Alan,

You won't find any real evidence because there isn't any. Your rational common sense approach to appraising the situation has led you directly to the root of the problems - cheap tat made to a minimum standard.

AFDDs won't pick up anything on the south side of a transformer and it's associated secondary electronics, nor will they pick up a serial arc on a ring final.

Someone is trying to make a lot of money very quickly.

As you say hard information is sparse. My guess is that the case of the kind that  "fused the cable to the terminal" is about all they'd be able to detect that other devices (e.g. RCDs) wouldn't have anyway.  If money were no object I'd probably have them on the basis that in the game of natural selection every little helps, but in the real world I'd probably think that spending that money on about 101 other things first would probably yield better results. (Personally I'd go for BS 8436 cables, either metal or thermosetting plastic enclosures for accessories, and better terminals - as being a more thorough approach to a wider range of problems.) 

The rather radical change of emphasis in their use BS 7671 between AMD1 (generally in fire vulnerable environments (sleeping accommodation, combustible materials) and AMD2 (just socket outlet circuits in high-rise and high density housing) it does seem to suggest that their emphasis is now more on portable appliances rather than fixed wiring. Not many portable appliances draw more than 2.5A (so unlikely to help with the knock-off hovverboard or electronic cigarette chargers that seem prone to starting fires), the likes of irons and vacuum cleaners are tend to be attended when in use, so likely to be unplugged if they start misbehaving, long before they'd get to the stage of starting a fire. I guess portable room heaters are the most likely candidate, but how often they suffer arc faults rather than just just falling over and overheating the fluff, I'd have my doubts.

I suppose my slight hope in all this is that if we start introducing devices that can very carefully monitor the details of load currents and use some decent intelligence in evaluating them, there might be the possibility of detecting other problems. E.g. the increase in resistance over time of a loose connection or overloaded conductor is likely to be quite different to that of a normal load (comparable with the arcing from motor bushes being distinguishable from arcing from a loose wire) so perhaps the basics of an AFDD could in time be adapted to something that is more generally useful.

   - Andy.

When they were first introduced in the wiring regs I did some research.

All the statistics I could find in America assumed that all electrical fires were caused by Arc faults and I could not see a correlation between the introduction of AFDDs and a decrease in electrical fires.  This exercise was made more difficult by the individual states using different terminology and collecting different statistics. Bear in mind that America used to use a lot of aluminium cable in house wiring  and the outlets are not fused. 

Digging deeper I found papers by an safety conscious American electrical engineer, who's name I cannot remember, who started out writing papers and having meetings with electrical distributors where he was trying to get them to install fuses at the house end of distributers cables. He was also trying to get the MCBs used in consumer units to be made more sensitive, from what I read the MCBs used at the time were similar to our D types. He then started on Arc fault detection having been involved in the examination of a number of electrical fires and I believe had a number of patents on the technology. His arguments mentioned appliance flexes a lot. 

So I am skeptical about the benefit of AFDDs in installations complying with BS7671 without AFDDs and I feel that we should have been given the statistics that prove the benefits of AFDDs before they were introduced into the standard

I managed to find the name:

It is actually more complex than the comments here. My personal view (having played around with some) is that they probably show zero benefit on fixed wiring because fixed wiring does not initiate arcs, even with loose connections. Loose connections either make a connection or not, and as 230V can only jump about 100 microns at best, and even if it does a big hot arc cannot start unless it is already a big enough arc to melt the wire or terminal (the electrodes) to produce an ion source, otherwise nothing much will happen. This is clearly illustrated by the product standard and demonstration jigs the manufacturers show around to get sales. The test rig and standard both depend on one electrode being carbon (which ionises very easily) and moving the electrode(s) to actually start the arc with an essentially short circuit. As we do not have carbon wires or terminals, and the manufacturers do the series test with a wire cut but the sheath burnt (carbonised) to start the arc (see product standard). Various makes detect arcs started in various ways but do not detect sparking of motor brushes, and vary a great deal in sensitivity even with the jig. There are lots of YouTube videos where they are tested by the trade. None of these manage to get a trip with a metal-metal arc however hard they try. Note that arc welders depend on the easily ionised electrode coating to get a reasonable arc, it is virtually impossible with uncoated wire in air. Burnt, welded etc terminals are caused by resistive heating not arcs, so almost never cause a fire unless there is much flammable material nearby. Thermosetting boxes and accessories are very hard to get burning (that's why that are made of thermosetting material).

Most of us who have been in the industry for 30 years or more will remember the long discussions on RCDs. In fact, in the 1st Edition of IET Guidance Note 3 Inspection and testing is the statement (2nd para, Section 15, page 81):

In order to test the effectiveness of such devices after installation a series of tests must be applied to check that they are within specification and operate satisfactorily. This test sequence will be in addition that proving that the test button, if fitted to the r.c.d., [sic] is operational.

I remember 'back in the day' there was so much concern about mandating RCDs, chiefly because of their cost (around £80 to £100 in the late 1980's), and the fact no-one trusted them to operate 'as sold'.

The reason we have continued to test RCDs (but not other protective devices) has been ongoing debate about their effectiveness, stemming back to this time. (I'm not, however, saying that all RCDs were always perfect, we all know they have occasionally failed to operate, but no-one seems to care about circuit-breakers which also may suffer from environmental influences on their mechanisms etc.)

I'm sure there will be those who will try and deviate from the following points, by trying to limit the parallels in the discussion on AFDDs. But I think these points are very important:

1. We should not miss the opportunity to introduce protective devices that can save lives and prevent damage to properties. I fully understand cost-benefit debate, but looking back I've heard this before with RCDs.
2. There is an agreed international (and British) standard for these products. That standard has gone through full rigour and public comment process.
3. Because of 1. and 2. above, pseudo-scientific approaches, inductive thinking, etc. used to answer the questions posed in the OP isn't really going to work.

Based on all of the above, I believe the question posed in this thread ought to be turned on its head, and ask the three questions posed on the OP in the following way:

• Is there any evidence that AFDD's would NOT stop these failures?
• What about internal appliance faults - would AFDDs NOT detect them?
• Wasn't Grenfell started in a fridge? If so, would AFDD NOT detect that fault?

Properly researched and supported answers on a post-card please ...

Graham, the parallels you draw with the introduction of RCDs is a red herring, as are your questions which ask the impossible, i.e. to prove a negative.

I have some questions of my own.

Where is the empirical evidence/data gathered by professional investigators which highlight that the lack of a AFDD was responsible for the outbreak of a fire? Or rather, how many fires over the last say, 10 years would have been directly prevented by a AFDD?

How many fires which have been fully investigated, can be directly attributed to an arc fault as being the root cause?

Where is the evidence?

Also, where is the evidence which proves that the decision to mandate the introduction of AFDDs was exclusively driven by sound scientific studies rather than hard lobbying from manufacturers?

Why do the test rigs purposely constructed by the manufacturers rely upon a pre-carbonised element in order to make their device look as if it is actually working properly? Something which is a rare event in the real world of electrical faults.

If the scientific studies were really carried out, then they should be available for public presentation and scrutiny.

Every time I have asked these questions before, I have been deafened by the silence which ensued. Care to make some noise?

Otherwise, all I get are aburpt denials - a bit like Boris denying he had a party at No 10, and we all know how that ended.

There is an agreed international (and British) standard for these products. That standard has gone through full rigour and public comment process.

But aren't product standards typically in effect sponsored by the manufacturers involved, primarily in order to refine the market into which they wish to sell? As far as I know there's no independent refereeing of their technical contents, as there would be for instance with a scientific paper published through a learned journal for example. Sure any tom, dick or andy can send in comments, but the committee has the final say, and often only ones that can really afford the time an expense of being on such committees are those whose employers will grant them the time (and expenses) to do so, and such employers thus tend to be self selecting to those than can see an advantage in doing so (e.g. manufacturers or their associated trade bodies). Rigorous perhaps, but to what intended end?

   - Andy.

Perhaps we ought to start by asking how fires with electricity actually start.

I have provided comment in the past on this Forum, with testimony as a first-hand witness of at least parallel arc that could easily lead to fire, which is swiftly met with "well, we don't wire houses with that cable".

Lots of questions one way, but none the other ...

Which is why I think the parallels with RCDs are very valid.

whjohnson said:

How many fires which have been fully investigated, can be directly attributed to an arc fault as being the root cause?

Where is the evidence?

So,  again, how does electricity cause fires? Above is a post from another contributor, which says a lot about how electrical products and wiring can't cause fires or help them spread, BUT I don't think any of us would doubt that fires in electrical products and installations happen.

whjohnson said:

a red herring, as are your questions which ask the impossible, i.e. to prove a negative.

I'm not asking anyone to prove a negative, I'm asking for some balance in the discussion.

From my perspective, this debate is too much "What have the Romans ever done for us?" at the moment.

AJJewsbury said:

But aren't product standards typically in effect sponsored by the manufacturers involved, primarily in order to refine the market into which they wish to sell?

Interesting point of view - but I know there are contributors out there who would be offended at the suggestion that they are driven by those motives.

It's easy to sit on the side-lines and say such things - equally someone might suggest that those representing test houses and similar verification bodies are only there to help veer the verification elements of standards in a more lucrative direction.

Where would that stop?

AJJewsbury said:

Rigorous perhaps, but to what intended end?

Perhaps, to try and improve things?