INTERNATIONAL ROUGHNESS INDEX

Will it work for project acceptance?

One of the struggles we have in our path to smoother pavements relate to ride measurement. This is a struggle because any measuring system should then relate to actual seat-of-the-pants detection of ride comfort, or the lack of it. Current practice in North Dakota and just about all other states is to use the Profile Index (PI) for project acceptance. However, 15 states are in various stages of converting to IRI, which is also the system normally used in Pavement Management Systems (PMS). North Dakota has even dabbled with the use of IRI for measurement for ride incentive payments. However, they have fortunately dropped its use and gone back to the PI system.

The PI system uses a 25 foot long California Profilagraph to measure surface deformations. The concept relates to measuring the deviations from a 25 foot long rolling straight-edge, and then summing the height of the valleys and bumps over a given distance along the roadway after making allowances for a blanking band. The result is the sum of the variations in an inches per mile format. The California Profilagraph has been in use for most of the 1990's and is the best understood measurement system by contractors and engineers.

Even though this system is prevalent, no one believes that it is the best system to reflect the ride felt by a traveling public. However, no other systems (such as IRI) are currently available which allows a proven measuring device to get on the new pavement the next day to accurately and repetitively check on smoothness characteristics. This is important from the contractor perspective, because he needs almost instant information on the status of pavement smoothness. Mix proportion adjustments, paver set-up, or other items a contractor can control are beyond manipulation, if information on pavement ride measurement is not readily available (or understandable) from the previous days production.

IRI has been around for quite some time. It is a mathematical representation of ride measurements. Data is gathered from laser sensors combined with accelerometer data to plot the road profile. Then, a 25 foot straight-edge model is used to produce a PI. Applying multiplication factors up to 1.8 to critical repetitive bump wavelengths, the mathematical model is generated representing IRI in an inches-per-mile format. The critical bump wave-lengths have been known in theory for quite some time. One critical wave-length relates to the response of a nominal vehicle to the road which puts the body's organs in an uncomfortable motion up and down. Another wave-length relates to the general characteristics of a wide variety of automobile response to suspension system overload. At these wave-lengths, the tire bounce becomes over-active, and the car experiences extreme vibration. However, even with IRI, there is controversy as a measuring system, depending on using half car or quarter car data sensing systems, which yield different results. In any case, the mathematical modeling present with IRI seems like it has potential to solve the problem of what the passenger feels in a more optimal way than PI.

The Pavetech/Pathways vans used for pavement management have many years of data on highway smoothness. The FHWA apparently mandates data be collected in IRI for PMS. But these vans are too heavy to drive fresh concrete, and usually unavailable anyway, as their cost is high, so each DOT can only own one of them. Therefore, the vans cannot serve the role of quick response measurement devices.

The advent of the Light Weight Profiler (LWP) was supposed to bring use into the early and quick mode of ride measurement, including IRI. These ATV's, with computers and laser sensors, come in many different variations from different companies and seem quite affordable. But only a handful of states have approved them for use in ride measurement. Calibration from one machine type to another has been impossible to reconcile thus far. Experience has shown that each different machine company uses different laser beam width, sampling rate and software algorithms. It is no mystery why calibration is difficult.

So, while these LWP's currently have some shortcomings, work in reconciling the calibration issue is moving ahead. Thanks to funding by ACPA National and certain state ACPA Chapters, Steve Karamihas from the University of Michigan is in the process of providing co-relation work. At Steve's invitation, 18 different LWP machines including the four major manufacturer's together with the full-scale vans were tested at two locations. Over a two-day period, the GM Automotive Test Track and a Michigan highway construction project were very busy with calibration runs. With the data in his office, Steve is now on a fast-track to evaluate the findings and make recommendations. Hopefully, cures may be as simple as a change in a computer program to meet the calibration requirements. However, if that won't work, it will be up to the respective company to secure the necessary changes to meet the "soon famous" Michigan calibration test.

No, IRI will not currently be useful for project acceptance measurements. However, when recommendations from the Michigan test are implemented, the LWP will be the key to this. Calibration fixes should be in place by the next construction season; good timing for a certain 15 states, and likely, for our North Dakota interests, also.