Sunday 12 March 2017

Reflections on a Piece of Bullshit PR

The National Grid, upon whom we all depend, glories in a comfortable statutory monopoly with its more-or-less guaranted rate of return and a very cosy relationship with government and regulators alike.  Still, it feels the need to boast.


OK, they posed the question and, given the monopoly thing, it's a very important question.  Intuitively the answer of course is - well, your being a monopoly and all, we confidently expect you are inefficient - it's just a question of scale.  As it happens, a lot of the old CEGB inefficiency (which was grotesque, BTW) was kicked out of the system in the decade or so following privatisation.  But it's been creeping back, courtesy of the renewables upsurge.  The aforesaid authorities rely totally upon the Grid to facilitate their manic policies, and don't feel inclined to question them particularly closely on how they go about it.** ++

Anyhow: having posed the fateful question, the Grid declines to answer it, relying on the usual PR sleight of hand.  Are we efficient?  Well, it's all very difficult as we hope you understand.

And - boy, oh boy - are we effective!!  99.9999% reliable!!  Six nines!!  Go Grid !!

This raises some interesting issues.
  1. Effectiveness has minimal bearing on efficiency, even if the first 3 letters are the same.  I assume, however, that loads of suckers are focus-group proven not to notice the difference.
  2. The whole matter is one of Gold Plating - the old CEGB disease.  We know you can be staggeringly effective if you throw money at it.  Our money.
  3. WTF does six nines mean in this context anyway?  One hour's unreliability in 120 years?  The Grid hasn't been around that long.  One second in 12 days?  (Incidentally, 'grid reliability' isn't what most people would associate with continuity in supply of their electricity.  Blackouts happen all the time, of course: but almost all of them occur at the distribution level - storm damage, pneumatic drill through cable etc - for which National Grid is not responsible.  'Grid reliability' is more to do with 'quality of electricity'.)
  4. Most weightily:  what is the optimum degree of reliability?  Thoughtful people realise that the cost of going from, say, five nines to six nines is many times greater than going from four to five.  Who needs six?  There may be an answer to that question: it will probably be the manufacturers of ultra-high-tech electrical components.  But they typically modulate their power supplies themselves, so as to be even more sure.
Yup, this is a big question, and one the politicians leave to the Grid to decide for themselves.  But there are potentially massive efficiencies to be had from a true optimisation of reliability.  Suppose hypothetically it was noticed that no-one drank tap water, but bought their own in plastic bottles.  What, then, would be the point in treating all mains water to be potable, as at present?  It's obvious that for some overwhemingly vast percentage of all usage of mains-water, its costly potable quality goes to waste.  It's the same with grid reliability: who really needs six nines?  Maybe they should get what they need in a different way.  Maybe the Grid should even pay for the cost of their doing so ...

Not something we can solve on a Sunday morning.  One more observation on the Grid, though.  Just a very few years ago they told us we needed a capacity margin (i.e. excess of power generating capacity over peak demand) of 20%.  Funnily enough, now that it's negative, they tell us that we only need 6% (and they procure the top-up in various costly ways). 

Yeah, funny that.  Efficient?  Effective?  Eff ....  Lay off the bullshit PR, eh, boys?

ND
______________
** although one-time Energy Secretary Amber Rudd - remember her? - announced this was going to change ...  We can but hope.

++ UPDATE: now here's a thing, from today's DTel.
National Grid’s ‘unnecessary’ power reserve cost £180m - and wasn’t used. In the winter of 2014/15 National Grid paid £23.5m to form a reserve bench of 10 power plants which could be used if cold, dark weather caused demand for power to surge higher than the UK’s power market could meet. The payment later spiralled because the rate per unit of generation capacity doubled, the report found. In the first winter of the scheme the rate averaged £15.57 per kW but by the third winter of the scheme National Grid was paying £34.21 per kW. 
(Of course, the Grid makes the obvious point about insurance payments ...)

11 comments:

Blue Eyes said...

Of course plenty of countries do take that view on water. It is only fairly recent in some Western European countries that it is safe to drink the tap water.

Interesting question overall. Not long ago one could not really use a computer in Europe that had been bough in North America. These days even bog standard laptop chargers can handle 100-250V at 50 or 60 Hz. Plug and play.

Obviously with climate change we will need more power for air conditioning but maybe less for aluminium smelting as we go up the value chain.

Demetrius said...

AS the late great James Hunt might have said, it is all getting a bit previous. When the smash then occurred he was all too often right. We have increased the risks of serious outages not only in the UK but across Western Europe. The big one could be coming.

Anonymous said...

The old three day week of Ted Heath was quite nice, as pubs tapped barrels on the bar and put candles on the tables. But I'd presume street lighting will be kept on, as I think it was then. Wouldn't like to be in many of our cities and towns during a total blackout.

Anonymous said...

Street lighting? Thats not the half of it.

http://www.mirror.co.uk/news/uk-news/government-contingency-plans-5-day-6767579

Electro-Kevin said...

Sunday posting ?

Shit. This must be serious.

Hopper said...

Nines of reliability in the software world are expressed in terms of the end user experience otherwise you run the risk of optimising a statistic that no-one actually cares about. If I was measuring "availability" of the National Grid, I would totally do it in terms of end-user availability - every black-out and brown-out would count in terms of minutes per affected customer downtime (and obviously I'd at least segment residential customers from businesses).

Measuring nines is also not terribly useful unless you have consequences - if you go outside your availability target for a measuring period, what do you do to fix it? In the software world, if you have repeated downtime taking you outside your availability target then you look at the root causes and then take action to fix them. Too many bad releases? Switch to a critical-bugs-only policy for releases until your availability returns to an acceptable level. In Grid terms, this might involve a re-examination of the root causes of brown-outs (too many bird mincers, solar not delivering through too many clouds?) and building out diverse capacity to fix it. I wonder how acceptable a coldly objective engineering approach like this would be to the environment fans.

ND's point about determining the maximum useful reliability is also well made. Once you're measuring the correct statistic, you want to make sure you exceed your availability target but not by too much because if you are then you are almost certainly wasting money.

I find it interesting that National Grid quotes gas distribution availability - five nines - but electricity transmission availability. I'm sure that they're not trying to hide anything: http://www2.nationalgrid.com/responsibility/how-were-doing/grid-data-centre/Customer-service-and-network-reliability/

Anonymous said...

Anon 8.42 - I'm trying to imagine what could take the grid out completely for 5 days. Not much I'd have thought. A cursory look at this link returns 0 countries with such a downtime that I could see. Even Lothar/Martin in 1999 which took out 1/4 of France's high-tension lines left most of the country on.

https://en.wikipedia.org/wiki/List_of_major_power_outages

Any ideas, ND? Huge (i.e. not seen in centuries) solar discharge?

Nick Drew said...

Just contingency planning, I guess

This one - https://en.wikipedia.org/wiki/Northeast_blackout_of_2003 - being the USA, probably troubles people more than (frankly) the southern Asian ones

Sometimes for a scenario, there's no need to put a name to a specific cause, just think through the consequences & see what it tells you, a useful exercise in its own right. One of the principles of Stress Testing (I've written abt this before in the banking context, e.g. here) is to go for something more extreme than anything previously encountered: and one of the usual failings of ST is that the scenarios are uninformative - constrained by the lack of imagination of those writing them

That said, I can tell you that they are planning for something extremely bad (dunno if it's as much as 5 days, but B-I-G, anyhow) in S.Germany, where there are all manner of grid constraints and nonsenses. This stems from massive industrial demand in that region, and the grid(s) being forced to accept large and unpredictable quantities of electricity from windfarms in the north of Germany

Alternatively, there must be worries about what happens in a very finely-balanced grid system with (a) lots of interconnections to other countries who will naturally cut off exports at the drop of a hat when the chips are down (pun intended); plus (b) minimal capacity margin; when (c) it's all overlaid by 'smart grid' stuff - with maybe (d) malicious cyber-interventions from who-knows-where

Not wrong to be thinking it through, I'd say ...

rwendland said...

In France they need ~2.3 GW of extra power for every degree Celsius of average temperature drop in winter. Dunno what that is in the UK, but I'd guess 1GW per every degree Celsius drop - the French are more keen on electric heating.

So maybe you could regard the ~3GW Supplemental Balancing Reserve as insurance against an unpredictable seriously cold (-3C) winter periods, as well as generation issues.

I'm disappointed the ECIU report does not examine average winter temperatures against demand highs over recent years, and evaluate those risks. It seems like it would be an important consideration.

Nick Drew said...

Agreed, Mr W. The whole qn of 'optimising' capacity / insurance cover is a complex one

I just don't like it (a) being left to the Grid to make it up as they go along; and (b) not being procured competitively.

At least the latter is being (somewhat) fixed. And maybe the Grid's conflict of interest will be fixed too ..?

Weekend Yachtsman said...

In the same way, there absolutely no nett gain whatever (except perhaps from the knacker-man's point of view) in all pet food being fit for human consumption - yet it is, and must be by law.

Thanks, EU! We won't miss you at all...