Estimation of Design Hourly Volume

Introduction to Design hourly volume

A real life example

In the construction season, Washington Avenue Bridge is under mantaince–again. Because the government believe the bridge is too narrow that congestion happens very often, so it needs to be expanded. The questoin is, how width the bridge should be?

Estimated traffic demand (volume) → Capacity of one lane → Bridge (road) width

Ideally, we want the road to accommodate traffic demand during most periods without experiencing serious congestion.

Design hourly volume (DHV)

The traffic volume that the road is designed to accommodate during a representative high-demand hour.

Estimation of Design hourly volume

How do we find a volume that can “accommodate during a representative high-demand hour” ?

Will the highest demand hour volume a good choice for the Design hourly volume?

The highest hourly volume may occur only once or during an unusual event. Designing the entire roadway to accommodate such an extreme condition could therefore be uneconomical.

If the highest demand hour volume is not appropriate, what would be proper one? 80% of the highest hour volume? 50% of the highest hour volume? 30% of the highest hour volume?

Rather than selecting an arbitrary percentage, a reasonable approach is to examine how hourly traffic volumes are distributed throughout the year.

In early 1950s, L. E. Peabody and O. K. Normann plotted ranked hourly volumes for hundreds of permanent automatic traffic recorders across the United States，and they found that: The hourly volumes change rapidly for the highest 30 hours, but change much more gradually after the highest 50 hours.

In addition, designing above the 30th hour yields severe diminishing economic returns, while designing below the 50th hour causes disproportionate delays. So they settled the 30th highest hour as the Design hourly volume.

Furthermore, the 30th-highest hourly volume is approximately at the 99.7th percentile of annual hourly traffic volumes, indicating that the demand is under this level during 99.6% of the hours in a year. Sounds good?

Recent research on Design hourly volume

Data availability: Geographical, climatic, economical, socio-demographical data, etc.

New approach: statistical models, machine learning models.