Prelude Processes Peter Salamon San Diego State University - - PowerPoint PPT Presentation

prelude processes
SMART_READER_LITE
LIVE PREVIEW

Prelude Processes Peter Salamon San Diego State University - - PowerPoint PPT Presentation

Prelude Processes Peter Salamon San Diego State University Optimization At A Small Scale UCSD 2009 Questions of Finite Time Thermodynamics What is the maximum power that can be delivered by a heat engine in finite time? Given A init ,


slide-1
SLIDE 1

Prelude Processes

Peter Salamon

San Diego State University

Optimization At A Small Scale UCSD 2009

slide-2
SLIDE 2

Questions of Finite Time Thermodynamics

  • What is the maximum power that can be delivered by

a heat engine in finite time?

  • Given Ainit, Afinal, and τ , what is the minimum entropy

that must be produced in changing the state of system A from Ainit to Afinal in time τ ?

  • Given Ainit and Afinal, what is the minimum time for

changing the state of system A from Ainit to Afinal?

  • How fast can we approach T=0?

∆ Su > L2/2n

slide-3
SLIDE 3

"The Quantum Refrigerator: The quest for absolute zero",

  • Y. Rezek, P. Salamon, K.H. Hoffmann, and R. Kosloff,

Europhysics Letters, 85, 30008 (2009) "Maximum Work in Minimum Time from a Conservative Quantum System", P.Salamon, K.H. Hoffmann, Y. Rezek, and R. Kosloff,

  • Phys. Chem. Chem. Phys., 11, 1027-1032 (2009)

QuickTime and a decompressor are needed to see this picture.

slide-4
SLIDE 4

Definition

A prelude process is a reversible process performed as a prelude to a thermal process.

slide-5
SLIDE 5

Ensemble of Independent Harmonic Oscillators Sharing a Controlled Frequency ω

Cool atoms in an optical lattice. Lattice created by lasers and having an easily controlled ω .

slide-6
SLIDE 6

The Heat Engine

  • Contact with T=TH at ω = ω 1.
  • Adiabatic change from ω = ω 1 to ω = ω 2.
  • Contact with T=TC at ω = ω 2.
  • Adiabatic change from ω = ω 2 to ω = ω 1.

Controls: It’s all in the timing. Time for thermal contacts and rate at which ω changes on adiabats. ω T Rate Limiting Step

slide-7
SLIDE 7

Finite-time Third Law

The second law limits the rate of cooling

For a cycle operating between Tc and Th and exchanging heat, the net entropy production rate is For bounded this rearranges to give

slide-8
SLIDE 8

The working fluid: Quantum Harmonic Oscillator

Hamiltonian Lagrangian

C is needed to close the Lie algebra

Correlation

slide-9
SLIDE 9

Heisenberg Representation

adiabats thermal contacts with Lindblad operator

slide-10
SLIDE 10

Dynamics on Adiabats

Sudden adiabats not optimal due to quantum friction.

  • r, using sudden jumps
slide-11
SLIDE 11

Fixed omega dynamics

slide-12
SLIDE 12

Dynamics for Heat Exchange

  • Lindblad dynamics

where

Heat bath = coherence decay

slide-13
SLIDE 13

Quantum Entropy

where P=diag(ρ E) and ρ E is the density matrix in an energy basis. The Von Neumann entropy is conserved. Effective entropy in contact with the heat bath is the energy entropy

slide-14
SLIDE 14

Quantum Friction

Problem: Changing ω at a finite rate or jumping from creates “extra” entropy by increasing SE. During a heat exchange, the energy in the LC oscillation becomes heat. This is Feldmann & Kosloff’s quantum friction. ω

slide-15
SLIDE 15

Adiabatic Switching

Sets energy minimum Thermal equilibrium at L=C=0

If we change ω infinitely slowly, we can keep SVN constant.

slide-16
SLIDE 16

Optimal Control

Easier and more powerful calculus of variations. The Problem: The Tool: The Optimality Conditions:

slide-17
SLIDE 17

Optimal Adiabats

Problem: How to choose ω(t)? augment with Optimal control Hamiltonian Linear in u!!!

slide-18
SLIDE 18

Singular Control Problems

σ = switching function σ > 0; u=uMax σ < 0; u=uMin σ ≡ 0; u=? This structure usually leads to turnpike theorems.

Theorem: Optimal control of the harmonic oscillator is bang-bang. Singular branches are never used.

slide-19
SLIDE 19

Best Adiabat

ω f ω i ω t1 t2 t3 Total time on the order

  • f one oscillation !!!
slide-20
SLIDE 20

The Magic

  • Fast(est) adiabatic switching.
  • Can only extract the full maximum work

available from the change if time > min time else must create parasitic oscillations.

  • - New type of finite-time Availability
  • Time limiting branch in a heat cycle to cool

system toward T=0.

– Implies

slide-21
SLIDE 21