From: Paul
Date: Fri, 25 Oct 2002 13:25:26 +0100
Subject: Re: [TSSP] Top Voltage
Hi All, Things do seem to be looking up. I'm pleased and relieved to see that there's a huge amount of professional research out there on the subject of breakout in long gaps. A few comments on what I've seen so far, (thanks Marco for Larsson, et al, that was bang on target) There appears to be off-the-shelf physical quantitative models available for each stage of the breakout process: corona onset, diffuse corona (streamers), streamer-leader transition, leader development. These models extend to discharges inhibited by finite source charge and resistance. They seem to involve straightforward integrals over field quantities, with things like coefficients for ionisation, recombination, etc. The characteristic time for leader formation seems to be quite long - several tens to a hundred or more uS, which gives plenty of scope for delayed breakout in say, OLTC. I begin to see now why breakrate, frequency, and coupling will have a big impact on the breakout. I'm coming to agree with Boris in that although 30kV/cm may predict the corona onset, the initial effects are pretty negligible, and a whole lot more physics happens (and a much higher voltage is required) before any distinctly visible leaders begin to propagate. We should, I think, take some time to study the available literature and set up some models of our own, before we propose too many more experiments. There's an awful lot of stuff out there and some effort will be required to filter out the relevant bits without drowning in information and getting hopelessly lost. Hopefully, we can achieve our aims without having to do any original research of our own. We can continue our tradition of working from well established physical principles - no need to launch off into the dark. To summarise our goals as a guide to what info is relevant: Our good understanding of the secondary needs to be extended to account for the interaction with breakout loading. We would like to be able to predict the breakout performance of a TC, given the geometry and firing voltage.[*] Predicting the performance, in this sense, means being able to estimate the typical number and length of inhibited leaders when discharging into air, for a given firing voltage. To do so, we will likely have to take account of the dynamically varying load applied to the secondary, working in the time domain to derive a numerical solution to the non-linear system. Only then will we cope with arbitrary topload waveforms (wide range of Ctop, Fres, and k). Thus we will need to assemble quantitative models for each step in the formation of inhibited leaders. We should be able to predict, not only the corona onset, but the voltage at which diffuse corona forms, and the voltage at which visible leaders begin to start, as well as work out the max leader length. One possible difficulty is that all the stuff I've seen so far deals with single polarity impulses. Our HF alternating waveforms, modulated by a beat envelope, then modulated again by a break rate, will add some complications. We'll just have to see how things come out. I'll start a new web page this weekend to summarise existing applicable research on breakout phenomena. I'll use it to keep track of which references we need to obtain or have got, so that Marco will know exactly what to go for. There's such a huge amount of material out there that we have to be pretty specific. [*] Given this, we would then be able to close the loop on secondary design and answer questions such as which combination of BPS, Fres, k, h/d and topload size makes the longest leaders from a given input power. -- Paul Nicholson, --
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.