Homeostatic Utility Control in Retrospect Gridwise Architecture Council

Homeostatic Utility Control in Retrospect Gridwise Architecture Council 4 th Transactive Energy Workshop J. L. Kirtley Jr. kirtley@mit. edu 12/10/13 Homeostatic Utility Control 1

Homeostatic Utility Control • Arose from an initiative of Fred C. Schweppe and his colleagues in the late 1970’s • Was actually named by Richard Tabors • Was part of Schweppe’s efforts to reinvent the electric utility system • Was intended to make the grid work better • Was developed in the context of regulated public utilities • When first presentation, was rather roundly ridiculed • Has provoked quite a lot of research 12/10/13 Homeostatic Utility Control 2

There were three basic elements of Homeostatic Utility Control: • Frequency Adaptive Power Utility Regulator (FAPER): Fast control • The Energy Marketplace • Marketplace Interface to Customer 12/10/13 Homeostatic Utility Control 3

• Frequency Adaptive Power Utility Regulator (FAPER): Fast control • The Energy Marketplace • Marketplace Interface to Customer 12/10/13 Homeostatic Utility Control 4

FAPER • Intended to replace (or supplement) ‘Spinning Reserve’ • Fast (virtually instantaneous) control • Takes advantage of ‘average power’ or ‘energy’ type loads • Works only within the hysteresis band of an energy load • Within that band, turns load on or off according to frequency 12/10/13 Homeostatic Utility Control 5

FAPER would help the system, but why would customers want to install them? • There is some expense • Probably negligible impact on comfort of equipment operation • Here is Schweppe’s suggestion for compensation • Charge less when frequency is high and more when frequency is low! 12/10/13 Homeostatic Utility Control 6

Later work on FAPERs at MIT: Kevin Brokish • Surprisingly large fractions of consumer loads are amenable to FAPER operation • Brokish assumed operation to be a modification of setpoint in a hysteresis type control 12/10/13 Homeostatic Utility Control 7

Brokish recognized • an instability that can arise with a lot of FAPERS controlling loads by switching them on and off • Loads synchronize with each other • Solution is like Ethernet communications: use probabilistic delay • White band is the dead band • Outside dead band, lighter blue is higher probability of switching 12/10/13 Homeostatic Utility Control 8

12/10/13 Homeostatic Utility Control 9

More Transactive Energy Work: Olivia Leiterman on Storage • FAPER like action need not depend on frequency • Signals from the utility system can initiate change in interchange • Energy Storage is the ultimate in ‘energy’ type load • Here is some motivation for involving real energy wiggles in ‘ancillary services’ 12/10/13 Homeostatic Utility Control 10

‘Power Signal’ could be derived from frequency Or it could be an area power error signal Separating high and low frequency signals 12/10/13 Homeostatic Utility Control 11

Separation of high frequency and low frequency variations • Energy Duration tells what your energy storage is doing • Ramp Duration tells what your other (slower) regulation resources are doing 12/10/13 Homeostatic Utility Control 12

Here is the bottom line: • More to this than can be quickly explained • Longer (slower) frequency cutoff reduces mean ramp rate for thermal units • But it also means more storage energy is required 12/10/13 Homeostatic Utility Control 13

• Frequency Adaptive Power Utility Regulator (FAPER): Fast control • The Energy Marketplace • Marketplace Interface to Customer 12/10/13 Homeostatic Utility Control 14

This is the Energy Marketplace as envisioned by Schweppe in 1980 Note ‘Utility Generation’ is in the ‘Regulated Industry’ 12/10/13 Homeostatic Utility Control 15

In The Energy Marketplace: • There would be a mix of regulated and unregulated generation • Regulation to ensure return to capital and prevent monopoly pricing • Separate ‘buy’ and ‘sell’ prices for customer generation • ‘Time of Day’ pricing was recognized as insufficient • Anticipated automation in customer premises • Recognized that there would be issues with customer acceptance and privacy • And anticipated (maybe incorrectly) that two-way communication with meters might be impractical 12/10/13 Homeostatic Utility Control 16

Later work on Spot Pricing: Jiankang Wang Here we have a pretty generic explanation of why you want to get the final price right 12/10/13 Homeostatic Utility Control 17

In a deregulated market, System operators must buy electricity and keep the system balanced Elasticity Matrix describes • instantaneous elasticity of demand • cross-elasticity (from one time period to others) 12/10/13 Homeostatic Utility Control 18

Demand is determined by a balance • Unit Commitment • Economic Dispatch • Price Elasticity predicts change in demand • But see there are numerous ways things might not balance 12/10/13 Homeostatic Utility Control 19

Better search methods can find the balance 12/10/13 Homeostatic Utility Control 20

• Frequency Adaptive Power Utility Regulator (FAPER): Fast control • The Energy Marketplace • Marketplace Interface to Customer 12/10/13 Homeostatic Utility Control 21

Market Interface to Customer (MIC) • Need to get ‘price’ or ‘prices’ to Customer • Requirement may be only 5 to 10 minutes • Anticipated that there might need to be a confirming signal in reverse • Automation at the customer premises was anticipated • Some form of ‘smart meter’ was also anticipated 12/10/13 Homeostatic Utility Control 22

Market Interface to Customer To avoid this fellow (who Scheweppe anticipated) • Smart Meter does not report on customers in real time • Measures power, multiplies by price, integrates the result 12/10/13 Homeostatic Utility Control 23

Later work at MIT: • ‘The Energy Box’ • Richard Larson and students: • Dan Livengood • Woei Ling Leow • Anticipates forecasts of price, weather, etc. will be required This is an image from Livengood’s thesis, (taken without permission) 12/10/13 Homeostatic Utility Control 24

Current work: using an office building for ancillary services (Young-Jin Kim) Solar Cells Electric Car Charging 12/10/13 Air Conditioner Modulation Homeostatic Utility Control 25

Discussion This work was presented to the IEEE Power Engineering Society Summer Meeting in Vancouver • Charles J. Frank of EPRI told us we were idiots • None of the group had any experience with running a utility • We don’t know what spinning reserve is (misuse the term) • FAPER would cost $60 k • Robert W. Alford of Siemens-Allis said that • Indirect load control would not be effective • Complicated pricing structures require too much customer participation • Confirmation of prices posted every 5 minutes would require too much bandwidth • The Electric Utility Business is not as mature as it used to be. 12/10/13 Homeostatic Utility Control 26