Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
For the out-of-band Slashdot experience (mostly headlines), follow us on Twitter, or Facebook. ×

Comment: Re:Why bother with installed capacity? (Score 1) 259 259

> Really? Every grid-tie inverter, ever, has this built in.

Exactly. This is why it has been a problem. In the event of a grid disturbance, a large amount of grid-tie inverters falsely detected an island and tripped out simultaneously.

> Pointer to the UK issue you're referring to?

Although the initiating cause was 2 large thermal plants tripping, the rapid reduction in grid frequency caused approximately 300 MW of grid-tie inverters to trip simultaneous, accelerating a grid frequency collapse event.

National grid has recommended new grid-tie inverter firmware configurations, which should be present in new inverters, but this reduces the sensitivity to true islanding, but this is now considered a lesser issue than loss of embedded generation.

Comment: Re:Why bother with installed capacity? (Score 3, Informative) 259 259

It depends on the nuclear plant design, but in general a LWR nuclear plant has better load following characteristics than a coal plant (the compact core has low thermal mass compared with the large furnaces and boilers of a coal plant). Load following, however, may not be permitted under regulatory regimes (for example, in the US). The availability characteristics of a typical LWR are not much different to coal - there are longer, less frequent maintenance/refuelling periods.

A 100 MW wind farm is not a perfectly good substitute for a 100 MW coal plant. Typically the load factor for wind is in the 25-35% range, with off-shore wind being higher. However, the "firm capacity" (i.e. the capacity that can be relied upon) is poor - about 2% in the UK, whereas for coal it is about 85%.

Both wind and solar have very poor load following capacity - as they cannot automatically respond significantly to changes in grid frequency (except in the case of overfrequency), whereas most thermal plants have the capacity to automatically increase power, provided they are not at capacity, in response to a drop in grid frequency. Im Germany, this is partly mitigated by requiring that rooftop solar installations be electronically limited to 70% of their nominal capacity and/or be able to receive remote configuration updates from the utility, so that there is frequency reserve margin.

Small scale embedded generation (i.e. rooftop solar) has an additional problem which is that of grid failure detection and anti-islanding (i.e. the embedded generators must not be allowed to supply energy to the local area in the event of failure of grid connection). The problem is that grid instability is not easily discriminated from grid islanding, hence there is a tendency for a severe grid imbalance to trigger cascading disconnections of small generators, which makes the imbalance worse.... This has happened in the UK, and very nearly caused a country-wide blackout. It was only arrested when underfrequency protection started blacking out regions of the country in an attempt to reduce load on the grid.

Comment: Re:Good god. (Score 1) 253 253

Except in an aircraft you have multiple redundant engines, so the idea of dropping a faulty engine to idle is not that unreasonable.

The problem here was 3 faulty engines, for which there was insufficient redundancy - in this case a "common cause" failure that's a much more difficult problem to deal with.

Comment: Re:Good god. (Score 3, Insightful) 253 253

My guess is that rather than "files" per se, these are look-up tables which were statically linked into the binary.

On this type of safety critical application, it's a key design aim to avoid code which might fail or throw an exception at runtime. So, rather than load data from a file, which could fail due to a memory allocation failure, a file system failure, etc. the relevant data is static linked, so if the executable successfully launches, it cannot fail to have the data available.

I don't know what these tables might have been mapping, but conceivably if they torque tuning parameters, the engine might still have run if the data was all NULLs, but delivered the incorrect torque in response to control inputs. Of course, if the missing data was things like fueling data, then the engine may have failed to start.

Comment: Re:Hack piece (Score 3, Informative) 126 126

There was a big change in design philosophy. Early reactor designs were intended to prevent meltdown and had limited mitigation. More recent designs now include substantial mitigation as well as more robust prevention strategies.

E.g. The fukushima accident occurred because of a "common cause" failure of multiple safety critical systems - the redundant diesel generators. This failure led to a "cliff edge" cascading failure of numerous safety systems, effectively meaning that core melt was inevitable. (This is in addition to the incorrect site risk assessment, where an incorrect tsunami risk was used when assessing the suitability of the site for a nuclear power plant, and the additional failure to mitigate that risk when the tsunami risk was recognised in the 1980s).

Most modern reactor designs (the EPR excepted) do not class their diesel generators as "safety critical", because they are not necessary to place the plant in a safe state and initiate adequate reactor cooling. In addition, nuclear regulators (Japan excepted) around the world started carefully investigating "cliff edge" scenarios following the 9/11 attacks, to see if deliberate sabotage could result in disproportionate failure of safety features. In the US, the NRC started mandating that "safety critical" diesel generators be heavily hardened against beyond design-basis natural events and other methods of attack, even if not originally conceived at design stage; that UPS batteries be upgraded to provide up to 24 hours of safety, in order to allow emergency assistance to be called in, and/or that additional electrical power sources (e.g. gas turbines) be installed in fortified near-site (to mitigate against local site damage) installations.

A similar set of upgraded mitigations have also been in place for a while - hydrogen catalytic recombiners (these are basically catalytic converters similar to those in a car exhaust which react hydrogen and oxygen at a low temperature and low hydrogen concentration, well below the minimum ignition level. Heat generated from the recombination is used to cause natural circulation of air through the combiner to accelerate hydrogen removal and stir up the air to ensure that hydrogen cannot pool away from the recombiners) have been installed in-containment, and in buildings close to hydrogen vent pipes. In Fukushima, no hydrogen recombiners were used, instead the main containment building was inerted with nitrogen. As a result, hydrogen (and steam) built up in the containment pressurising the building. In order to reduce pressure to prevent rupture, the containment building was vented into the main reactor building, where the hydrogen mixed with air and later ignited. More modern designs vent directly outside through filters, or vent through hydrogen recombiners.

The other complicating issue is that at Fukushima unit 1, the reactor core appears to have completely melted through the reactor vessel into the containment building, severely contaminating the water in the containment building which was being used for cooling (and also leaked through minor damage to the containment). Again, modern designs try to mitigate this. The AP1000 design fills the bottom of the reactor vessel with low-melting point, sacrificial material into which molten core material will melt, resulting in dilution, prevention of re criticality, and spreading of the decay heat. Then by flooding the containment building and submerging the reactor with water, "melt through" is prevented because of combination of external cooling water and the diluted core material, as a result the containment building itself is not contaminated. The EPR instead, has a special chamber beneath the reactor intended to spread and retain molten core material, in such a way that it would not contaminate the containment building.

Comment: Re:Pointing out the stark, bleeding obvious... (Score 1) 247 247

The grid technical paper specifically listed multiple different sizes of OCGT and their ramp rates, so I presume that they do matter.

However, checking the specification sheet for a state of the art large turbine (GE 7HA), the ramp rate is 40 MW/min for the 275 MW model, with a manufacturer claimed startup-signal to full load time of 10 minutes.

By contrast, checking the spec sheet for the same manufacturer's small turbine, they claim that the turbine can ramp to 20 MW (45%) from idle within 5 seconds. I could well imagine that such a turbine could start, synchronize and ramp to full power within 1-2 minutes.

Comment: Re:Pointing out the stark, bleeding obvious... (Score 1) 247 247

My figures were taken from a 2012 report by my local grid operator, based upon operational data supplied by the power plant operators.

The figures are conservative, but they are based upon the figures declared by the plant operators, based upon existing plant, but some consideration has been given to new-build plant. I accept that I omitted the issue of ramp "elbow" for CCGT, but that was for simplification.

As to the ramp rate of OCGT, it varies with size. Aero-derivative OCGT (20-60 MW range) can certainly come to full power within 3 minutes. Large frame OCGT (200 MW range) are slower. Even a state-of-the-art turbine needs at least 10 minutes to come to full power from cold shutdown. Most existing plant is slower.

Comment: Re:Pointing out the stark, bleeding obvious... (Score 1) 247 247

Nuclear plants are modestly controllable, but it is rarely done, because the cost savings of ramping down are negligible, so typically nuclear plants only ramp down for operational reasons, or grid acceptance reasons. In countries with large amounts of both nuclear and renewable power, nuclear plants operate in a load-following mode, ramping up and down with demand/renewable supply.Old nuclear plant normally offer ramp rates of 2.5% of nameplate rating per minute, with more modern plants offering 5% or greater. There can be some issues with ramping older plants because of temperature changes in the reactor which can contribute to fatigue and limit the reactor life time. Modern plants are designed for isothermal ramping to prevent reactor thermal fatigue from load following operations.

Large coal plants typically can achieve approximately 2% per minute, with the more modern coal gasification combine cycle plants achieving approximately 3%. The big problem with coal plants is start up time after a shut down. A hot start (48 hours) can incur an 8-12 hour delay.

Most existing combined cycle gas turbines can ramp at approximately 3% of rating per minute, with a 60 minute start up delay from warm, or 3 hours from cold. Modern (new build) combined cycle gas turbines can ramp at approximately 5% of rating per minute (when hot), or approximately 1% per hour from cold start with a 15 minute start delay.

Open cycle gas turbines can ramp at approximately 10% of rating per minute, with a cold start delay of approximately 10 minutes.

The big advantage of OCGT is that they can start from cold with minimal notice, so for short-term peaking, they are excellent.

Modern CCGT has most of the benefits of OCGT, but a very much higher capital cost - so there needs to be adequate baseload demand to make the economic case for CCGT, even though efficiencies can be considerably greater with CCGT (62% for a state-of-the-art CCGT, compared with 38% for state-of-the-art OCGT).

In California, utilities are building OCGT like crazy, because it's the cheapest way to provide rapid start standby capacity when the Spring/Autumn Sun starts to go down, just as demand starts to peak.

Comment: Re:Can you get this in concentrated form? (Score 2) 190 190

It is possible to sorb ethanol into a dextrin. The problem is that the volume/mass of sorbent is much larger than the amount of alcohol that can be bound.

So, if you want to bind 10 ml of ethanol (approximately 1 shot), then you may need 100 grams of powder. Which makes the product of limited value.

If, however, you want something iso-intoxicating to 10 ml of ethanol, you can reasonably safely do that with about 500 ul of 2-methyl, 2-butanol, which could be sorbed in 5 grams of powder. The latter is a practical product which meets the description of "palcohol"

Comment: Can you get this in concentrated form? (Score 2) 190 190

Never mind the powdered form, what about getting the liquid in concentrated form?

"Palcohol" is not ethanol, but the highly intoxicating 2-methyl, 2-butanol, which is about 30x as potent at causing intoxication as ethanol. Despite being termed one of the "toxic alcohols", it probably has lower chronic toxicity than ethanol, as being a tertiary alcohol, it cannot be oxidised to toxic aldehydes/ketones.

Comment: Re:What about military satellites (Score 1) 178 178

Not quite correct. Galileo (when it is eventually commissioned) will specifically have the ability to detect the signal from 406 MHz from emergency locator beacons.

Because existing beacons use signals not designed for time of arrival detection, location would still rely on Doppler processing techniques, but location to within 1 mile or so should be achievable with this system. There are plans to change the modulation of emergency locator beacons to permit time-of-arrival localisation with 10 meter precision.

Comment: Re:Are we sure it went south? (Score 3, Informative) 208 208

There are several sources of Doppler shift and compensation. There is Doppler shift between aircraft and satellite, and between satellite and ground station. The ground station automatically compensates for all the Doppler shift between GS and satellite.

The Doppler shift between aircraft and satellite is partially compensated by tracking the Doppler shift in transmissions from the satellite to the aircraft. Without compensation by the aircraft, Doppler shift would be in the region of 300-400 Hz, which exceeds the bandwidth of the channel allocation. The compensation is subject to local oscillator error in the aircraft transceiver, hence individual aircraft will apply the compensation slightly differently.

Although the degree of compensation varies between aircraft to aircraft, it could be fitted with a standard linear regression. This method was apparently verified by Inmarsat on several other aircraft with similar transceivers, and was calibrated based upon transmissions with known locations/velocities.

Comment: Re:I have said it before (Score 1) 384 384

Indeed, most of the issues with the EPR could have been predicted and prevented - it has been a clear example of inept project management. The major problems have been:
  • Out of spec concrete: Nuclear grade concrete needs strict porosity control, and very large seamless pours. Conventional concrete formulations and QA techniques (slumping) are not feasible, and advanced formulations blended to strict proportions are needed. The problem is that the contractors they employed to do the concreting lied about their ability to make nuclear grade concrete. When the regulator inspected the site in Finland, they found that the concrete contractors were blending the aggregates and cement without taking into consideration the water adsorbed onto damp materials. As a result the concrete did not meet the porosity specification. The foundation slab had to be relaid.
  • Out of spec welding: This was further compounded by the fact that many of the required welds were first-of-a-kind, requiring welding of unique dissimilar metals at unprecedented scales in very difficult configurations, requiring the development of new welding equipment as it became clear during dummy runs, that existing equipment could not achieve the quality required
  • Problems with scale: The sheer size of the EPR and volume of concrete and steel for the containment made it difficult to source enough workers of any skill. There were frequent communication problems due to language barriers, and it was difficult to ensure that all staff were kept aware of issues. For example: Areva were unable to source enough welders locally. Welders from as far away as Bulgaria were brought in. However, due to language difficulties and inexperience with the QA required, many welds were not made to adequate quality and had to be remade.
  • Overly aggressive construction schedule: The planned construction would have been the fastest construction of a nuclear plant ever, quite a bold claim, considering that hte EPR is also the most complex ever builtstarted.
  • Insufficient skilled staff: Construction started before design was complete. In particular, control systems had not completed design and verification. The prime contractor also had insufficient architects and engineers to ensure that all designs had been completed to the level of detail required for construction.
  • Failure to validate the supply chain: Construction started before the designers had adequately assessed the global supply chain for parts. There were numerous delays due to excessive lead times for parts which had not been planned for, particularly as many part manufacturers had wound down their facilities due to the death of nuclear plant construction in Europe.

Similar issues, but to a much smaller extent have also cropped up in France on their EPR construction. They've had problems with poor quality welding too, as well as difficulty finding enough competent staff.

Comment: Re: What a wonderful name! (Score 1) 267 267

You mean 200%. Electricity costs in the UK are $0.06/kWh in bulk. Offshore wind has a breakeven price of $0.25/kWh. Things may change in time, and I wouldn't suggest withdrawal of subsidies until we see the potentials and unavoidable drawbacks of the technology. However, to illustrate the point, the UK government recently announced a plan to cut the offshore wind guaranteed purchase price to $0.24/kWh and suddenly a whole bunch of investors threatened to pull out of projects, so the govt backtracked.

The sooner you fall behind, the more time you have to catch up.