shows that the dry-band arcing level dropped to zero after the washing operation.

This is the presentation of the sensor of our example before and after the washing operation:


Sensor Status on the MetryView map – (a) 19/9/12 –  before washing, (b)  26/9/12 – after washing

In this example, the status was “yellow”, and there was no urgency to wash the line. In fact, rains in this region of Israel starts at November, and this washing operation could have been saved in this case.