Ninety years for a microsecond
Testing and inspection of equipment has certainly been traditionally the main focus of KEMA Laboratories; it is even expressed by the first letter of the acronym KEMA, the Dutch word “Keuring”. Though most customer’s aim is to get equipment certified in order to increase market acceptance, through the decades a large number of research oriented tests for product development were conducted. In many cases, customers do not have the laboratory facilities available they need for product development, so they called upon KEMA Laboratories. Research and development means not only the result of a test counts, but also that having ‘’looking glasses’’ to observe processes inside equipment during a test becomes relevant.
The first tools that became available were measuring equipment to observe ever faster and smaller signals. Special apparatus that can record lightning impulses of a millionth of a second, very short lasting partial discharges in cable were already quite common in the seventies. When these data became available as “digitized” information, the next step could be made because digitalization makes it possible to “do anything you like”, e.g. extract special features hidden in the signals.
The first practical use is to automate measurement procedures, whereby smart algorithms process signals and extract the key features that need to be reported. The AURA system was developed in the eighties, and made the reporting process more efficient by removing manual evaluation. Great effort was put into setting up automated laboratory data storage and analysis systems, and the roll out of this “MARS” system in a uniform way to all the laboratories. By now, the amount of data is gigantic: an estimated 50 TB of data is collected, stored and processed in a typical single testing day, only in the Arnhem laboratories.
Next the question was raised on what information can be retrieved on the critical processes during a test. Circuit breakers need to interrupt current, and their behavior around the current zero crossing is of crucial importance. In the nineties a project was set up, supported by EU, to develop a method to characterize the breaker’s performance in the few microseconds around “current zero”. This technology was refined and transferred to a number of manufacturers for application in their product development. Later, the development started to combine this information with the monitoring of transient pressure inside the arcing chamber of SF6 circuit breakers. Such analysis tools will be helpful to characterize the specific features of SF6 alternatives, a buzz word these days.
Also, a method was developed to measure the very small current that is flowing after interruption of vacuum circuit breakers. High-power laboratory test engineers frowned their eyebrows when confronted with micro-amperes instead of the usual kilo-amperes.
The availability of all these data is partly needed to validate models that manufacturers use in the design of their equipment.
This triggers the discussion “why not replace physical testing completely by model verification?” With the advent of cheap and massive computer power and advanced multi-physics simulation software it is generally recognized that stresses due to voltage, heat, mechanical tension, thermal, pressure etc. stresses can be predicted quite accurately, even in complicated designs. However, the yield to such stresses, causing disruptive, local phenomena like electrical breakdown, melting, rupture, explosion, break, deformation etc. cannot be predicted by simulation.