2003-09-25_PWETC_AgendaPacket - Laserfiche WebLink

Home

13 curves. Lamp correction factors were applied, based on tbe types of lamps actually in use at the time of measurement. Measured and calculated data were then compared to produce typical examples of in-service- depreciation. Comparison of the point-by-point grid measurements witb calculated initial design data for tbe 10 locations showed light losses ranging from 18 to 72 percent. The aver- age was 49 percent for an average system age of nearly 8 years. Depreciation for in- stallations as old as 7 years averaged 47 percent, and systems 8 to 15 years old aver- aged 54 percent. All of tbe below-luminaire readings in 10 sections were compared witb calculat~d initial illumination for this point, and depreciations of 35 to 73 percent were found. The overall average light loss was51 percent, as measured by this rough approximation method, which thus compared favorably witb tbe 49 percent average from tbe grids. In-service uniformity ratios (average-to-minimum) found in the grid measurements were worse tban the calculated initial values in all but 2 cases. The average change was 58 percent for those that became worse. At one location where grid readings were made, the nearby contributing luminaires were cleaned and new grid readings taken. The direct effect of this cleaning was to increase tbe average illumination from 0.45 to 0.50 HFC (an increase of 10 percent). However, tbe cleaning wors~ned tbe average-to- minimum uniformity ratio, from 9:1 before tt} 10:1 after. The first 10 test sections involved practices of burnout replacement of lamps ratber tban group replacement. In practically all of tbese cases, tbe only luminaire cleaning performed was at the time of lamp replacement (or pole lmockdown). In fairness to the cooperating agencies in this study, it should be reported that several were beginning programs of improved maintenance. Deprectation factors were also compared at 4 locations of equal age (4.5 to 4.7 years). Two systems had burnout replacement and luminaire washing only at burnout, and the otber 2 had group replacement at 16,000 hours (about every 4 years) plus annual wash- ing. An average depreciation of 54 percent was found for the poor maintenance systems versus only 36 percent for the group maintenance. This offers evidence in favor of group replacement programs, which are endorsed by most engineers. However, de- preciation in uniformity ratio was just as bad under one maintenance system as another. The city of Philadelphia requires its maintenance contractor to regularly test mer- cury lamps removed from service during group replacement programs. The testing involves comparison against "standard" lamps, and data are tabulated on the percentage of lumen output of tbe lamp being.tested. The date of lamp installation is placed on tbe base, and the date of removal is known. It is, therefore, possible to compare months in service witb lumen output (expressed as percentage of a 100 percent standard). Data from 804 such tests were secured from tbe contractor. Findings are given in Table 2. Wide variations can appear early in lamp service. Most group replacement is performed at a 3- to 4-year lamp life. The extremes in output of 40 to 94 percent, as compared witb tbe 78 percent average for this length of life, are equivalent to a range of -49 to +21 percent of tbe average. The range for all lamp ages, as compared with the 67 percent overall average, is an example of what might be expected in the field under a burnout replacement program; this variation was -52 to +33 percent. These laboratory findings were compared with the below-luminaire readings from each freeway route section. The field readings were expressed as a percentage of- the average of the 2 lamps at tbe "typical" test locations. This was done separately for each of the 8 routes so tbe problem of different types of luminaires would be eliminated. The extreme reading variations were tben averaged. The limits tbus found ranged from a low of -49 percent to a high of +66 percent for tbe routes with burnout replacement systems. The range was -40 to +69 percent for tbe group replacement routes. The field data show an excellent fit witb tbe laboratory findings on tbe low end of tbe scale but a poor fit on the upper end. They are sufficiently in agreement, however, to suslain tbe hypotbesis that freeway illumination usually varies about 50 percent on either side of any "nominal" value measured in the field. The :1::50 percent variation occurs under ideal conditions of constant geometry and spacing to mounting height ratios. When the actual spacing and variable-roadway width differences (such as those that occur at ramp entry or exit points) are considered, an even wider rangt exists in the hypothetical "average" HFC. ,~ '.I~ 'I 1 "1 1,1 , I I I i i .1 ,i,l , , :i "1 -!!'~':, j