(John Kemp is a Reuters market analyst. The views expressed are his own)
LONDON, Aug 18 (Reuters) - California’s recent blackouts have illustrated the risks when the transition to renewable generation proceeds faster than measures to make electricity consumption more responsive to available power supplies.
In the long term, California has ambitious plans for a grid dominated by wind and solar generation, backed up by variable pricing, load-shifting and lots of grid-scale and customer-scale storage.
But the state has failed to sequence the transition properly, phasing in wind and solar generation and phasing out other sources too quickly, before consumption has been reshaped and storage has been scaled up sufficiently.
For more than a century, the centralised electricity industry has been based on the principle generation-follows-load (i.e. production follows and responds to consumption rather than the other way around).
Consumers decide how much and when they want electricity, and then generators and transmission operators respond to ensure there is always sufficient power available to meet demand at the flick of a switch.
Individual consumption by households and businesses is highly variable but aggregate power use is fairly predictable based on time of day, day of the week, time of year and weather forecasts.
Electricity system operators are able to produce accurate load curves forecasting power consumption – for example the daily curve published on the internet by the California Independent System Operator (CAISO).
The task of generators, transmission operators, power distributors and reliability organisations is to match production with the load curve, hour by hour, minute by minute and second by second.
For the most part, the modern power system is based on passive load-following. Prices play a limited role in balancing short-run electricity demand with supply, at least for most retail and small business customers.
For example, California’s predicted load curve for Tuesday Aug. 18 forecasts system-wide load will hit a low of 27.1 Gigawatts at 0420 local time before rising to a peak of 50.5 GW at around 1740.
In most cases, load-following is achieved by ramping up and ramping down generating units over the course of the day to match the predicted curve.
Some additional capacity is also maintained to act as a reserve to deal with unexpected generation outages such as the loss of a power station or a transmission line, or short-term variability in consumption.
While some generating units operate non-stop to provide base load, others cycle on and off over the course of 24 hours to follow the curve and provide short-term reserves.
Gas-fired units are ideally suited to load-following and provide short-term reserves because their output can be ramped up and down rapidly in a time period ranging from minutes to around an hour and a half.
Coal-fired units take much longer to ramp up and down (3-12 hours) but can still follow load if given enough notice based on the forecast curve.
Nuclear generators are designed to operate at constant output and cannot efficiently follow load so are normally used to provide base load only.
Wind farms and solar generators produce variable output depending on time of day, season and local weather conditions, so they too cannot follow load.
Some of the variation in wind and solar output is predictable, based on seasonality and weather forecasts, but there is also a random short-term element which is harder to control.
As the proportion of wind and solar power integrated onto the grid increases, the traditional system will have to be reversed so load-follows-generation to a much greater extent.
Electricity consumption will have to be reshaped so the load curve is adapted to conform more closely to the varying availability of wind and solar generation.
Reshaping the curve will involve much greater use of variable prices to encourage retail and smaller business customers to use electricity when it is most plentiful and avoid using it when the system is stretched.
Reshaping the curve will also involve much greater use of battery storage at grid-scale and in individual homes and offices, for example through electric car batteries.
Reshaping the curve is not a new idea. Britain’s massive expansion of nuclear generation in the 1980s was accompanied by promotion of night-storage heating to soak up excess generation in the middle of the night.
But the reshaping that will be needed to accommodate a very high share of wind and solar on the grid is far more ambitious than anything the industry has attempted before.
California’s power blackouts illustrate the scale of the challenge.
Solar output peaks in California around the middle of the day, while load peaks around six hours later in the early evening and remains high for several hours afterwards.
On Aug. 14, CAISO issued a warning about system capacity in the early morning (0900 local time), then stepped up the warning about the risk of forcible load shedding in the late afternoon (1646).
But with solar generation fading rapidly, while air-conditioning load was still very high, the system operator initiated forcible load shedding in the evening (1925).
The blackouts had nothing to do with intermittency (the difficulty of predicting short-term variations in wind and solar output). The problem stemmed from the well-understood ramp down of solar output during the evening.
The system failed either to reshape the load curve to match the solar generation curve or to carry enough spare capacity from other generators to make up the evening shortfall.
Proponents of wind and solar have outlined plans to move to a system where load follows generation through smart meters, smart appliances, time-of-use pricing, grid-scale batteries and grid-connected electric car batteries.
But the introduction of more non-dispatchable generation has so far outstripped efforts to make load conform to the available generation.
California’s regulators now have four options:
(1) slow the rollout of renewables and changes in the generation mix;
(2) increase the availability of dispatchable gas-fired electricity reserves;
(3) accelerate the introduction of battery storage; and
4 accelerate the introduction of variable electricity pricing to compel more load-shifting. Editing by Susan Fenton
Our Standards: The Thomson Reuters Trust Principles.