Radio Frequency Identification Applications in Pavements

CHAPTER 5: RFID TRACKING OF PCC PLACEMENT

INTRODUCTION

To ensure and ideally to improve the performance of concrete pavements, it is vital that the influence of material properties on performance be clearly understood. Correlations between as-constructed properties of PCC in construction databases and field performance in PMSs can quantify the link between material quality and pavement performance.

PCC is produced at a production facility and then trucked to the highway construction site for offloading. As part of an overall QA program, PCC producers sample their production periodically from delivery trucks as they leave the plant to perform various quality control (QC) tests to ensure that the mixture properties remain within acceptable limits. Agencies typically take additional samples for quality acceptance testing to corroborate the producer's QC test results and to establish pay factors.

Knowledge of where the sampled truckload of material is deposited along the roadway is critical when attempting to correlate as-constructed material properties with measured performance from a PMS because pavement management data are typically referenced to a specific spatial location (milepoint, latitude/longitude). Unless the PCC QA data can also be tied to a spatial location along the pavement, they cannot be correlated accurately with pavement management data.

Similar to the HMA tracking application considered in chapters 2 through 4, sensors based on RFID technology could be used to tag truckloads of PCC as they left the production plant. The sensors would be cast into the slab during the concrete paving operation. After construction, a vehicle-mounted scanner would be used to electronically "read" the identity tags to link them to the material properties measured from samples obtained from the truck that brought the concrete to that location. The data could then be directly linked to future pavement performance data in the agency's PMS, enabling robust statistical analyses of the correlations between material properties and actual performance.

LABORATORY EVALUATION

A preliminary laboratory evaluation of read range was conducted by embedding conventional UHF passive RFID tags in 4-inch (100-mm) diameter by 12-inch (300-mm)-high concrete cylinders. Alien® Gen 2 2x2 and Alien® Gen 2 1x1 tags were encapsulated using the same techniques as for the HMA tracking tags and then embedded in the middle of the cylinders in an upright position. Two replicates of each tag type were embedded in a one tag per cylinder basis. After the tags were embedded, read ranges were measured periodically as the cylinders were curing. All readings were conducted using a Mercury® 5 ThingMagic reader. No relationship between the curing of the cylinders and the read range of the RFID tags was observed. To determine the orientation effect on the readings, the read range was recorded for the cylinders in vertical and horizontal positions relative to the antennae. The results are shown in table 11 and table 12 for readings taken at 20-s intervals after 13 days of curing. The best read range results obtained when the RFID tags where in front of the transmit side of the antenna are given.

Table 11 . Read range for vertically positioned cylinders.

Table 12 . Read range for horizontally positioned cylinders.

Based on these results, it can be seen that the larger format 2x2 tags have much longer read ranges through concrete compared with the smaller format 1x1 tags. Orientation is also a very important factor in the consistency of the readings. The read range for the second replicate of the 2x2 tags in the horizontal position was only 1 ft (0.3 m). Because the orientation of the tags in the field is purely random, it should be anticipated that some of the tags could end up in the worst orientation and may not be readable; this appears to be what happened for one of the 1x1 tags. Nonetheless, these preliminary results show the potential of this application in PCC tracking.

FIELD EVALUATION

The concept of PCC tracking was evaluated in the field in conjunction with the reconstruction of a section of the New York State Thruway I-90 outside Syracuse, NY. The Syracuse Project was located between exits 39 and 40 on I-90 as shown in figure 55. Three different types of tags were studied: Alien® Gen 2 Higgs® 2 2x2, Alien® Gen 2 1x1, and Gen 2 Titan™ tags. A total of 80 tags were used in this project, with seven Alien® 2x2, seven Alien® 1x1, and six Titan™ tags in each truck.

Original image: ©2015 Google®; map annotations provided by University of Maryland
Figure 55. Photo. Location on the Syracuse project where the RFID tags were embedded.⁽⁹⁾

Methodology

The Syracuse field trial was performed in two steps. The first was the embedment of the RFID tags in the concrete pavement, which took place on October 8, 2009. Four sets of tags, each consisting of three different types of tag, were used for this project. The Alien® Gen 2 Higgs® 2 2x2 and 1x1 tags were pre-encapsulated in CPVC pipes with high-temperature epoxy following the procedure explained in chapter 2. The tags were curled inside of the CPVC pipes in an effort to keep the final product in the size and shape of an aggregate, thus maintaining the integrity of the concrete. The Titan™ tags were already encapsulated by the manufacturer to endure high humidity and pressure.

The first three sets of tags, each including seven Alien® Gen 2 2x2 tags, seven Alien® Gen 2 1x1 tags, and six Gen 2 Titan™ tags, were sown in three consecutive trucks as they were being loaded with concrete at the plant. These were the last trucks to unload before a bridge deck. After unloading, the tags in the concrete mix went through the slip form paver and were finished into the 13-inch (325-mm) slab. Some concern was raised about whether the concrete unloaded from these three trucks would be cut and replaced with manually finished concrete next to the bridge deck because bridge obstacles might interfere with the progress of the slip form paver. However, it was later confirmed that there was no slip form concrete removed from in front of the bridge. The fourth set of tags was sown in the sixth truck after the bridge. The finished pavement with the tags embedded is shown in figure 56.

The second step was to read the embedded tags after the concrete had cured for 24 days. Two antennae were connected to the reader and mounted on the outside of a cargo van at a height of approximately 6 inches (150 mm) from the pavement, as shown in figure 57. The antennae each contain a receive side and a transmit side. In the laboratory evaluation, the best readings were received when the tags were in front of the transmit side of the antenna. Therefore, the most efficient way to cover the pavement area was to position the antenna sequence across the width as T/Transmit-R/Receive-T/R. Overlapping passes at very low speeds were performed to cover the entire width of the pavement. The overlapping passes were repeated three times. The pavement area was also manually checked for tags by moving the antennae by hand over the pavement surface.

Figure 56. Photo. Finished concrete pavement.

Figure 57. Photo. Antenna configuration.

Field Test Results

None of the embedded tags were read in any of the passes. This was not unexpected for the Alien® Gen 2 1x1 tags because as shown in table 11 and table 12, their read range was much lower than that of the Alien® 2x2 tags. However, the inability to read any of the Alien® Gen 2 2x2 or the Titan™ tags was a surprise. The Titan™ tags could be read up to approximately 6 ft (1.8 m) in air.

It is unlikely that the failure to read any tags was due to equipment malfunction. The reader was checked and worked in the laboratory immediately before and after the test, and it passed all diagnostics in the field on the day of the test. The possibility that all of the tags ended up in the bottom of the 13-inch (325-mm) slab and could not be read because of severe attenuation was also highly unlikely.

The inability to read any of the tags in the field was particularly surprising because RFID tags have been used successfully in tracking precast concrete. Trackcon®, a division of International Coding Technologies (ICT), uses passive UHF tags to track precast concrete components. The tags are UPM Raflatac® Dogbone tags incorporating the Monza® 3 chip, the newest chip at the time from Impinj®, Inc. They are molded into the surface of precast concrete components. Trackcon® has also tried to embed them in concrete, and according to Tom Tilson, CEO of ICT, the tags can be read through 6 inches (150 mm) of concrete and sporadically even through 12 inches (300 mm) of concrete. However, for these tags to be read, the 9090 Motorola® handheld reader had to be no farther than 3 to 4 inches (75 to 100 mm) away from the tag. Falken Secure Networks, Inc. markets another commercial system for tracking precast concrete components using RFID. Company representatives confirm reading through precast concrete of thicknesses of 6 to 8 inches (150 to 200 mm) using the latest generation tags such as Alien® Higgs® 3 or UPM Raflatac® Dogbone Monza® 3 with the 9090 Motorola® handheld reader.

The following are two possible reasons that these two companies were able to read tags through concrete but the current study could not:

• The two companies used tags incorporating the newest chips available such Monza® 3 and Higgs® 3.
  
  
• The tags they used are encapsulated flat instead of curled as the ones used for PCC tracking in this study were.
  
  

Even though ICT was able to read some tags through concrete, the company is not pursuing the idea of inserting passive tags in concrete because it has concluded that the read ranges and read success are not good enough for successful application. ICT continues to place RFID tags only on the surface of precast components.

Because these puzzling insights from similar applications were inconclusive, further investigations were conducted to determine why the tags in the present study could be read satisfactorily through HMA but not at all through PCC.

FURTHER INVESTIGATIONS

Laboratory Evaluation

Encapsulated RFID tags were embedded in a concrete block in the laboratory to simulate conditions in the field. Only the Alien® Gen 2 2x2 tags were used for the laboratory evaluation because they had the best read range in air. An 8-ft³ (0.23-m³) wooden box was built for the laboratory evaluation.

Because coarse aggregate is the largest component of concrete and often has a high dielectric constant, 12 tags were first placed in known positions inside the box as it was filled with coarse aggregate. The coarse aggregate was a limestone from the Millville quarry in West Virginia. This aggregate had an estimated dielectric constant of 7 (dry) or 8 (wet).⁽¹⁰⁾ The RFID tags were placed in three layers as shown in figure 58 through figure 61. The box was then read with the antenna positioned at the mid-height of the box at distances of 1 and 2 ft (0.3 and 0.6 m) away from the box. The antenna was positioned horizontally with the transmitter on the right side. Readings were taken with the middle of the antenna located between 24 inches (600 mm) to the left of the box and 24 inches (600 mm) to the right in 4-inch (100-mm) intervals for all four sides. All the read ranges were conducted in the laboratory, so there was the inevitable random interference due to metal objects.

Because the box is symmetrical, the readings on each side should have theoretically been identical. However, owing to interferences and possible inconsistencies when placing the tags in the box, there were slight variations in the readings from one side to the other.

The number of tags read at each offset averaged for each side is shown in figure 62. It can be seen that the number of tags read is skewed, with more tags being read on the left side of the box (negative values for distance). From previous read range studies, it was known that the RFID tags read best if they are closer to or in front of the T (transmit) side of the antenna. Because the T side of the antenna was on the right, increasing the offset on the right side increased the distance from the T side to the tags, which explains the skewed readings.

Another important fact is that, although the antenna was placed only 1 or 2 ft (0.3 or 0.6 m) away from the box, that is not the actual antenna-to-tag distance. The distance to the tags also included an extra 4 to 20 inches (100 to 500 mm) of coarse aggregate based on their position inside the box.

The maximum distance at which any of the tags was read was generally on the order of 6 ft (1.8 m).

Figure 58. Photo. Tags located 8 inches (200 mm) from the bottom and 4 inches (100 mm) from the sides.

Figure 59. Photo. Tags located 12 inches (300 mm) from the bottom and 12 inches (300 mm) from the sides.

Figure 60. Photo. Tags located 8 inches (200 mm) from the top and 8 inches (200 mm) from the sides.

Figure 61. Photo. Box filled with aggregate and 12 RFID tags.

1 ft = 0.305 m
1 inch = 25.4 mm
Figure 62. Chart. Average read range of four sides of the box for the RFID tags embedded in coarse aggregate.

Next, the same tags were placed in the same locations as in the aggregate tests as the box was filled with concrete. None of the tags could be read after encapsulation in concrete, even after the concrete had cured for 1 month. To check the survivability of the tags, they were extracted from the concrete block via demolition. The tags could be read after extraction, so survivability was not the issue.

Additional tests were performed in an effort to determine the reason these passive RFID tags could not be read through concrete. Because the tags could be read through the coarse aggregate, the next logical step was to determine whether the moisture present in the concrete attenuated the electromagnetic waves enough so that the chip could not be activated. The dielectric constant of water is about 80, which makes it one of the most difficult media for the electromagnetic waves to penetrate.

A 32 gal (121 L) plastic bucket was filled with sand at about 9.5-percent water content, which was similar to the water content for the concrete mixture. Sixteen tags were embedded at various depths in the sand at a maximum of 7 inches (175 mm) in from the sides of the bucket. All of these tags were read successfully from at least 2 ft (0.6 m) away. This finding shows that moisture by itself is not the reason there were no readings through concrete.

As a final test, six tags were embedded at three different depths in a 5-gal (19 L) plastic bucket of cement paste prepared at a water-to-cement (w/c) ratio of 0.4. The tags closest to the surface were embedded only to about 3.5 to 4 inches (87 to 100 mm). None of these tags could be read even after 12 days of hydration. These findings clearly indicated that the cement paste was the reason for the unsuccessful readings through concrete.

The Dielectric Constant of Cement Paste

Using mixture theory, the dielectric constant of concrete is approximately 11, which alone would not account for the poor read performance. Therefore, in an effort to determine the reasons that the RFID tags could not be read through the cement paste and ultimately through the concrete block, a thorough literature review was performed on the dielectric constant of cement.

Zhang et al. used a microwave technique to determine the dielectric constant of cement pastes with different w/c ratios at frequencies between 8.2 and 12.4 GHz for up to 30 hours of hydration.⁽¹¹⁾ The w/c ratio of 0.4 is the closest to the one used in the present study. Because attempts to read the tags embedded in the concrete block were made after 1 month of hydration-i.e., after most of the water was hydrated-the dielectric constant of interest from Zhang et al.'s study was the latest one after 30 hours of hydration. Zhang et al.'s measured dielectric constant for the cement paste was about 20 at 8.5 GHz, which is the closest frequency to the 915 MHz UHF frequency used in the present study. Other results by Zhang et al. show that the dielectric constant increases as frequency decreases, which means that the dielectric constant will be greater than 20 for frequencies close to 1 GHz.

Wen and Chung examined the dielectric constant of cement pastes with different admixtures.⁽¹²⁾ Table 13 summarizes their key results. The first row gives the dielectric constant of Portland cement mixed only with water, which is the value of interest for the present study. At 1 MHz, this value is 23.7±2.8. For frequencies of about 1 GHz, the dielectric constant would be slightly lower, approaching perhaps 20.

From these two papers, it is seen that the dielectric constant of cement is much higher than that of the coarse aggregate and coarse aggregate and water as estimated using mixture theory. Although the coarse aggregate is the main component of concrete, the cement paste covers all of the aggregate. The electromagnetic waves coming from the antenna must first go through the cement paste, which could strongly attenuate the signal due to its high dielectric constant. After that, the electromagnetic waves experience still more attenuation from the aggregate, making it hard for the signal to reach the RFID tags and power the chip.

Table 13 . Dielectric constants for various cement pastes (after Wen and Chung⁽¹²⁾).

- Not applicable
Admixtures: L = Lime, SF = Slag Flyash
KHz = kilohertz
MHz = megahertz

CONCLUSIONS REGARDING PCC TRACKING

Preliminary laboratory evaluation of concrete cylinders suggested that the Alien® Gen 2 2x2 tags could be read through about 2 inches (50 mm) of concrete at a distance of up to 5 ft (1.5 m). Although this suggests that PCC tracking should therefore be successful, negative results were obtained during the Syracuse project field trials. None of the encapsulated RFID tags embedded in the concrete pavement were read even after 24 days of concrete curing.

In an effort to determine the reason for this failure, an extensive laboratory evaluation was conducted. Twelve encapsulated RFID tags were embedded in three layers in an 8-ft³ (0.23-m³) box filled with coarse aggregate. These tags were all successfully read as the antenna was moved around all sides of the box. This suggests that attenuation of the radio signals by the coarse aggregate, the main component of concrete, is not the reason the tags could not be read in the field.

Next, the same configuration as for the aggregate box was applied to the box filled with concrete. None of the tags could read after 1 month of curing, even though four tags were embedded only 4 inches from the box side. All tags could be successfully read when extracted after demolition of the concrete box.

The next test consisted of inserting 16 tags in a bucket filled with sand and water. All of the tags could be read from at least 2 ft (0.6 m) away. Finally, six RFID tags were embedded at different depths in a 5-gal (19-L) bucket of cement paste. None of the tags could be read even after 12 days of hydration. This indicates that the cement paste, which has a dielectric constant of about 20, is the main factor attenuating the electromagnetic waves in the concrete.

In light of these results, it can be said that passive RFID tags, although successful in tracking HMA, are not suitable for tracking PCC. Active tags, with their much higher transmitting power, could possibly work for this application; however, because they are needed in large quantities, their high price would make them unsuitable for the PCC tracking application.