Elsevier

Building and Environment

Volume 43, Issue 10, October 2008, Pages 1633-1646
Building and Environment

Model-based demand-limiting control of building thermal mass

https://doi.org/10.1016/j.buildenv.2007.10.009Get rights and content

Abstract

This paper describes the development and evaluation of a model-based approach for minimizing peak cooling demand using energy storage inherent in building structures. On any day where the strategy is invoked, the building is precooled with zone temperature setpoints at the low end of comfort prior to a demand-limiting (DL) period. The zone temperatures are then adjusted upwards during the demand-limiting period following a trajectory that keeps the peak cooling load below a specified target. The cooling demand target and setpoint trajectory are determined using a building model that is trained using field data. The overall approach was demonstrated for a building representative of a small commercial facility. The first step involved training the inverse model using a few weeks of hourly data. The model was then used to study the potential for peak load reduction and to determine setpoint trajectories that were implemented at the site. The demand-limiting strategy resulted in approximately 30% reductions in peak cooling loads compared to night setup control for a 5-h on-peak period of 1 PM to 6 PM.

Introduction

It is generally necessary to consider the effect of thermal storage in building structures to estimate cooling requirements for sizing equipment. In particular, concrete floors, walls, and roofs used in commercial buildings dampen the effects of heat gains to the structure on the resulting gains to the interior air, leading to reduced peak cooling loads compared to less massive structures [1]. It is possible to obtain additional reductions in peak cooling through adjustments in zone temperature setpoints within bounds of thermal comfort.

Fig. 1 qualitatively compares zone temperature setpoint and cooling load variations for three different control strategies: conventional night-setup (NS) control, load-shifting control, and demand-limiting (DL) control. With NS control, the setpoint temperature is typically maintained in the middle of the comfort range during occupied periods and set up to a high setpoint during unoccupied periods. This strategy tends to minimize the total integrated cooling load and is the most common strategy for commercial buildings. Load-shifting control uses precooling (PC) with a setpoint temperature near the lower end of the comfort range prior to the occupied, on-peak period and then resets the setpoint to a fixed setpoint near the higher end of comfort during the on-peak period. As a result of the cooled thermal mass, less heat gain occurs to the air during the hours following the rise in setpoint than occurs for NS control. As the temperature of the building thermal mass increases, the effect of heat absorption decreases resulting in an increase in cooling load. This control strategy maximizes use of stored energy in the building mass and is appropriate for minimizing on-peak period electrical energy charges. With demand-limiting control, the building is precooled prior to an on-peak or critical peak-pricing period and then setpoints are adjusted in an optimal way so that the absorbed energy into the building thermal mass is controlled and the peak cooling load is minimized. This type of strategy is appropriate for minimizing on-peak demand charges or for use in a utility program where the utility takes active control for the end-user during periods of critical demand.

As noted by Braun [2], there are four opportunities for reducing operating costs associated with load-shifting and demand-limiting strategies: (1) use of low temperature nighttime air for ventilation precooling, (2) improved mechanical cooling system efficiency due to more favorable operation at lower ambient and part-load conditions, (3) reduction in on-peak electrical energy charges, and (4) reductions in on-peak electrical demand charges. The first two opportunities lead to reduced building energy usage, whereas the second two incentives change the time variation in electrical energy requirements but can result in increased overall usage due to lower average zone temperature setpoints. However, an increase in electrical energy usage at the building can actually result in a reduction in primary energy usage at the power plant if the electrical energy is used at times of lower power plant capacity when more efficient equipment is employed. This is one of the reasons that electric utilities provide rates that encourage users to reduce electrical usage and peak demand during on-peak periods. The current paper addresses control strategies for achieving demand-limiting control and not load-shifting control.

There have been a number of studies that have considered the impact of zone temperature setpoint adjustments on load shifting and peak load reduction, including studies by Braun [2], Ruud et al. [3], Rabl and Norford [4], Reddy et al. [5], Keeney and Braun [6], [7], Braun et al. [8], and Xu et al. [9], [10]. In particular, these studies demonstrated very significant potential for peak demand reduction when setpoints were adjusted in an optimal manner. However, these papers did not present general approaches for determining setpoint trajectories that minimize peak electrical power or cooling load.

The current paper describes development of a model-based demand-limiting control strategy that utilizes an inverse building model presented by Chaturvedi and Braun [11]. The use of an inverse model trained with field data allows determination of setpoint trajectories that are near-optimal for a specific building. The overall approach was demonstrated and evaluated for a building representative of a small commercial facility. The inverse model was trained using hourly data from a previous study. The model was then used to study the potential for peak load reduction and to determine setpoint trajectories that were implemented at the site. Measured cooling loads for demand-limiting control were compared with those for night setup control to evaluate peak load reduction. In addition, an occupant comfort survey was performed. Details of the experiments and results were previously presented by Lee and Braun [12], whereas the current paper focuses on the method development and performance evaluation.

Section snippets

Description of building

The facility used in this study is the Energy Resource Station (ERS) building that houses the Iowa Energy Center (see [13] for a detailed description). The building includes a set of unoccupied test rooms that are well instrumented for comparative testing as well as occupied areas for building employees. Fig. 2 shows a schematic floor plan of the ERS building that identifies the test rooms and occupied areas. The eight test rooms are organized in identical pairs (labeled “A” and “B”) with three

Model development

Fig. 3 shows the thermal network employed for the inverse model of the ERS building that was developed from the model of Chaturvedi and Braun [11]. All test zones (west, east, south, and interior rooms) were modeled as a single zone using a single external wall, an internal wall, a roof and a floor. Predicted and measured cooling loads were for the combined test rooms at the ERS, which is representative of a small commercial building.

Each of the walls was represented with two capacitors and

Model-based demand-limiting control methodology

The load-shifting strategy used to obtain the results of Fig. 5 was not designed to maximize peak load reduction. Much greater peak load reduction is possible if the zone temperatures were varied within the comfort region rather than being held constant. Such a strategy would keep the zone temperatures low at the beginning of occupancy and raise them later in the day. This would have the effect of increasing the early occupancy loads and decreasing the late occupancy loads. This section

Approach

The trained model was used along with the demand-limiting algorithm in order to investigate the impact of different control variables on demand reduction and determine an appropriate strategy to test at the ERS building. The last day of night setup testing from 2001 (August 11) was chosen for comparison with a base case of night setup control. Every simulation was conducted for an identical set of 10 days in a row to erase the effects of initial conditions and reach a steady-periodic condition.

Test procedures

In order to validate the level of demand reduction predicted with the inverse model for demand-limiting control, testing was conducted over 14 days from August 14 to 27. The setup of the tests was identical to the setup used by Braun et al. [14] with the same internal gains, clear windows having no blinds, and with no additional internal mass. The testing began on August 14 with a conventional night setup strategy in order to establish a good baseline. The strategy was switched to the

Conclusions

A demand-limiting control methodology was developed that uses an inverse building model trained with field measurements. The methodology was applied to a building representative of small commercial applications. The inverse model was able to predict cooling loads within about 5% when trained with 14 days of data from Braun et al. [14] where precooling and night setup control strategies were applied to this facility. More importantly, the model very accurately predicted load shifting and peak

References (17)

  • T.A. Reddy et al.

    Shaving residential air-conditioning peaks by intelligent use of the building thermal mass

    Energy

    (1991)
  • Z. Zhou et al.

    A numerical and experimental investigation of ground heat transfer including edge insulation effects

    Building and Environment

    (2002)
  • M.J. Catani et al.

    Heavy building envelopes and dynamic thermal response

    Journal of the American Concrete Institute

    (1976)
  • J.E. Braun

    Reducing energy costs and peak electrical demand through optimal control of building thermal storage

    ASHRAE Transactions

    (1990)
  • M.D. Ruud et al.

    Use of building thermal mass to offset cooling loads

    ASHRAE Transactions

    (1990)
  • A. Rabl et al.

    Peak load reduction by preconditioning buildings at night

    International Journal of Energy Research

    (1991)
  • K.R. Keeney et al.

    A simplified method for determining optimal cooling control strategies for thermal storage in building mass

    International Journal of HVAC&R Research

    (1996)
  • K.R. Keeney et al.

    Application of building precooling to reduce peak cooling requirements

    ASHRAE Transactions

    (1997)
There are more references available in the full text version of this article.

Cited by (124)

View all citing articles on Scopus
View full text