This past construction season, several new technologies for quality control of hot-mix asphalt (HMA) were tested on a project in Alabama. The project was a first look at using automated sampling and testing of the materials before and after mixing at an HMA plant. The Alabama Department of Transportation (ALDOT) and FHWA funded the project, which included the purchase and installation of the automated devices, and contracted with the National Center for Asphalt Technology (NCAT) to conduct the evaluation.
The new technologies evaluated on the project included automated belt samplers, moisture content probes, an automated drying and gradation device, an in-line binder viscometer and a robotic truck sampler. Several of the devices were first-generation equipment.
The goal of this project was to evaluate the technical functionality of these devices and see if the information gathered was useful in the monitoring of the HMA production process. In the future, it is hoped that such technologies will help provide better control of mix production without added technician manpower.
A quantity of quality
The new equipment was installed on East Alabama Paving Co.'s asphalt plant in Opelika, Ala. Evaluations were conducted primarily during the spring and summer of 2005 and included the production of three different types of HMA.
Automated belt samplers were used to obtain samples of aggregate or RAP from moving conveyor belts. When a belt sampler is activated, an open box rapidly sweeps transversely across the belt, closely following the contour of the belt so that all of the material in the cross section is removed. The speed of the sweep is very fast to obtain an even cross section of material and minimize the potential influence on the plant's belt scales. Belt samplers have been used by other industries, particularly the mining industry, for several decades, so this technology is mature and the equipment is robust enough for the HMA industry.
On this project, the samples obtained by the belt sampler went straight into an automated drying unit. As an option, the sample could have been deposited into a prepared container for later testing. The mass of the sample obtained by the automated belt sampler depends on the amount of material on the belt and the size (width) of the box. Typically, the sample size obtained by the belt sampler was 20 to 30 lb. For this project, belt samplers were installed on the virgin aggregate conveyor and the RAP conveyor. Although a formal analysis of the belt samplers was not conducted, a video appears to confirm that the materials on the belt were completely removed during the sweeps.
For continuous mix plants, moisture contents of the aggregate and RAP must be determined in order to correct the mass measurement (i.e., tons/hour) of the belt scales. Two technologies were evaluated on this project to determine moisture content. The first technology utilized probes that were inserted into the stream of material traveling on the belt.
These probes use microwave-based technology which instantaneously senses the microwave energy absorbed by the material. The energy absorbed is proportional to the moisture content of that material. This technology has been used in several other manufacturing applications, including the ready-mix concrete industry. For this project, moisture probes were installed on the aggregate conveyor belt and the RAP belt.
Although the moisture probes were easy to set up and operate, unique calibration factors are necessary for each material. Consequently, each mix requires a different calibration, which is a time-consuming process. However, due to mix production scheduling issues, the instruments were not recalibrated for each mix. Therefore, the moisture content results obtained with the microwave probes show a consistent offset from the measured moisture contents determined with laboratory tests. This problem can be corrected by establishing and following a calibration process for each mix.
The second method for determining moisture content was with automated sample driers. The driers used on this project were first production units. The driers use electric heating elements to heat the sample to a temperature around 218ºC (425ºF) until the sample reaches a constant mass. The drying units were suspended on load cells so that the sample mass could be monitored by a programmable logic controller (PLC) and the moisture content of the sample automatically calculated. Drying times for a 20- to 30-lb sample were in the range of 30 to 100 minutes, depending on how wet the sample was at the start of the test.
Moisture content results with the automated driers were much more erratic. There were a variety of issues with these driers that apparently resulted from the fact that they were first production models, and these issues may have contributed to the erratic results.
There were programming issues dealing with time intervals between mass measurements, heating element problems and probably some wind effects on the chamber. There also is a question about comparing results of the automatic driers, which heated samples to above 200°C, and the laboratory method, which uses a much lower temperature. All of these issues should be easy to resolve in future models.
Another new automated device evaluated on the project was an aggregate gradation unit. After aggregate samples were dried by the drying unit, they were directed into an automatic gradation device. The gradation device was similar to laboratory sieving equipment and was equipped with six standard sieve screens (12.5 mm, 9.5 mm, 4.75 mm, 2.36 mm, 1.18 mm and 0.3 mm). Other sieve screens can be used.
Shaking of the screens was accomplished with variable-frequency vibrators. After vibrating for a programmed interval, the screening unit was rotated 90° and each screen was emptied, one at a time, into a catch pan. The catch pan was suspended on three small load cells connected to a PLC which calculated the gradation as percent passing each sieve. The gradation unit used on this project was one of the first built for use at an asphalt plant. Similar gradation units have been installed on aggregate crushing plants.
Overall, the automated gradation device provided results similar to the laboratory dry gradation test results on the same samples. Eighty-two percent of the results from the automatic gradation unit were within the multilab range of acceptable results from the national standard gradation procedure. This repeatability needs to be improved. Some of the error may have been due to incompletely dried samples from the automatic drier unit. The observed differences between the automated gradation unit results and lab tests did not indicate any trends with specific sizes, and an inspection of the automatic gradation unit did not reveal any problems with the screens.
A viscometer installed in the asphalt supply line from the plant's tanks to the point of mixing in the drum was used to indicate if the correct binder grade (e.g., PG 67-22 or PG 76-22) was being used in the mix being produced. The in-line viscometer uses a magnetically excited vibrating rod in the flow of the asphalt binder. The dampening effect of the asphalt on the vibrating rod is proportional to the viscosity of the asphalt. To compensate for the effect of temperature on the viscosity of asphalt, the instrument also measures the temperature of the binder and a PLC interface corrects the viscosity to a standard reference temperature. For this project, the reference temperature was set at 135°C (275°F).
In-line viscometers are used in other industries to monitor and control chemical mixing processes and to optimize fuel viscosities for large combustion engines and burners. The measurements of the in-line viscometer were successful in differentiating the unmodified PG 67-22 binders from the polymer-modified PG 76-22. Although more testing is needed to evaluate other grades and sources of asphalt, the rapid viscosity check provided by this instrument appears to provide excellent process control information.
A standard item on this HMA plant was an infrared temperature sensor located at the discharge of the drum where the mix drops into the slat conveyor. Mix temperature was continuously monitored using this sensor as a plant process control function to provide information to the plant operator for mix startup and burner adjustments. Most HMA plants are equipped with this type of sensor or another type of mix temperature sensor. The infrared temperature sensor at the point of discharge from the drum was found to be reasonably accurate.
Another location for automated monitoring of mix temperature for quality control would be the point of discharge from silos or surge bins when loading haul trucks. This location would provide a better sample of mix temperature, compared with sticking a stem-type thermometer into the side of the truck bed or shooting the top of the load with a hand-held infrared temperature gun.
A robotic mix sampler for retrieving HMA samples from haul trucks also was evaluated on this project and compared with two other sampling methods. One was traditional haul truck sampling by a technician from a sampling stand using a shovel, and the other was sampling behind the paver using a template. The robotic truck sampler consisted of a large steel frame with a hydraulically operated telescoping arm and sampling probe. The robotic sampler was able to penetrate deep into the load to obtain mix samples, which was believed to provide a less segregated and more representative sample of the mixtures. Due to mechanical problems with the robotic sampler, only one mix was sampled with this device.
The test results indicate that the samples obtained with the robotic truck sampler were more similar to samples of HMA taken just behind the spreader. Compared with samples taken with a shovel from a sampling stand, the mix samples taken with the robotic device were coarser and had lower asphalt contents. The difference in results between these two sampling methods is likely due to some segregation during sampling with the shovel from the top of the load.
A few modifications
As might be expected with most new technologies, numerous modifications and adjustments were made to the automation equipment during the progress of the project. Overall, some of the automated testing devices evaluated in this study provided good data and some need further improvements. The concept of automating testing for some aspects of process control for HMA production is viable. Continued development of automated sampling and testing technologies is essential to provide data that can help advance HMA quality.
