Benjamin F. HobbsTheodore K. and Kay W. Schad Professor of Environmental Management

Whiting School of Engineering, JHUFounding Director, E2SHI

Chair, CAISO Market Surveillance Committee

Workshop on Economic Approaches to Understanding and Addressing Resilience in the Bulk Power System

RfF 30 May 2018

I. Traditional resource adequacy (independent or correlated forced outages)

III. Long, wide-area network outages

II. Extreme supply outages, multiple stressors

General principles

Three classes of disruptionsMainly markets Primarily Planning

When rely on markets?

◦ When solutions not obvious, require private information

◦ Actions by many (price-taking) parties needed

◦ Responsibility/property rights can be assigned & traded

◦ When events are (relatively) frequent, not severe

When regulation/central planning?

◦ Solutions obvious

◦ Actions by few or 1 party needed

◦ Public good, markets likely to be suspended in event

◦ Events rare, potentially catastrophic

Assume: ◦ Generator outages are random ..

..and (conditionally) independent of each other & of load

Can be extended to: ◦ Correlated outages, e.g.: 100 MW unit, 0.1 FOR, 0 correlation ELCC = 92 MW

Classic engineering methods: ◦ LOLP by convolution methods (Billinton/Allan)◦ Expected Load Carrying Capability (Garver, IEEE TPAS, 1966)

= Increase in peak load that can be accommodated by adding resource … while maintaining LOLP

If 0.3 correlation ELCC = 55 MW

(Based on N load, 2000 MW peak hour, identical units, LOLP = 2.4 hr/yr)

Saint Fred (Schweppe): Spot markets with appropriate scarcity pricing alone can incent optimal investment and flexibility

Capacity markets can too, in theory◦ Desirable if: scarcity underpriced in spot markets, or long run contract

markets absent◦ General result (Bothwell, Hobbs, Energy J, 2017): There exist supporting prices for optimal generation mix:

Capacity market pays each generator: (marginal capacity contribution) * (market price for capacity)

MCC = ∂(expected unserved energy)/∂(capacity)

Need good rules◦ Credits based on marginal contributions◦ Forfeit payments if unavailable when needed Simulate impact of efficient spot market

◦ Look-ahead, locational

4 calls/mo demand response Fly wheels (15 minutes stored)

However, awkward way to incent flexible investment◦ CAISO FRACMOO. But how do you compare the following?

Fully dispatchable CTs Renewables that can turn down 1 start/day resources

◦ Problem: What’s the marginal contribution when: operating conditions & system response are complex? event probabilities, failure mechanisms poorly known?

◦ Belts and suspenders (Bushnell, Harvey, Hobbs, CAISO MSC): Improve spot markets to reward output when system values it

California 2000-01: Seven Plagues of Egypt ◦ Fuel (compressor outage)◦ Hydro shortage◦ NOx allowance shortages◦ Kelp

Compounded by market design failures

caiso.com

“When sorrows come, they come not (as) single spies. But in battalions.” (King Claudius, Hamlet Act 4 Scene 5,Thx to Kousky and

Cooke, RFF-DP-09-03.pdf)◦ Multiple events without precedent, no agreed upon probabilities

Feb. 1-4, 2011 cold snap 210 units on outage High load 4 GW curtailed

www.powermag.com/prepare-your-gas-plant-for-cold-weather-operations/

www.carbonbrief.org/analysis-the-legacy-of-the-fukushima-nuclear-disaster

Fukushima: 17% of Japan’s capacity lost

Not spot markets alone. For extreme events: ◦ Probability estimates unreliable◦ Insurance unavailable or costly

“...three particular phenomena of climate related risks that will require a change in our thinking about risk management: global micro-correlations, fat tails, and tail dependence.

“(Their) consideration …will be particularly important for natural disaster insurance, as they call into question traditional methods of securitization and diversification” (Kousky & Cooke, RFF-DP-09-03-REV.pdf, 2009)

How to prepare against the unimagined risk?◦ Inherently stable systems: diverse sources more redundancy, island-able

◦ Achievable by: Baskin & Robbins capacity markets higher reserves, grid reinforcements black-start, microgrids

◦ Yet what are we willing to pay for that stability? Is the cure worse than the disease (regulatory rent seeking, capture…..)?

(Mukherjee, Nateghi, Hastak, Reliability Engineering & System Safety, 2018)

Cascading outages, system collapse◦ Managing frequency excursions

in a renewable heavy system◦ Cyber insecurity What is Russia really capable of?

www.cartoonistgroup.com/properties/anderson/art_images/030823grid.jpg

Electromagnetic disturbance◦ Solar flares◦ Twitchy fingers

Natural disasters

Fire Earthquakes &

transformer replacement(Enders et al. 2010 Energy Systems)

www.businessinsider.com/pictures-of-californias-latest-wildfire-2016-6

www.tdworld.com/test-monitor-control/new-seismic-shock-system-helps-protect-power-grid

Public good of network reliability◦ Central planning◦ NERC rules◦ Tight FERC oversight In identifying risks, remember Tversky/Kahnemann biases:

imaginability, over-confidence, ….

(Quasi) market roles◦ Bidding to provide equipment, services◦ Performance-based ratemaking for grid owners

I. Traditional resource adequacy (independent & correlated outages)

Mainly markets Primarily Planning

III. Large & long network-caused outages

II. Extreme supply outages due to multiple stressors, extreme events