Research Papers

Modeling, Control, and Stability Analysis of Heterogeneous Thermostatically Controlled Load Populations Using Partial Differential Equations

[+] Author and Article Information
Azad Ghaffari

Department of Mechanical &
Aerospace Engineering,
University of California, San Diego,
La Jolla, CA 92039-0411
e-mail: aghaffari@eng.ucsd.edu

Scott Moura

Department of Civil and
Environmental Engineering,
University of California, Berkeley,
Berkeley, CA 94720
e-mail: smoura@berkeley.edu

Miroslav Krstić

Department of Mechanical &
Aerospace Engineering,
University of California, San Diego,
La Jolla, CA 92039-0411
e-mail: krstic@ucsd.edu

Contributed by the Dynamic Systems Division of ASME for publication in the JOURNAL OF DYNAMIC SYSTEMS, MEASUREMENT, AND CONTROL. Manuscript received May 4, 2014; final manuscript received June 6, 2015; published online July 27, 2015. Assoc. Editor: Umesh Vaidya.

J. Dyn. Sys., Meas., Control 137(10), 101009 (Jul 27, 2015) (9 pages) Paper No: DS-14-1197; doi: 10.1115/1.4030817 History: Received May 04, 2014

Thermostatically controlled loads (TCLs) account for more than one-third of the U.S. electricity consumption. Various techniques have been used to model TCL populations. A high-fidelity analytical model of heterogeneous TCL (HrTCL) populations is of special interest for both utility managers and customers (that facilitates the aggregate synthesis of power control in power networks). We present a deterministic hybrid partial differential equation (PDE) model which accounts for HrTCL populations and facilitates analysis of common scenarios like cold load pick up, cycling, and daily and/or seasonal temperature changes to estimate the aggregate performance of the system. The proposed technique is flexible in terms of parameter selection and ease of changing the set-point temperature and deadband width all over the TCL units. We investigate the stability of the proposed model along with presenting guidelines to maintain the numerical stability of the discretized model during computer simulations. Moreover, the proposed model is a close fit to design feedback algorithms for power control purposes. Hence, we present output- and state-feedback control algorithms, designed using the comparison principle and Lyapunov analysis, respectively. We conduct various simulations to verify the effectiveness of the proposed modeling and control techniques.

Copyright © 2015 by ASME
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Fig. 4

Population variation of on units over an infinitesimal time–temperature window. The graph is presented for a homogeneous population.

Grahic Jump Location
Fig. 3

Distribution functions of TCLs in three regions

Grahic Jump Location
Fig. 2

Characteristic of the switch modeled as a Schmitt trigger

Grahic Jump Location
Fig. 1

Equivalent electrical circuit of a TCL unit

Grahic Jump Location
Fig. 5

The method of finite difference applied for discretization of the transport PDEs to use in numerical simulations

Grahic Jump Location
Fig. 6

Normalized physical parameters

Grahic Jump Location
Fig. 10

For a constant N, normalized pole locus remains the same regardless of parameter variation as shown in Fig. 6

Grahic Jump Location
Fig. 9

TCL population uniformly distributes after transient is passed

Grahic Jump Location
Fig. 7

Variation of error between PDE and MC model versus TCL population

Grahic Jump Location
Fig. 8

System response to a step change in the set-point temperature from 24.5 °C to 24 °C with σ = 1 °C for (solid) our proposed PDE-based and (dashed) the MC model for 40,000 HrTCL units

Grahic Jump Location
Fig. 13

(Top) Hourly variation of environmental temperature in Phoenix, AZ, from July 13th 6:00 a.m. until July 14th 6:00 a.m., 2013 [18]. (Bottom) Variation of power versus time for (dashed) open-loop and (solid) closed-loop system. The customer designs the power steps according to his/her priorities.

Grahic Jump Location
Fig. 14

Evolution of the deadband and TCL population versus time. White line shows set-point evolution.

Grahic Jump Location
Fig. 15

Power outage happens from t = 1 hr to t = 1.5 hr. Power consumption for (dashed) open-loop and (solid) closed-loop system.

Grahic Jump Location
Fig. 16

Evolution of the deadband and TCL population during power control. White line shows set-point evolution.

Grahic Jump Location
Fig. 17

Static deadband and TCL population without power control. A large TCL population is turning on as power is restored.

Grahic Jump Location
Fig. 11

Reference tracking for step changes in power level

Grahic Jump Location
Fig. 12

Spatial distribution of TCL units remains practically uniform in the deadband. White line shows set-point evolution.




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In