| Episodes | 47 |
| Mentions | 84 |
| Cited here | 16 |
| First — last | #44 — #729 |
| Top guests | Ari Gerstman, Paul Zawada, Jeff Keyzer |
| Related | solar power · solar panel |
The power grid is the interconnected system of generation, transmission and distribution infrastructure that supplies electrical energy, and it is plausibly the largest and most complex machine ever built.[583] Its defining operational problem is the continuous matching of generation to load, a balance communicated to every participant through a single physical quantity: grid frequency.[583] Modern society depends not merely on electricity but on its consistency — power available instantly, constantly and cheaply, such that operating a switch is not a decision — and that consistency, rather than raw capacity, is the property that separates a functioning grid from a nominal one.[102]
Operation and control
Grid frequency is the signal that indicates whether generation matches load: every generator tracks it in a feedback loop, so the balance between supply and demand is communicated across the entire system by a quantity that every participant can measure locally.[583] Control is layered and distributed. Balancing authorities — sometimes non-profit entities, sometimes utilities — ensure generation matches load at a high level, while minute-by-minute regulation happens inside each generator’s own control system, whether that generator is a steam turbine, a wind turbine or a solar inverter.[583] There is no central authority operating the grid; energy policy instead works by incentivising technologies that make the system more efficient, more reliable and less in need of upgrades, so change arrives through economics rather than through instruction.[630]
Energy entering the grid joins a pool. No particular joule can be traced from a specific plant to a specific load, which means claims about consuming power from one source are statements about accounting rather than about physics.[583]
The computing and communications attached to physical plant constitute operational technology, a discipline distinct from both power engineering and information technology. The same skills apply across power, transportation and pipelines, and the work sits where embedded systems meet equipment measured in megawatts.[583] The industry is unglamorous and struggles to attract engineers, a problem that matters because the expertise is not easily replaced.[583]
Loads and scale
A single domestic appliance switching on is noise once a thousand houses are connected in parallel: people switch off as others switch on, and the aggregate is smooth. The loads that actually move the system are industrial, such as a factory bringing its lighting up at once.[96] At the largest scale, a single site becomes a grid-level actor. John Davis described an aluminium smelter paying seven million dollars a month for electricity, which carried monitoring on its incoming supply because a large motor going to ground could force the shutdown of a substantial fraction of that state’s grid.[385]
Remotely controllable appliances create a new class of coordinated demand step. If a large population of connected air conditioners were commanded on simultaneously, the resulting demand would appear instantly rather than following the diurnal curve the system is designed around.[303]
Generation mix and economics
Generators earn disproportionately at peak: bringing an extra turbine online on a hot afternoon commands the highest rates of the year, while the steady baseload business is comparatively unremarkable. Solar generation arriving precisely at that afternoon peak therefore attacks the most profitable hours rather than the average ones, which is why it disrupts grid economics faster than its share of energy suggests.[150]
Nuclear generation is designed for steady state — start it and run it until refuelling, avoiding fast changes — which makes it the opposite of the newer intermittent sources in the respect that matters to an operator: it is consistent, and it does not follow load.[365] Operators object to renewable generation on grounds of intermittency rather than principle, because a grid is built to match a varying load with controllable supply, and a supply that varies independently of load makes that job harder.[610]
Distributed generation
Distributed generation behaves like bulk capacitance on the system: it smooths the ripples as loads switch on and off, and adds capacity where it is consumed.[150] When distributed generation exceeds what a feeder can absorb, the available responses begin with curtailment — simply turning the generation down — which is the outcome least attractive to the owner and cheapest for everyone else.[630] A utility’s real interest in an interconnection application, beyond installation safety, is retaining the ability to shut the generation off; a feeder carrying more supply than local load is a genuine problem, and curtailment is the mechanism of last resort.[630]
A microgrid is not simply sharing power with neighbours. It is a defined sub-portion of the larger grid that can be understood and operated as a unit, which makes it a planning tool rather than a description of what already happens between adjacent houses.[630]
Against a genuinely reliable grid, domestic battery storage is difficult to justify on economics alone, because the grid is already performing the storage function more cheaply; the argument for it is autonomy or an unreliable supply, and should be made on those terms rather than on payback.[150]
Interconnection and regional variation
Grids are not one system even within a country: Texas operates a grid deliberately separate from the rest of the United States. Internationally the differences extend to phase arrangements and frequency, and even physically interconnected regions can run incompatible metering standards.[677]
Reliability and failure modes
Reliability varies enormously by region. One major metropolitan grid, the Sydney Basin, has never gone down in its entirety in fifty years, with outages confined to pockets of tens of thousands of customers lasting under a day.[317] By contrast, a whole-state blackout followed a severe storm that took down major interconnectors: automated protection tripped correctly, but insufficient internal capacity remained to carry the load alone. Robustness is a property of how the system is architected — including how much local generation can stand alone — rather than of the generation mix in isolation.[317]
A major grid failure traced back to frozen natural gas lines, because gas provides much of the fast-responding instantaneous generation. The lesson generalises: a grid’s resilience depends on the fuel supply chains beneath it, which are rarely modelled alongside the electrical system.[656] Geomagnetic disturbance remains a live disaster scenario for transmission infrastructure, with historical precedent for a single event affecting the grids of multiple countries — a risk that scales with the length of the conductors rather than with the sophistication of the equipment attached.[265]
Metering and energy accounting
Under net metering, generation consumed on site never reaches the meter, while excess is exported at a much lower rate than importing costs. The rational response is to move consumption into the generating hours — running the washing machine at midday — because using one’s own energy is worth more than selling it.[205]
Older installations used two separate meters, one counting import and one counting export, where a modern smart meter reports both through a portal as a daily figure; that daily record is what makes it possible to notice a small persistent anomaly at all.[677] A battery inverter typically has a programmable trickle setting that draws a small continuous amount from the grid to keep the connection alive, and anyone investigating an unexplained standing import can test that hypothesis directly by changing the setting and watching whether the daily figure moves proportionally.[677] A small persistent discrepancy may also simply be the meter itself: at roughly one percent tolerance, an apparent standing import of a fraction of a kilowatt-hour a day can fall inside the instrument’s specification.[677]
Electrification of transport
Electrifying transport is a grid problem more than a vehicle problem: moving that power through the network requires building new infrastructure and retiring old, and industry estimates of how long that takes have compressed from decades to a small number of years, which is itself a source of planning risk.[495] A well-to-wheels accounting that followed the diesel for the mining equipment, the transport, the generation, a 7.5 percent delivery loss to the garage and the charging loss found 27 percent less carbon dioxide per mile than a gallon of petrol — even assuming every kilowatt-hour came from coal, and at a measured 3.5 miles per kilowatt-hour taken from the grid.[112]
References
| Episode | Title | Date |
|---|---|---|
| 96 | Senseless Saccadic Shemozzle | |
| 102 | Gouging Green Gardyloo | July 1, 2012 |
| 112 | An Interview with Bob Simpson - Ardent Automotive Artisan | September 9, 2012 |
| 150 | Solar, FPGAs and Maxim Integrated - Solar Shopper Sickness | June 17, 2013 |
| 205 | Solar Factories and HVDC Lines - Pollent Power Pushing | June 30, 2014 |
| 265 | A Security Update with Michael Ossmann | September 2, 2015 |
| 303 | An Interview with Dmitry Nedospasov | June 14, 2016 |
| 317 | A Decoupled Episode | September 28, 2016 |
| 365 | Wait, why is Jeff glowing? | October 30, 2017 |
| 385 | An Interview with John Davis | March 25, 2018 |
| 495 | An Interview with Eric Klein | June 7, 2020 |
| 583 | The Smart Grid with Paul Zawada | March 27, 2022 |
| 610 | Picking a Pick and Place Pickiness | November 20, 2022 |
| 630 | Renewable Energy Policy with Ari Gerstman | May 2, 2023 |
| 656 | Pneumatic Tubes, Straight To The Home | January 22, 2024 |
| 677 | Watt Is The Deal | September 23, 2024 |