I am disturbed by the inaction of the pavement, asset management, and safety communities on every level for the lack of progress on measuring this extremely important metric in an efficient, accurate and effective manner. As I presented at the RPUG in Illinois this spring, The technology is here, all we need is for the equipment providers to build it and the agencies to use it.

On another front, I believe we should migrate from Skid Resistance to Traction as the title for this metric. There is no question that Friction should not be used since the interaction between tire and pavement is dramatically more complicated than simple friction. I am suggesting Traction since it is the term utilized by the vehicle and tire manufacturers to identify the tire-pavement interface force. while most often associated with acceleration, it is also associated with braking and control.

The E-274 Trailer has been and still is a centerpiece of the Skid Resistance testing process on our nation’s highways. The basics of its design date back well over 50 years. I think this shows how solid the concept was when it was created and how well the manufacturers have maintained its essential qualities. Since Skid Resistance is such a complex metric, it is essential that we test it in manner that is as similar as possible to vehicles that rely on it. The E-274 test system comes closest to accomplishing that goal. While it does give us a very good indication of the Skid Resistance property, as we get more focused on this very important metric of our pavements, the trailer’s limitations are becoming more apparent. We now see the need for additional data in curves and near intersections. This requirement necessitates a higher sample rate than the traditional E-274 design can accommodate. Recently, this weakness has been addressed by updating the mechanical, control, and electronic systems enabling a much shorter and more rapid test sequence. A rate of twenty tests per mile is now readily achievable on a network basis and much higher test rates are possible for shorter sections. Resource limitations (water, tire wear, compressed air, etc.) now control the effective rate for network testing. The future integration of highway system data into the Skid Truck control system will allow variable rate sampling based on the network geometry. Another concern for the E-274 design is the measurement of the non-longitudinal horizontal forces on the test tire/pavement interface which are currently unmeasured yet may affect the values collected. While these forces are most probably not consequential when measuring a straight section, they could become significant when measuring around a sharp curve. Resolving this issue will require a redesign of the trailer so that the transverse forces can be measured. I anticipate that the equipment providers will be addressing this concern in the near future as highway safety moves ever so slowly from good rhetoric to actual practice.

In the effort to relate Skid Resistance to the surface properties of pavement, the measurement of macro-texture is an important step. We now have effective methods of estimating the macro-texture of a series of pavement segments even while traveling at near traffic speeds. While the probable relationship seems obvious and data supports that a positive relationship does exist, the actual relationship between high speed skid resistance and the traditional metrics for macro-texture is not the same for different pavement types and installations. I believe much of this difference can be explained by the skewness in the elevation profile. Below are two graphs that illustrate in a very stylized fashion what I am attempting to explain. While both shapes have the same Mean Profile Depth (MPD), they have very different Skews and will definitely impact the tire very differently. There are also many other metrics for describing the prevailing macro-shape of the pavement surface such as Kurtosis most of which can be calculated relatively easily and should certainly be considered as we attempt to link surface properties to skid resistance. Including Skewness in evaluating the contribution of macro-texture to skid resistance seems essential and relatively easy to do.

There have been numerous research studies that have attempted to establish an algorithm for the relationship between multiple pavement surface characteristics and skid resistance on active pavements. All the studies that I am familiar with are limited by one or more of the following circumstances:

1.       The resolution of the texture measuring equipment is insufficient for the metric(s) being analyzed.

2.       The varying impact of contamination (mostly water) that occurs in testing skid resistance.

3.       The influence of a particular feature or metric on skid resistance does not completely stop or start at a particular wavelength.

4.       The influence of a particular feature of the measured skid resistance likewise does not completely stop or begin at a particular wavelength.

5.       The only “standard” available does not separate the impact of different texture wavelengths.

6.       Profile wavelengths impact the tire-pavement interface differently at different travel speeds.

People smarter than me have established that the break point between macro and micro texture occurs at 0.5 mm. The shorter wavelengths are considered to be about aggregate shape and surface. Wavelengths longer are about aggregate size, shape and exposure. Even from this very limited description, it is clear that the shape of the aggregate can influence both macro and micro texture as they are currently defined. This problem is compounded by the reference measurement challenges. The most broadly accepted device for establishing the skid resistance of significant pavement segments is the locked rotor device described in ASTM E-274. It is a summation device measuring the total resistance supplied by all components of the pavement surface and modified by the applied water contamination. The test is only valid at a specified speed and under sliding conditions. While the ribbed tire focuses more on the micro-texture based component, all surface characteristics and test conditions still play a significant role.

Since the “Gold Standard” we have available today measures the total Sliding Skid Resistance that is a blend of all contributing factors, we must wait until there is a device that can measure the surface profile including micro-texture wavelengths before we can attempt to develop a reliable algorithm that relates pavement surface texture to the “standard” skid resistance as we measure it today. The algorithm will be very useful for comparing pavement surfaces but will only apply directly under a very limited set of conditions including, at a specific speed, with a specific tire compound and tread pattern. Initially, for algorithm development purposes, the profile measurements including micro-texture need not be at or near traffic speeds but to be usable on a network basis, they must be capable of collection at least 40 mph. Current attempts to develop algorithms that link Skid Resistance to the surface profile measurements we have available today face insurmountable challenges. I expect that the equipment community will continue moving towards the objective of ultra-high resolution and ultra-high speed surface measurements that are an order of magnitude better that anything we have today. Until then, if we are truly concerned about safety, the proactive testing of our pavements is essential especially in critical areas such as ramps and curves. The development of a high resolution skid resistance measurement device and method that can be effectively applied to a complete network should be and can be rapidly developed because the technology currently exists. To be effective, it should resemble as closely as practical, a typical vehicle travelling down the pavement and be able to reliably measure skid resistance every 0.01 mile (16 meters). It need not totally lock the wheel and will create a new metric which closely resembles the values of the E-274 method but will better relate to the vehicles currently on our highways.

It was a great pleasure to attend PE2019 in Roanoke VA this September. It was very ably hosted by VTTI and Gerardo and Edgar were exceptional conference coordinators. This conference is joined by the Road Profiler Users Group (RPUG) as its annual gathering. Most enjoyable of all, the annual gathering of pavement data collectors and analyzers is like a family reunion. We all share a common interest in moving our profession forward and are comfortable discussing our various and differing opinions on the best path to improvement. I find that we, as a group, value and care about every individual regardless of any professional disagreements that we may have. Since I no longer represent a state agency, my role has, by necessity, changed. It was encouraging for me to feel that my input was still valued and that I can effectively participate in the process of moving us forward.

I applaud the focus of VTTI and this conference on safety and in particular as it relates to pavement Skid Resistance. I believe this pavement metric is at least as important as roughness and, despite much research, is not well understood by most highway professionals. I also believe that I sense a change towards enlightenment in this area and hope to lend whatever help I can provide to move Skid Resistance to the prominence it deserves.

A big “THANK-YOU” to all who hosted dinners and to those that made me feel so AT HOME. We are indeed a FAMILY that has its disagreements but truly cares about each and every member including, I think, old goats like me.

I am looking forward to seeing many of our “family” members again at Safer Roads 2020 and at the next RPUG.

The quality of the skid resistance data currently collected has been a subject of considerable debate. This situation is exacerbated by the various types of equipment involved and the variations in the methods employed when they are used. The ASTM E-274 standard has attempted to resolve the problem by specifying the equipment design and its application, but there are still modifications in the details of its application. Even when the E-274 equipment is applied identically on the same pavement section at different times, it does not provide the same results. This fact has been used to discredit the test and as one of the reasons to discontinue or not begin network testing. In reality, the variation in results is primarily the result of changes in the pavement surface condition itself. Surely the test tires are not quite identical and testing temperature plays a role, but it is the change in surface contamination that is the primary contributor to the differences experienced. Once we understand the cause of the variation and accept that the values collected represent the pavement Skid Resistance at the time of collection, we can confidently use the data to manage the Skid Resistance property of the network. The innate variability of the Skid Resistance property should certainly be taken into account in the process of discovering sections that need attention due to low Skid Resistance. Maryland’s experience with the variability of same pavement testing using E-274 locked rotor testing is that a CV of 5% can be expected when the same sections are measured over multiple summer months and several years. The highest values were measured immediately after significant summer rain storms and the lowest after long hot, dry spells. Both asphalt and concrete sections were evaluated in this study and displayed similar results.

In summary, while we are limited to an indirect test to measure pavement Skid Resistance, the data is actionable for identifying sections that need attention for low Skid Resistance.

Pavement Skid Resistance is measured by equipment utilizing both stationary and traveling processes. The stationary processes typically sample a very small patch of the surface and therefore focus primarily on micro-texture. An exception to this limitation are the laser based systems that measure the surface contour. Due to speed, spot size and sensor size limitations, the laser based devices are currently limited to macro-texture but may well get down to micro-texture resolution in the near future when applied in stationary mode and in several years when applied in traveling mode.

The most popular traveling equipment currently used to collect pavement Skid Resistance in the United States is based on the ASTM E-274 standard and typically is used in LOCKED ROTOR mode. There are a number of other devices utilized, especially for airport runway evaluation that create a different challenge to the pavement/tire interface and measure the forces involved. A number of programs and research efforts have attempted to harmonize these different devices and the E-274 device over multiple pavement surfaces with little overall success. While the values delivered are typically proportional, their relationships vary significantly when applied to multiple pavement types. The measurement of pavement Skid Resistance is clearly a work-in-progress.

Clearly the preferred method to measure the pavement’s contribution to Skid Resistance would be to measure directly the applicable range of pavement contour and relate it to the Skid Resistance provided to a typical vehicle on the highway. This process will need to be developed after a reliable and accurate method of measuring micro-texture on a large scale is developed and the research is completed that relates the various aspects of pavement contour to skid resistance. This appears to be a decade or probably more in the future.

We are therefore left with measuring the Pavement Skid Resistance by applying a challenge to the tire/pavement interface in a controlled manner and preferably in a very similar fashion to the typical vehicle on the highway. When the E-274 standard was developed, anti-lock braking systems were not on the horizon much less deployed. Now, they are standard on the vast majority of vehicles travelling our roads. I believe we need to modify the E-274 system to measure the skid force prior to full lock-up. I have proposed averaging the measured force between 33% and 66% lock-up. We need to begin the measurement beyond 20% lock-up to minimize the impact of the construction of the measurement tire. At 40 mph, today’s equipment can typically collect 10 or more data points in that range and average them for a Skid Resistance value which I suggest we call SSN (Slip Skid Number). Several thousand miles of pavement have been collected in this fashion and the SSN was typically proportional to, and roughly 125% of the SN (locked rotor} value with an R squared above 0.99 across all pavements. However, on an individual pavement by pavement basis, the relationship varies by over 10%. I believe this difference reflects the fact that this test is impacted by slightly different aspects of the pavement contour. If we agree that this is the case, the SSN value should better relate the pavement’s performance to the needs of current vehicles. Other advantages of collecting the SSN is that it is a very rapid test and it that uses much less resources than the locked-rotor test. As often happens when new technology is employed, the difficulty will be in relating the SSN to SN values collected in the past. In addition, significant modification needs to be made to both the hardware and software of an E-274 device to collect and report the SSN, but the first versions have been produced and are working. Future improvements to the E-274 device should include a trailer mounted, computer controlled, water delivery system and an electromagnetic or hydraulic braking system eliminating the use of pads and rotor.

Another potential area for improvement in testing for Skid Resistance is adjustable test intervals based on the length of anticipated homogeneous pavement surface. Currently, the equipment tests on a fixed interval (unless manually over-ridden) which is set at the start of the road to be tested, The sections at ramps, curves, and intersections can have quite different SN or SSN values than the remainder of the pavement. Utilizing the network database, these sections can be identified and the information transferred with the test requirements. The appropriate test interval can be set and transferred to the collection equipment so that the test resources can be deployed in an optimum manner. Test intervals could be varied from every 0.3 miles for long straight stretches of pavement down to every 0.01 miles.for critical sections.

In summary, the available equipment for measuring Skid Resistance is functional, but there is much room for improvement and for new technological development. There is also much room for a better understanding of the traction challenges faced by today’s pavement users and how the various characteristics of the surface contour impact the resultant Skid Resistance.

The VALUE impact of our projects or programs is not something we routinely think about as a measurable metric. We just know that we are “doing good things” and tend to focus our daily efforts on reducing the cost side of the value vs cost relationship because cost is something upon which we can typically have some direct impact.

Value, however, is something we should have at least a sense of its magnitude or we will potentially ignore processes and programs that can positively impact the quality of life we provide our ultimate customers, the users of the highways. Below is a list of potential areas in which we can have influence on their quality of life:

• Life itself

• Permanent Disability

• Personal Injury

• Vehicular Damage

• Infrastructure Damage

• First Responder Costs and Risks

• Congestion Costs and Risks

• Hazardous Materials Risks

• Vehicular Wear

• Cargo Damage

• Passenger Comfort

From this list, lets extract a few as examples that we can assign values for analytical purposes.

First, while we would all generally agree that a human life is priceless, many entities have given it a dollar value so that analyses can be performed. The value ranges from 5 to 10 million dollars. The EPA assigns 9.1 million, while the federal Department of Transportation assigns roughly 6 million. The value given to permanent partial disability varies widely based on severity and other variables but averages at roughly a tenth of the life value. Congestion is said to have cost Americans 305 Billion dollars last year with the vast majority due to insufficient peak network capacity, but a significant portion was due to maintenance imposed reduced capacity and traffic incidents.

An example of where understanding the value of an activity can impact our decisions is in testing for skid resistance. Many agencies only test reactively when there is an elevated accident rate in a particular section of highway. Others test proactively collecting network skid resistance data on a routine basis. In deciding the approach to take, each agency should be considering that one in every 140 accidents is fatal and, on average, a single accident has a direct cost of $7,500 that needs to be combined with the associated costs (First Responders, Congestion etc.) to determine total impact on the user’s quality of life. Should we be using accidents as the skid resistance measurement method to locate substandard sections when nondestructive methods are available?

Another case where a value analysis should come into play is when we consider the development and implementation of measurement technologies that minimize the impact on the traveling public. These programs can be quite expensive since they typically involve developing both new collection and analysis methods. These costs, however, can be fully justified when the value associated with the process is considered.

I suggest that we keep investing to find better ways to provide the optimum value - cost relationship available for those that use our highways.

Skid Resistance vs Friction

We often find ourselves talking about pavement friction in the pavement management and design arenas because, I believe, pavement skid resistance is so similar in concept to the frictional properties of two materials in contact. For example, the lateral force encountered is proportional to the perpendicular (normal) force (I am not convinced it is a linear proportionality in this case) and the sliding value is typically less than the static (rolling) one. What we typically measure, however, is something quite different from traditional friction. We attempt to measure the lateral force relationship between tire and pavement much as a typical user would encounter it during use. The force we measure relates to surface properties that are much larger than the molecular sized features associated with traditional friction. The lateral force we are measuring are considered to be primarily related to the micro and macro texture features of the two surfaces and the material properties of the test tire employed. To complicate matters further, we apply a surface lubricant (water) to the interface during measurement that significantly impacts the interaction between the two surfaces and the resultant measured force.

It is for these reasons that I suggest we, as a community, agree to refer to this pavement metric as Skid Resistance and the scale for the measured property , when measured by a locked wheel testing device, as Skid Number. I believe that other, non-locking type measurement equipment should also utilize the term Skid rather than Friction in the name of their output value since many of the same concerns apply to these alternate measurement devices.

The first issue I would like to bring up is the matter of Skid Resistance Measurement.

The some key issues that come to mind are:

• What’s the value and what’s the cost?

• Does the current collection equipment meet the need?

• Is the information derived from the collected data actionable?

• What’s the value of proactive measurement vs reactive measurement and what is the relative cost?

• What is the required accuracy and resolution for the data to be informative and actionable?

As you can see from this list, there is much to delve into relative to this pavement metric alone.

In the coming weeks, I will attempt to present my thoughts on each of the issues above, not as golden solutions, but as jumping off points for in-depth discussion that will improve both the state-of-the-art and the state-of-the-practice.

The plan for this blog is to present (on an irregular basis) my thoughts on current activities in our field (or the lack thereof) and on recent accomplishments that have positively impacted the value/cost relationship that we all strive to improve.

I hope that they may both, to some degree, inform and primarily inspire thought and conversation on the issues presented.