Power Systems Without Fuel

I. INTRODUCTION

• Fuel-based generators shape system physics and economics, so removing them raises underexplored questions about objectives, pricing, and operating practices.The authors also identify potential benefits, including eliminating unit commitment and enabling predominantly DC infrastructure.Source: §I · p. 1
• No-fuel systems are characterized by shrinking timescales and numerous small components replacing a few large ones.Renewable intermittency links hourly dispatch to regulation needs, while reduced mechanical inertia makes transient behavior depend on many small, fast power-electronic interfaces.Source: §I · p. 1

A. The soft energy path and the status quo

• The soft energy path envisions energy from numerous small, simple, diverse renewable sources, unlike a few large, complex technologies.The paper assumes almost 100% renewable production is viable for a large share of power systems and excludes nuclear generation, biomass, and biofuels from its scope.Source: A. The soft energy path and the status… · p. 1, B. Setup and organization · p. 2

B. Setup and organization

• The paper surveys a no-fuel future as a limiting case for understanding a potentially different energy future, while also considering mostly renewable systems.Its discussion is organized across planning, steady-state operation, and transient dynamics; flexibility is assumed available, with storage, flexible loads, renewable curtailment, and power-electronic devices among the mechanisms considered.Source: §I · p. 1, B. Setup and organization · pp. 2–3

A. Economies of scale and geographic constraints

• Renewable installations can be built incrementally over shorter periods, while their geographic constraints can require additional transmission investment.Weather-dependent siting and rooftop solar integration may require network upgrades; non-hydroelectric storage has few, if any, geographic constraints.Source: A. Economies of scale and geographic… · p. 3
• Smaller renewable units improve continuous approximations to planning problems, reducing the difficulty of associated integer programs.Finer installation increments reduce sensitivity to discrete installation choices and can improve integer-programming methods such as branch-and-bound.Source: A. Economies of scale and geographic… · p. 3
• Renewables’ shorter planning horizons reduce uncertainty and the risk of wasted investment from long-range forecasts diverging from reality.The paper describes this as reducing the variance of uncertain future information, without necessarily changing expected objective cost.Source: A. Economies of scale and geographic… · p. 3
• More numerous, smaller renewable market participants may be closer to price-takers, enhancing competition and easing market entry.Capacity payments may become more important than coordinated spot-energy markets; energy-only markets may need redesign to maintain reliable capacity expansion as energy-market revenues fall.Source: A. Economies of scale and geographic… · pp. 3–4

B. Coupling to other infrastructures

• Without fuel-based generators, systems avoid fuel transport, waste disposal, and thermal-plant cooling water, reducing planning complexity.These changes remove considerations such as collocating fuel-based generators and desalinization plants.Source: B. Coupling to other infrastructures · p. 4
• Power systems without fuel become more coupled to transportation and communications, with transport integration adding complexity and smart-meter control raising security and privacy vulnerabilities.The transport coupling can decouple transportation from fuel infrastructure; real-time control can enhance performance.Source: B. Coupling to other infrastructures · p. 4

C. AC versus DC and the relevance of frequency

• Power electronics can provide voltage conversion for both AC and DC, while the absence of synchronous generators makes machine synchronization less useful, supporting DC as a viable paradigm.The cited discussion notes that most other loads tolerate AC or locally convert it to DC.Source: C. AC versus DC and the relevance of… · pp. 4–5
• DC transmission offers one-phase operation, efficient wire utilization, enhanced stability, and reduced communication needs; DC is also viewed as advantageous for small systems and high-voltage transmission.For small systems, cited advantages include lower source-to-load losses, no synchronization or reactive-power considerations, and improved isolation from main-grid faults.Source: C. AC versus DC and the relevance of… · p. 4
• DC steady-state operation is mathematically simpler, and convex relaxations are exact under more general conditions than for AC, making standard nonconvex optimal power flow tractable across a wider range of DC systems.The principal DC drawback is protection: persistent current can cause arcing, and viable schemes are more expensive than AC counterparts.Source: C. AC versus DC and the relevance of… · pp. 4–5
• Renewables require new transmission investment for geographically constrained resources, but finer-sized units and shorter installation times make their planning easier than fuel-based generation.Smaller units better approximate continuous planning decisions, while shorter build times reduce uncertainty from long planning horizons.Source: C. AC versus DC and the relevance of… · p. 5

• Steady-state operation concerns matching bulk power supply to demand and how removing fuel-based generators affects that task.Source: §III · p. 5

1) Unit commitment:

• Without fuel-based generators, startup and shutdown constraints fall to minutes or less, eliminating unit commitment and the need to procure reserves far in advance.Wind and hydro can reach maximum output in minutes, while solar and power-electronic storage can engage nearly instantaneously.Source: 1) Unit commitment · p. 5
• Eliminating unit commitment avoids solving the associated combinatorial integer programs, reducing the computational complexity of power-system operations.Source: 1) Unit commitment · p. 5
• Shorter scheduling lead times reduce forecast error, while storage can supply reserves without the standby costs incurred by fuel-based generators.The cited storage-cost comparison assumes low leakage, described as realistic for batteries and pumped hydro.Source: 1) Unit commitment · p. 5

2) Load-shifting:

• The generic load-shifting problem minimizes production costs over hours to days, subject to storage dynamics, storage limits, supply-demand balance, and physical power-feasibility constraints.The storage model represents state of charge and power injection or extraction, with leakage and conversion losses.Source: 2) Load-shifting · pp. 5–6
• Load-shifting provides little efficiency benefit without fuel-based generators when system efficiency depends less on output levels and power prices may not vary over time.Source: 2) Load-shifting · p. 6
• Load-shifting can make operation feasible under production-capacity limits, and matching supply and demand is especially relevant with high shares of non-dispatchable wind and solar.Source: 2) Load-shifting · p. 6
• Peak-shaving shifts consumption to reduce peak demand and system capacity requirements, and remains useful with or without fuel-based generators.Source: 2) Load-shifting · p. 6
• Because storage can change output in seconds, schedules spanning more than a day can be computed and updated close to real time, for example every five minutes.The paper suggests model-predictive or receding-horizon control for these updates.Source: 2) Load-shifting · p. 6
• Load-shifting is critical for accommodating non-dispatchable supply and reducing capacity requirements, while actions can be implemented near real time despite schedules spanning days.The section concludes that there is little, if any, need for day-ahead scheduling without fuel-based generators.Source: 2) Load-shifting · p. 6

B. Dispatch

• Dispatch optimizes device settings through optimal power flow, minimizing operating costs subject to network constraints; no-fuel systems also require new objectives to replace fuel costs.Renewables, storage, and power-electronic devices introduce dispatch decisions such as storage schedules and converter set-points.Source: B. Dispatch · pp. 6–7
• Multiperiod optimal power flow incorporates storage dynamics and serial correlations in renewable inputs while remaining tractable in convex formulations.It couples single-period dispatch problems across time; stochastic and robust models can extend to this setting.Source: B. Dispatch · p. 7

1) Uncertainty:

• Renewable intermittency introduces forecast errors that dispatch models must represent, alongside existing uncertainty from loads and contingencies.Renewable plants can be dispatched up to maximum outputs set by random factors such as wind speed and solar intensity.Source: 1) Uncertainty · p. 7, B. Dispatch · p. 7
• Replacing fuel-based generators trades intractable discrete unit-commitment constraints for renewable uncertainty, which offers more modeling choices and tractable convex approximations.The paper presents this as a hypothesis needing further theoretical and empirical quantification.Source: 1) Uncertainty · p. 7

2) Dynamic constraints:

• Storage and demand response require dynamic modeling because their finite energy capacities affect flexibility; multiperiod optimal power flow couples dispatch decisions over time.Despite scaling to approximately NT variables and MT constraints, the multiperiod version of a second-order cone OPF remains a second-order cone program.Source: 2) Dynamic constraints · p. 7

3) Decentralization:

• Limited communications and computational constraints may require decentralized dispatch of low-voltage devices, with lower online computational requirements but performance loss versus centralized operation.Existing approaches include consensus-based algorithms and the alternating direction method of multipliers.Source: 3) Decentralization · p. 8

4) Dispatch criteria:

• Fuel-free dispatch needs objectives beyond fuel costs, including maintenance, reliability, power quality, and demand-side utility considerations.The paper advises avoiding objective functions that add complexity beyond the physical constraints, such as objectives requiring Monte Carlo integration.Source: 4) Dispatch criteria · p. 8
• Renewable forecast errors must be modeled, while storage dynamics and serial correlations call for multiperiod optimal power flow.These are among the four ways the paper says dispatch differs from current practice.Source: 4) Dispatch criteria · p. 8

C. Economics

• Without fuel costs, applying nodal pricing as-is would produce zero prices; alternative monetary objectives could yield price interpretations, but nonmonetary objectives would not.For nonmonetary objectives, dual multipliers indicate sensitivities of the optimal objective rather than marginal-cost-based prices.Source: C. Economics · p. 9
• It is unclear whether fuel-free systems’ marginal operating costs vary on minute-to-minute or hourly timescales, unlike current systems’ time-varying physical efficiency.The paper notes that demand and system capacity still vary over time.Source: C. Economics · p. 9
• Maintenance, reliability, and power-quality costs are harder to quantify and audit than fuel costs, potentially making related bids more susceptible to market power and gaming.The paper says accurately quantifying operating costs across timescales will be important for new market mechanisms.Source: C. Economics · p. 9
• Demand-side bidding alone reflects willingness to pay, not the cost of supplying power, and can produce zero payments to producers in the described bidding model.The paper cites a pure-strategy Nash equilibrium in which all bids are zero.Source: C. Economics · p. 9
• High fixed and low marginal costs make current nodal pricing dubious for power systems without fuel, and the paper calls for new market mechanisms.Bilateral contracts and spot markets should remain in some form, while flexibility should also be represented in markets.Source: C. Economics · p. 9

IV. TRANSIENT BEHAVIOR

• Fuel-free systems lack the mechanical rotor inertia that dominates present-day dynamics, while power-electronic interfaces provide limited innate inertia.Batteries and flywheels are exceptions among storage devices identified as electrochemical or electromechanical.Source: §IV · p. 9
• Steady-state dispatch has greater effects on transient stability, since greater PV output also increases the magnitude of disturbances contributed by the plant.The paper notes that reduced inertia may increase disturbance sensitivity, while power-electronic interfaces can act on faster timescales than synchronous generators.Source: §IV · pp. 9–10
• Quasi-stationary phasor models lack the fidelity needed for fast-switching converters, motivating revised modeling, analysis, and control frameworks.The paper identifies fast converter dynamics as an analysis challenge.Source: §IV · pp. 9–10

A. Modeling

• Inverter models are difficult to abstract like swing equations because digital controls, switching behavior, and proprietary nonlinear device characteristics complicate representation.Switching-time-scale behavior is modeled with hybrid automata, but scalable simulation platforms are lacking.Source: A. Modeling · p. 10

B. Analysis

• Quasi-stationary Kron reduction cannot capture transients faster than AC-cycle timescales, motivating scalable network reduction methods that work across timescales.Kron reduction retains specified terminals and eliminates other nodes using the admittance matrix; that representation relies on quasi-stationary sinusoidal steady state.Source: B. Analysis · p. 10

C. Feedback Control

• Without fuel-based generators, inverter filter elements provide limited inertia, making electrical dynamics faster and distributed feedback control more important for stable voltages and frequencies.Grid-forming controls regulate terminal voltage and frequency without a stiff grid voltage; time-domain methods act directly on waveforms rather than assuming sinusoidal steady state.Source: C. Feedback Control · pp. 10–11

D. Coupling between the transient and steady-state timescales

• Steady-state dispatch affects disturbance locations and transient regulation needs, so the section questions strict separation of energy management from real-time control.Jointly addressing both timescales is presented as a route to better performance, and more frequent energy management is also described as beneficial.Source: D. Coupling between the transient and… · p. 11

1) Feedback control and energy management:

• Dynamic KKT controllers use states that proxy for Lagrange multipliers and evolve toward optimal dual solutions, deriving physical-layer reference inputs from those states.They seek the KKT conditions for optimality while integrating optimization-based controllers with existing real-time controllers.Source: 1) Feedback control and energy… · p. 11

2) Economics:

• Fast regulation can incur resource maintenance costs that affect steady-state dispatch, so economic mechanisms must account for costs incurred on transient timescales.The section also says renewable disturbances should be paid for by both renewables and the loads they serve.Source: 2) Economics · pp. 11–12

V. SUMMARY

• Removing fuel-based generation requires more cost-effective storage and demand response, new optimization and control algorithms, and complementary economic mechanisms; the authors consider these challenges surmountable.The paper also notes potential advantages including closer-to-real-time decisionmaking and the viability of predominantly DC infrastructure.Source: §V · p. 12