for copies of this report, go to:

http://www.dot.state.oh.us/divplan/research

or call 614-644-8173

Ohio Department of Transportation

1980 West Broad Street

Columbus OH 43223

Project Description

Problem

Maintaining traffic signals is a significant expense at ODOT. Bulbs fail and need immediate replacement, which entails the expense of traffic control and may involve hazards such as heavy traffic or inclement weather conditions.

Two possible technologies to reduce this problem are improved load switches and group relamping.

The improved CTCLS series load switches were developed by Conservation Load Switch (CLS) of Westerville, and they were designed to extend the life of traffic signal bulbs by a factor of two to four. This was achieved by gradually “soft-starting” the bulb over 45 milliseconds rather than sending full power immediately and by using a microprocessor circuit to regulate voltage fluctuations. The CTCLS series load switches had been tested in laboratories and in very small scale field tests, but a large-scale test was needed to verify the manufacturer’s claims of doubling bulb life.

Group relamping is the process of replacing all traffic signal bulbs at a fixed interval, rather than simply waiting for them to fail before replacement. Most of the useful life can be obtained from the bulbs, while the cost of changing bulbs is reduced by replacing them all at once instead of making several spot replacement trips. The occasional failures that do occur before the group relamping interval is up are replaced immediately, but these are minimized because bulbs are not kept to the point where failures become common.

Objectives

The original objective was to assess the effect of the CTCLS load switches on bulb life in the field on four different major brands of bulbs as used in six states in varied climate regions in the United States.

The final objectives were to determine the best brand of bulb for use with conventional load switches and to develop a group relamping

model that could be used to determine the group relamping period that would minimize traffic signal maintenance costs.

Description

Approximately 144 intersections in each state were selected for the test, half to use CTCLS load switches and half to use conventional switches. Each state used a different brand of bulb as follows: Ohio (GE-Real), Mississippi (Sylvania), Wisconsin (Sylvania), Michigan (GE-Phillips), Waterloo IA (Duro-test). In each state there were at least 1728 bulbs operated by conventional load switches and a similar number operated by CTCLS load switches.

After installation, a number of the CTCLS load switches gave dual indications or other malfunctions in the field. It was eventually decided by the lead state (Ohio) that the CTCLS switches presented a safety and liability hazard, and therefore they were removed from the study. The study then converted to a pure bulb life study with about 3456 bulbs in each state operated by conventional load switches.

Whenever a bulb failed, it was to be tagged by maintenance personnel and then sent to the study subcontractor, the Operator Performance Laboratory at the University of Iowa, for cataloging and analysis. A spreadsheet cataloging the failed bulbs with survival curves was maintained.

Additionally, another spreadsheet was developed at ORITE to model group relamping. The spreadsheet would take as input various costs, distances involved in emergency spot and group relamping, and bulb survival rates (taken from existing study records for the example in the report). The output would be a table of annual cost of group relamping, averaged over a ten year period, as a function of group relamping interval in the range of 1 to 24 months. The optimum group relamping interval can then be found by selecting the interval with the minimum annual cost. A sensitivity analysis was conducted by modifying the values of the input parameters to see how the output changed.

Conclusions & Recommendations

The CLS claims that their CTCLS series load switches extended bulb life and reduced maintenance costs could not be investigated since the switches were removed early in the study.

Overall, the bulb life portion of the study produced little useful data. In all cases, the states and cities tagged and returned at most a fraction of the failed bulbs. It appears that the extra work caused by the study may have been asking too much of already busy maintenance personnel. Thus it seems that even well designed studies that rely on extra efforts by field personnel over a long period of time are at best difficult to carry out.

The Duro-test bulb manufacturer went bankrupt after the study was underway, making the results from Iowa at best academic, since the bulb would be no longer available for use by state DOTs.

Ohio used GE-Real bulbs, which had an extremely poor field performance, with nearly 4% (3.99%) failing by the end of the first year, nearly twice ODOT’s 2% maximum failure amount after one year. It can be concluded that the GE-Real bulb is not satisfactory and should not be used in Ohio. Nearly 30% (29.7%) of the bulbs failed at the end of two years.

In Jackson, Mississippi, a similar result was obtained with Sylvania bulbs. According to official records, the 2% first-year failure limit was crossed at no more than ten months into the study. The two-year failure rate for the Sylvania bulbs was 10.5%

The group relamping model was evaluated, using Ohio District 4 as a test case for a limited sensitivity analysis. The model shows a relatively shallow minimum at about 10 months, providing a minimum annual cost of $38,719 to the district for maintaining their 273 signalized intersections.

Implementation Potential

Group relamping spreadsheet esults will vary somewhat in other districts or states, but the software can be applied to those cases as well. The greatest difficulty involved with the software is the need to input the emergency spot relamping distances from the central office or garage to each signal, to cluster the signals into groups that can be group relamped in a single day, to input the distances to each cluster, and to input the distances between the signalized intersections within each cluster.

The software will provide a relamping interval that minimizes overall cost, and an estimate of that cost on an annual basis based on group relamping intervals from 1 to 24 months. Because of the relatively shallow minimum in the cost as a function of relamping interval, it may be advantageous to use a slightly longer group relamping interval at a slightly greater cost, for instance an annual relamping that may fit in better with other existing maintenance tasks, such as inspecting suspension wires holding up traffic signals.