Performance of Retrofit Optical Reflectors

Jeffrey Kessel (retired)
University of California
Physical Plant-Campus Services
Energy And Engineering Services
Berkeley, CA 94720

January 10, 1991

Related Reports:

Managing a Large Lighting Retrofit Project

Field Experience with High-Frequency Ballasts

ABSTRACT

The illuminance distributions of two typical fluorescent luminaires, a 2F40 and a 4F40, were measured before and after removal of one-half the lamps and stallation of various specular optical reflectors. The modified fixtures' fluxes were calculated by area-weighting the measured illuminance distributions, and were found to be 59%-73% of that of the unmodified fixture. This was accompanied by a 16%-19% decrease in the distance perpendicular to luminaire axis at which the illuminance dropped to one-half of its value directly under the fixture. These results were used to predict the life-cycle costs of the candidate reflectors when used with commercially available partial light-output ballasts.

INTRODUCTION

The installation of a specular optical reflector into a fluorescent luminaire, accompanied by removal of some lamps and ballast, has gained acceptance as a means to decrease lighting energy use by nearly one-half, without losing a proportional amount of useful light.[1] Other investigators [2,3] have found that illuminance levels decreased to 58% - 65% of original levels after similar modifications. These results agree with photometric data from one luminaire manufacturer [4] who offers several luminaires in two versions, differing only in the reflector material. The CU tables show an increase of 11% - 26% in the CU values of the specular fixture relative to the new fixture with standard white diffuse reflector. Other manufacturers offer new fixtures with specular reflectors [5]. Because specular reflectors are incorporated into the design from the beginning, rather than as a retrofit add-on, we may expect to see still higher efficiencies and greater uniformity of illuminance.

As part of the contracting procedure in the Lighting Modification Program of the Department of Facilities Management at the University of California at Berkeley, we carried out a procedure to quantify this effect. In order to show the range of possibilities for this type of retrofit, we selected two different luminaire types in the same building for the performance test. Four reflector manufacturers produced for testing prototypes encompassing design and material of their choice, and submitted bids to supply the reflectors. In order to encourage use of superior design and materials we incorporated into our request for bids a procedure to give performance-based credits to the bids of the better performing reflectors.

Our results should be characteristic of the performance obtainable from retrofit specular reflectors designed under competitive conditions. Two fixtures, (Type A, a 2'x4' 4F40 with a wrap-around clear prismatic pattern #12 lens, and Type B, a 1'x4' surface-mounted 2F40 with a flat clear prismatic pattern #12 lens), were loaned in the summer of 1987 to four reflector contractors for design of prototype reflectors, to be submitted with a bid to supply and install 1550 and 1000 units, respectively. The request for bid specified the manner in which a credit, proportional to performance, would be applied to the bids. The reflectors produced were of different design and materials, and are representative of typical commercially available products. The fixture and reflector characteristics are listed in Table 1.

Table 1. Optical Reflector Performance Measured at the University of California, Berkeley, Summer,1987

Notes to Table 1:

• Type A is a 2' x 4' pendant-mounted 4F40 with wraparound prismatic lens.

• Type B is a 1' x 4' 2F40 surface-mounted fixture with flat prismatic lens.

• Partial or Full Coverage refers to the size of the reflector relative to the size of the fixture. Relocate indicates that the original socket holders were removed and replaced with relocated sockets.

• Anod. Alum. is high quality anodized aluminum. DEO indicates the addition of a dielectric overcoat to increase reflectance. Silv. Laminate refers to a coating of elemental silver deposited on plastic sheet which is adhered to aluminum substrate.

• The Bid prices quoted were to clean the fixture and to supply and install the reflector (and socket mounting brackets, if any), to supply and install lamps, and to install a ballast supplied by the University. The Adjusted Bid is the performance-adjusted bid price as explained below in the section on Reflector Selection. R=% Base Flux = % of flux from original fixture (after being cleaned and relamped).

PROCEDURE

The test procedure was similar for categoriesA and B. Both test luminaires were over twenty years old, and had rarely been cleaned. A classroom measuring 26.5 ft. x 37.0 ft. x 14.7 ft. high, with fixtures suspended at 9.0 ft. height, was selected for the CategoryA test. An office measuring 17.0 ft. x 21.0 ft. x 9.8 ft. high, with fixtures suspended at 9.5 ft. height, was selected for the Category B test.

First a fixture located near the center of the test room was thoroughly cleaned, and fitted with a full complement of 40 w. F40/CW lamps which had been operated under controlled conditions in a lamp-life test rack for 4000 hours. The fixture's 277 v. ballast was removed and replaced with energy-efficient 120 v. core & coil ballast (Advance R-2S40-1-TP for category A, and Universal 412-L-SLH-TC-P for category B) operating initially the full lamp complement, and subsequently one half the lamps in the modified fixture. The test ballast was powered by a line voltage regulated power supply to ensure that illuminance measurements of modified fixtures would be obtained at the same power level.

Power consumption was monitored with a TIF 2000 Wattprobe, and found to remain within a 2% range during illuminance measurements on the test grid by means of a Tektronix J65 illuminance probe. All luminaires in the room were switched off except for the test fixture. Air temperature near the test fixture was monitored with a Fluke Y8103 Type K thermocouple, and building fans were switched off in a successful effort to maintain the room at constant temperature (20.5 C).

A test grid of twelve points, 2.5 ft. above the floor, was laid out in one 8.75 ft. x 8.75 ft. quadrant of the luminaire. Figure 1 shows the grid for the two tests. Each reflector contractor in turn installed his reflector, most choosing to relocate lamp sockets as part of their design. The test lamps and the sockets were marked to ensure that each reflector was tested with the same lamps installed with the same orientation (end-for-end and rotationally about lamp axis). There was, of course, some lateral displacement of the lamps depending upon the contractor's design location for the lamp sockets. After waiting 10 minutes until the power consumption stabilized, illuminance measurements were taken on the grid.

RESULTS

The measured power decreased, for the category A luminaire, from 163 w. (unmodified, 4 lamps) to 91 w. (modified, 2 lamps). The modified category B luminaire operated at 47 w. (with one lamp). The unmodified category B luminaire's power was not measured because it was powered by its original 277 v. 2F40 ballast during the baseline (unmodified) illuminance measurements.

In Table 1 the photometric results are summarized by stating R, the percent of unmodified (base) luminaire flux delivered by the modified luminaire. This ranged from 59%-71% for Category A, and from 61%-73% for Category B. Figure 2 and Figure 3 plot the illuminance along a line perpendicular to the luminaires' major axis. The decreased lateral light distribution attributable to the reflectors is indicated by a 16%-19% decrease in the distance at which the illuminance drops to one-half its maximum value (as indicated by the arrows labelled 1/2 Max).

REFLECTOR SELECTION

Rather than award the contract to the low bidder, the request for bids specified award to the low bidder after a downward adjustment that was proportional to the relative performance of the reflectors. This adjustment factor is derived as follows:

Let R be the ratio of flux delivered to four quadrants by the modified luminaire relative to the unmodified luminaire: (delamped & reflectorized)/(fully-lamped basecase).

Consider two reflectorized fixtures, 1 and 2, with ratios R1 < R2, (R calculated relative to the fully-lamped basecase). By using commercially available [6] "tuned" partial light output ballasts we can operate the poorer performing reflector at full power Power1 = W, and the better performing fixture 2 at partial power Power2, so that its resultant ratio R2 equals that of fixture 1 operated at full power. That is, decrease Power2 until Power1 / Power2 = W/Power2 = R2/R1.

Because the ballast used with fixture 1 consumes W watts, we have:

Power1 - Power2 = W * ( 1 - (R1/R2) ).

Therefore the lifecycle avoided energy cost of retrofit fixture 2 relative to retrofit fixture 1 is, per fixture:

[1-(R1/R2)]*(W watts)*(0.001 KWH/watt)*(H hrs/yr)*(E $/KWH)*[P/A] ,

where W=ballast power, H=annual operating hours, E= current cost of electricity, and [P/A] is the present worth factor encompassing anticipated lifetime, discount rate, and energy cost escalation rate.

In our contract award, this avoided energy cost factor was used to adjust downward the bids of the better performing reflectors. Thus all retrofit fixtures were evaluated as if they operated with partial power so as to have the same illuminance ratio as the worst performing reflector (which itself in effect receives an adjustment of $0.00). The rationale for this adjustment is that we select as candidate spaces for reflector retrofit those spaces calculated (or measured) to have a factor of excess illuminance which we anticipate will be "corrected" by delamping and installation of a poorer performingreflector. The extra illuminance provided by a better performing reflector may be captured as energy savings through the use of partial light output ballast, or may be allowed to remain as an amenity with the use of full power ballast. These adjustments ranged from $10.55-22.82 for category A (W=60, H=3000, E=0.075, and [P/A]=10), and from $5.54-$14.79 for category B (W=30, H=4000, E=0.075, and P/A=10). The adjusted bids are shown as the last entry for each reflector in Table 1.

CONCLUSIONS

The removal of one-half the lamps combined with the installation of specular optical reflectors as a retrofit modification for fluorescent luminaires resulted in useful delivered flux in the range of 60% - 73% of that obtained with a cleaned older fully-lamped luminaire. This performance range is a function of reflector design and material. Lighting energy use can be approximately halved in spaces that can tolerate both a decrease in illuminance of this magnitude, and a decrease in uniformity characterized by a decrease of around 16% in the half-peak spacing. This range in reflector performance can be quantified and used as input to a life-cycle cost analysis which in turn can be used to adjust the bid prices for award of a contract.

ACKNOWLEDGEMENTS

The author wishes to thank Robert Clear and Rudy Verderber of the Lighting Systems Research Group at Lawrence Berkeley Laboratory for their encouragement and helpful suggestions and comments.

REFERENCES AND FOOTNOTES

1) Chester K. Johnston, Lighting Energy Management - With Reflectors, Facilities Manager, Vol. 1, No. 4, Winter 1985

2) D. L. DiLaura and D.G. Kambich, Luminaire Retrofit Performance - Commercial Building Lighting Systems, March, 1987,
EPRI Report EM-5094, Palo Alto, CA 94304

3) T.K. McGowan, Fluorescent Fixture Reflector Inserts, General Electric Technical Information Series, No.4162-871A, January 22, 1987.

4) Fixtures 101, 103, 325 are offered with and without a silver-laminate specular reflector. Wellmade Metal Products, Oakland, CA.

5) Brayer Lighting, San Francisco, CA ; Maximum Technology, Brisbane, CA ; Wellmade Metal Products, Oakland, CA

6) Electronic Ballast Technology, Inc., Torrance, CA

FIGURES

FIGURE 1. PHOTOMETER LOCATIONS

FIGURE 2. ILLUMINANCE ALONG PERPENDICULAR TO FIXTURE MIDLINE

FIGURE 3. ILLUMINANCE ALONG PERPENDICULAR TO FIXTURE MIDLINE