Advances in Civil Engineering

Advances in Civil Engineering / 2010 / Article

Research Article | Open Access

Volume 2010 |Article ID 579631 | https://doi.org/10.1155/2010/579631

Osama Y. Abudayyeh, Joseph Barbera, Ikhlas Abdel-Qader, Hubo Cai, Eyad Almaita, "Towards Sensor-Based Health Monitoring Systems for Bridge Decks: A Full-Depth Precast Deck Panels Case Study", Advances in Civil Engineering, vol. 2010, Article ID 579631, 14 pages, 2010. https://doi.org/10.1155/2010/579631

Towards Sensor-Based Health Monitoring Systems for Bridge Decks: A Full-Depth Precast Deck Panels Case Study

Academic Editor: Farid Taheri
Received14 Jun 2010
Revised06 Dec 2010
Accepted17 Dec 2010
Published17 Feb 2011

Abstract

Traffic and variable loading conditions greatly influence the performance, durability, and safety of a bridge structure throughout its service life. Continuous monitoring can provide the basis for determining the deterioration rate and for estimating the remaining service life, thus assisting in making important decisions regarding bridge maintenance. This paper presents the design and implementation of a health monitoring system for condition assessment of full-depth precast concrete bridge deck, which was developed for the Parkview Bridge in Kalamazoo, Michigan. This system is composed of a remotely accessible on-site data acquisition system and a sensor network of vibrating wire strain gauges to monitor strain and temperature over given time increments. The system relies on the sensor network embedded in the bridge deck to gather static performance data under different loading conditions to provide condition assessment by collecting, storing, analyzing, and reporting relevant performance data over time. The paper presents a practical case study that mainly focuses on describing the initial steps in the development of the sensor network system, namely, the design (system architecture and data structures) and construction along with examples of how the data is acquired, organized, presented, and analyzed, keeping in mind that the bridge is still early in its life-cycle and has not yet experienced any structural problems.

1. Introduction

Bridges are a critical component of the transportation infrastructure. There are approximately 600,000 bridges in the United State according to the U.S. Department of Transportation Federal Highway Administration [1]. Four billion vehicles traverse these bridges daily [2]. Regular inspections and maintenance are essential components of any bridge management program to ensure structural integrity and user safety. This is a grand challenge due to the enormous number of existing bridges.

To assess the condition of a bridge, a few approaches are commonly used in practice. Visual inspection has a long history in bridge condition assessment and documents any sign of cracking, spalling, leaching, deflection and vibration, accidental damage, and deck surface damages. However, assessing the extent of structural deficiency of a concrete bridge is usually unreliable through visual inspections [2]. Coring samples provide a supplementary approach, in which small cores are drilled and concrete samples are obtained and tested in a laboratory. Since samples are taken from small selected portions of the concrete, erroneous conclusions might be reached due to the lack of overall behavioral information throughout the concrete structure. Diagnostic testing is another bridge condition assessment technique in which a bridge is exposed to varying loads and its responses measured and analyzed [3]. Diagnostic testing faces many constraints related to cost and traffic interruption. More importantly, diagnostic testing lacks the capability of continuously monitoring the bridge performance, which is the key to determining the remaining bridge service life [4].

Even though intensive bridge inspection and maintenance are being performed nationwide, the outcomes are not impressive. It has been reported that of the 600,000 bridges, 12% have been deemed structurally deficient while another 13% have been deemed functionally obsolete [1]. With these statistics, 25% of the nation bridges require immediate attention or repair and may present safety challenges. The appearance of widespread failures in bridges emphasizes the importance of effective continuous monitoring systems so that problems can be identified at early stage and economic measures can be taken to avoid costly replacement and catastrophic bridge failures [5]. Therefore, there is a need for bridge health monitoring technologies to enable continuous monitoring and real-time data collection.

This paper presents a sensor-based bridge health monitoring system developed for the Parkview Bridge in Kalamazoo, Michigan. In this study, a bridge health monitoring system was designed and deployed for a newly constructed bridge that adopted rapid bridge construction techniques using precast concrete technology. Sensors were installed at strategic locations and connected to a remote computer workstation via telephone lines. Continuous bridge condition data are being collected in real time, archived in the laboratory computer workstation, and analyzed to assess the structural performance and integrity.

2. Concrete Bridge Health Monitoring

Traffic and variable loading conditions greatly influence the performance, durability, and safety of a bridge structure throughout its service life. In addition to carrying traffic loads, a bridge is exposed to its surrounding environment. Environmental factors such as temperature, humidity, and precipitation can all significantly affect the structural integrity and performance of a bridge as well. The condition of a bridge is never constant and therefore needs to be monitored continuously. Continuous monitoring can provide the basis for determining the deterioration rate and for estimating the remaining service life, thus assisting in making important decisions regarding bridge maintenance.

Among many technologies that have been developed to aid bridge condition assessment, sensor technology has attracted enormous research interest due to its capability of continuously monitoring the bridge condition [47]. With the advancement in sensor technology, many sensors can be embedded in concrete at different locations, comprising a sensor network for structural health monitoring (SHM). These sensor networks can aid in the determination of the true reliability and performance of a structure by interpreting the data collected.

Applications of sensor-based SHM have been expanding in bridge health monitoring to increase safety and help engineers develop models for determining how a structure is behaving internally. At this moment, however, only few studies on sensor-based SHM in practice are available. Chajes reported on a study, in which strain gauges were used to remove restricted load limitations on three bridges in Delaware [8]. Casas described the usage of fiber optic sensors for bridge monitoring including crack detection, strain monitoring, and temperature monitoring [5]. Howell reported the development of an in-service strain monitoring system and its application in a number of bridges in Delaware [4]. Olund reported a series of bridge health monitoring studies conducted in Connecticut using the sensor technology, with recommendations of sensor selection and system configuration [7]. Table 1 provides a summary of the characteristics of some examples of these sensor networks that are already in use [911].


Case studyTypes of sensors used*#of sensors usedSensor placementData collection time intervalsLoad type monitored

Confederation bridge (Canada)(1),(2),(3),(4), (5),(6),(7), (8)389Deck, Beam, PierVariableStatic/dynamic
Pemiscot county bridge, Missouri (U.S.)(1), (3), (4)64Deck, BeamStatic
North halawa valley viaduct, Hawaii (U.S.)(1), (4)200Deck, Beam5 minutes/2 hoursStatic
Parkview bridge, Michigan (U.S.)(1), (4)184Deck10 minutesStatic

*Note: sensor types: (1) Vibrating wire strain gage, (2) Fiber optical, (3) Resistance strain gage, (4) Thermocouples, (5) Accelerometer, (6) Tilt-meter, (7) Displacement, and (8) Ice-force.

These studies demonstrated the applicability of using sensor systems in continuously monitoring the condition of concrete bridges. Observations and findings in these studies laid a foundation for the design and implementation of the sensor network for the Parkview Bridge in Michigan. All the case studies above, except for the Parkview Bridge, had cast-in-place decks. The Parkview Bridge, on the other hand, had precast, full-depth deck panels, making this SHM study different from the others and posing unique challenges.

3. The Parkview Bridge SHM System Design and Installation

The Parkview Bridge is the first prefabricated bridge in Michigan to take advantage of rapid bridge construction techniques. A sensor network was designed and deployed to provide an SHM system for monitoring the performance of the bridge’s full-depth deck panels. The bridge has four spans and three lanes, with all its major bridge concrete elements including piers, abutments, I-beam girders, and full-depth deck panels prefabricated off-site. The superstructure is composed of type IV AASHTO girders, and the deck is composed of forty eight, nine-inch thick precast reinforced concrete panels. These panels are categorized as North and South. Once the North and South panels were installed on-site, they were joined by a cast-in-place grouted joint. The deck is posttensioned with an added three-inch asphalt wearing surface. An overview of the bridge layout is shown in Figure 1.

3.1. SHM Sensor Network Configuration

Properly placed strain and thermocouple sensors can provide valuable information about structural performance. Additional measuring devices can be used, but only strain and temperature measurements were chosen in this project. The selection was based on the location of the bridge (relatively rural with light traffic) and cost, as well as on the reliability of such sensors, particularly the successful long-term performance of similar sensors (nine years in one instance [10]). To further ensure reliability in the sensor network, a number of redundant sensors were installed to allow for continued monitoring in case a sensor stops working. Therefore, the SHM system for the Parkview Bridge is composed of 184 vibrating-wire strain sensors (VWSG) with built-in thermocouples (thermistors) installed in the bridge deck panels. Additionally, the system hardware includes 12 multiplexers, 2 data loggers, 2 modems, a remote computer workstation in a laboratory, and the necessary wiring for communication and data transfer. The sensors and data loggers were calibrated and used per the manufacturer recommendation. They were tested in the laboratory to insure that they were reading correct values before they were placed in the bridge panels. They were installed by the research team and by a certified electrician during the construction phase. The initial temperature and strain values were recorded before placing the sensors in the panel forms. These initial readings are used later in calculating the strain and temperature values after recording the actual readings from the sensors. Figure 2 provides a schematic view of the designed system.

In order to effectively monitor the structural performance, sensors must be placed at strategic locations. The locations of the sensors were based on the structural analysis information provided by the bridge designer. They were installed at and around the locations of maximum stresses and at locations with anticipated future durability issues (i.e., at panel joints and around the closure grout section). Additionaly, redundant sensors were also installed to compensate for the possibility of losing some sensors during construction or beyond. Four groups of strain and temperature sensors were installed at: (1)midspans and supports to monitor longitudinal stresses, (2)midspans in the transverse direction to monitor lateral stresses, (3)edges of deck panels to monitor the joints between panels, (4)along the two sides of the grouted joint between the North and South panels.

Figure 1 shows the locations of the sensors installed at the Parkview Bridge. These sensors were used to capture data throughout the day at ten-minute increments to determine maximum and minimum values of stresses and temperatures recorded. Since the deck and beam act as a composite section after construction, sensors were placed near the top fiber of the section. Due to limitations by the owner, sensors were only placed in the deck.

3.2. Sensor Hardware Installation

Attaching VWSG sensors to reinforcing bars must follow a few precautions to ensure proper operation. In this study, all sensors were attached to the top reinforcement using zip ties with foam spacers to provide cover. Figure 3(a) illustrates the details of a properly secured sensor and how this configuration allows for the free flow of concrete mix while protecting the sensor and its wire during casting.

Once the sensors were properly attached to rebar, the wires connecting them to multiplexers were loosely coiled around the reinforcement to allow concrete bonding between the wires and the reinforcement and to prevent any damage that might occur to the wire during the placement of concrete. The wires were run to a four-inch diameter PVC pull boxes to protect them from the concrete during the pour and to provide accessibility to the wires after the installation of the deck panels at the bridge site. Each wire was labeled to indicate the sensor location and orientation after casting. Figure 3(b) shows a completed sensor network for a panel along with wire routing and pull box placement. Figure 3(c) illustrates the exposed pull box underneath the deck panels for access and splicing. Sensor wires were spliced together and run through PVC conduits underneath the deck panels to the data logging equipment.

In this project, the equipment was housed in three cabinets to protect the electrical equipment from varying environmental conditions, which were secured to the pier of the bridge as shown in Figure 3(d). Each data logger contains a modem for remotely communicating with the laboratory computer workstation for data transfer.

4. The Parkview Bridge SHM Data Structure

To effectively monitor the bridge performance under varying load conditions, sensors were grouped depending on their locations. In this study, four groups of sensors were used to monitor the bridge performance.(1)Group 1: longitudinal stresses at mid spans and over the piers.(2)Group 2: transverse stresses at mid spans.(3)Group 3: stresses at joints between panels.(4)Group 4: Stress at both sides of the cast-in-place grout between North and South panels.

Figure 1 shows the locations and labels of all the sensors in the panels and provides the group number for each sensor in parenthesis. If a sensor belongs to multiple groups, the numbers are separated by commas; for example, Group 1 refers to those sensors placed near mid spans and at pier locations, and orientated longitudinally near the traffic lanes to monitor longitudinal stresses. Table 2 summarizes the list of sensor labels that contribute to Group 1 in each panel.


North panelSensorSensorSouth panelSensor

Span 1
1N_1_C1S_1_A

Pier 1
4N_4_C4S_4_A

Span 2
7N_7_CN_7_F7S_7_F
8N_8_CN_8_F8S_8_F
9N_9_CN_9_F9S_9_F

Pier 2
12N_12_C12S_12_A

Span 3
15N_15_CN_15_F15S_15_F
16N_16_CN_16_F16S_16_F
17N_17_CN_17_F17S_17_F

Pier 3
20N_20_C20S_20_A

Span 4
24N_24_C24S_24_A

Group 2 includes those sensors that are used to monitor the bridge performance under transverse loading. The locations are similar to those in Group 1, but oriented transversely. Table 3 lists all Group 2 sensors. Group 3 refers to those sensors along the edges of critical panels. The main reason for having this group of sensors is to monitor the bonding and load transfer of deck panels. Theoretically, the deck and girders should behave as a composite section, but environmental factors and loading may cause the composite section to behave as smaller sections if joints fail or show fatigue over time. Table 4 displays the sensors used for this category. Group 4 refers to those sensors that are placed along both sides of the center grout joining North and South panels. They are oriented longitudinally. Table 5 lists all sensors belonging to this group.


North panelSensorSensorSensorSensorSouth panelSensorSensorSensorSensor

Span 1
2N_2_AN_2_BN_2_D2S_2_AS_2_CS_2_D

Pier 1
4N_4_DN_4_F4S_4_ES_4_F

Span 2
7N_7_DN_7_D'N_7_GN_7_G'7S_7_ES_7_ES_7_GS_7_G'
8N_8_DN_8_D'N_8_GN_8_G'8S_8_ES_8_E'S_8_GS_8_G'
9N_9_DN_9_D'N_9_GN_9_G'9S_9_ES_9_E'S_9_GS_9_G'

Pier 2
12N_12_DN_12_F12S_12_ES_12_F

Span 3
15N_15_DN_15_D'N_15_GN_15_G'15S_15_ES_15_E'S_15_GS_15_G'
16N_16_DN_16_D'N_16_GN_16_G'16S_16_ES_16_E'S_16_GS_16_G'
17N_17_DN_17_D'N_17_GN_17_G'17S_17_ES_17_E'S_17_GS_17_G'

Pier 3
2020S_20_ES_20_F

Span 4
22N_22_BN_22_D22S_22_CS_22_D
23N_23_BN_23_D23S_23_CS_23_D


North panelSensorSensorSouth panelSensorSensor

Span 1
1N_1_B1S_1_B
2N_2_C2S_2_B

Pier 1
44

Span 2
7N_7_B7S_7_B
8N_8_EN_8_B8S_8_DS_8_B
9N_9_E9S_9_D

Pier 2
1212

Span 3
15N_15_B15S_15_B
16N_16_EN_16_B16S_16_DS_16_B
17N_17_E17S_17_D

Pier 3
2020

Span 4
22N_22_A22S_22_A
23N_23_CN_23_A23S_23_BS_23_A
24N_24_D24S_24_D


North panelSensorSouth panelSensor

Span 1
1N_1_C1S_1_A

Pier 1
4N_4_C4S_4_A

Span 2
7N_7_C7S_7_A
8N_8_C8S_8_A
9N_9_C9S_9_A

Pier 2
12N_12_C12S_12_A

Span 3
15N_15_C15S_15_A
16N_16_C16S_16_A
17N_17_C17S_17_A

Pier 3
20N_20_C20S_20_A

Span 4
24N_24_C24S_24_A

5. Parkview Bridge SHM Data Acquisition and Analysis

The deployment of the Parkview Bridge structural health monitoring system enabled the remote collection of continuous strain and temperature data. Real-time data are being collected at ten-minute intervals. The two data loggers are contacted weekly through the modems and dedicated telephone lines to download and archive the sensor data for future analysis. The raw strain data are converted into stress data using material properties based on the laboratory testing of concrete samples collected during casting. The SHM system started to function in December 2008. Data archiving for a period of three years is currently underway to develop a solid baseline for future continuous monitoring of this bridge’s condition. Examples of data analysis are presented in this section to illustrate how such data is interpreted.

Figure 4 shows the longitudinal stress monitoring for the North panels of Span 2 (Group 1) in January 2009. Note that a negative stress value represents compression. Also, note that the bridge deck is designed to be in compression (deck panels are posttensioned) at all times and that the maximum compression allowed is −3600 psi (−24,821 kPa) and the maximum allowable tension is +537 psi (+3702 kPa). The coinciding temperatures recorded are illustrated in Figure 5. Based on the limited information (one month in the first year), it is observed from Figure 5 that the differences in magnitudes between sensors as well as the slope of the lines over similar time periods are fairly similar (almost identical) with respect to temperature, suggesting a uniform behavior. The trend patterns for each sensor in Figure 4 demonstrate a uniform behavior as well.

Since all deck panels are fully restrained between supports, examining Figures 4 and 5 reveals that as temperature decreases, tension increases, reducing the total compression in the deck panels. It also reveals that as temperature increases, compression increases. This is clearly amplified in Figures 4 and 5 in the period from Tuesday 1/13/09 to Saturday 1/17/09 where the temperature has decreased by 17 degrees Celsius, resulting in a decrease in compressive stress of about 500 pounds per square inch (3447 kPa). The fluctuations between different locations are very minimal, indicating how little effect daily traffic has over the given time period and suggesting that temperature variation is the controlling factor in stress variation. The relationship between stresses and temperatures is key in this analysis, and any change in observed patterns over time may suggest that cracks are beginning to develop in the deck or that the deck is not acting as a fully composite unit (loss of bonding between joints).

When significant variations are noticed between stress lines of different sensors in similar locations over the same time period, further analysis should be performed to distinguish between abnormal and normal behavior. Examining Figure 4 reveals a steep change in stress over a single day from 1/17/09 to 1/18/09. Figures 6 and 7 isolate this day and display the data on an hourly scale. The trend is clearly similar for the sensors under consideration; temperature increased causing compressive stresses to increase. The slopes of the lines over the given time period in similar locations are very close and consistent. The difference in locations caused a slight difference in temperature, but the behavior is the same. When changes in consistency are noticed, a further investigation must be performed. Over the month of January, similar analyses were performed for all locations under this group and no concerns for safety or maintenance were noticed, which is expected since the bridge is new.

The stresses at the joints between panels are very important to monitor due to the unique nature of the Parkview Bridge design. To transfer stresses efficiently, the bond between panels must be maintained so that the deck behaves as a uniform composite member. Once again temperature had the greatest impact on stress fluctuation. The stresses are within allowable limits causing no concern for performance at this time. This investigation focused on the behavior of stress readings between adjacent panel edges. Figures 8 and 9 show the recorded values for the month of January. The stresses at the joint between North panels 7 and 8 (N_7_B and N_8_E sensors) were compared to each other. These sensors should provide similar stress patterns to demonstrate that bonding remains intact between the two different panels. If stress patterns were to change, then a closer look would be needed to determine causes for the change in pattern. If changes were to occur, the prediction would be that bonding had failed or cracks had developed to weaken the bond between the two panels. This analysis was performed on all sensors in this category (Group 3), and no concerns for performance or maintenance were noticed. Similar analyses were performed for the other groups of sensors and similar observations and conclusions had been obtained.

Tables 6 and 7 and Figures 1013 illustrate the monthly maximum and minimum stress levels experienced at critical points (mid spans and piers) in longitudinal and transverse directions during the one-year monitoring. While there is no observed abnormal bridge deck behavior, there are a few instances that have experienced small tensile stresses in the transverse direction. Note that in these tables we are reporting the absolute maximum and minimum stresses experienced during a given month and that some of these readings may not necessarily be correct when the temperatures approach the lower sensor limit (−20°C or −4°F).


North
MonthLongitudinalTransverse
S1P1S2P2S3P3S4S1P1S2P2S3S4

Dec-08Max−1736−1602−1717−1485−1780−1511−1697−1242−1055−1113−1154−1557−768
Tem2.322.663.083.083.263.122.752.272.663.353.443.263.08
Min−1500−1342−806−1198−736−1155−1449−949−863−521−853−422−409
Tem−8.66−8.76−8.76−9.7−9.15−10.27−10.15−7.64−6.35−9.98.66−10.09−10.27

Jan-09Max−1638−1516−1594−1369−1659−1377−1586−1162−961−997−1067−1394−670
Tem−0.34−0.8−0.09−0.30−0.1−0.09−0.57−1.04−0.08−0.170.05−0.5
Min−1155−957−539−820−424−735−1090−694−609−224−596−128−66
Tem−19.1−19.6−18.5−20.2−19.5−20.8−20.8−16.8−18.9−20.7−17−20.74−21

Feb-09Max−1843−1780−1927−1690−2037−1726−1825−1325−1105−1150−1240−1567−870
Tem12.211.3713.7813.1714.1713.7213.4512.0810.5513.6313.6913.7213.63
Min−1310−1211−706−1013−643−974−1278−833−749−367−746−262−254
Tem−16.1−13.3−14.28−12.52−13.86−13−14.18−10−9.38−13.86−9.7−13.86−13.54

Mar-09Max−2005−1944−2109−1865−2210−1934−1918−1459−1266−1277−1359−1691−1006
Tem19.318.2820.6920.0420.920.6620.2820.791821.22120.6921.31
Min−1325−1200−713−1029−654−1010−1322−744−775−356.6−684−180−264
Tem−13.01−11.23−11.65−11.84−12.13−12.24−11.45−13.05−10.18−13.15−12.14−13.15−12.74

Apr-09Max−2224−2155−2365−2129−2468−2205−2129−1618−1346−1430−1496−1845−1176
Tem26.82528.127.0328.1927.9827.7428.7124.5228.8328.2227.8629.2
Min−1664−1518−939−1382−985−1416−1604−997−962−669−930−511−549
Tem−0.530.08−0.35−0.44−0.52−0.61−0.43−0.710.51−0.956−0.61−1.01−0.78

May-09Max−2399−2237−2545−2331−2666−2410−2282−1742−1604−1553−1521−1969−1292
Tem32.833233.1432.6933.133.0532.4234.533.7532.8332.5632.5634.1
Min−1867−1740−1116−1611−1226−1639−1761−1155−1080−824−1158−654−709
Tem9.5910.039.418.978.928.18.119.688.667.9311.027.587.49

Jun-09Max−2572−2523−2769−2567−2908−2627−2456−1828−1667−1660−1697−2067−1402
Tem39.5939.5439.8839.8439.8840.2139.4540.2139.3838.0239.2337.6240.54
Min−2090−1939−1276−1856−1473−1882−2003−1296−1292−994−1246−802−906
Tem13.5414.6514.3714.2814.0113.6414.3713.5515.1213.0914.0913.0913.64

Jul-09Max−2570−2449−2663−2459−2798−2518−2433−1819−1630−1686−1683−2078 −1383.2
Tem33.2432.933.0232.9532.8333.0233.2434.6433.832.733.9132.2934.79
Min−2190−2118−1324−1939−1655−1978−2094−1345−1316−1039−1301−942−936
Tem14.4715.414.1114.2814.113.9114.2514.115.4913.1713.6313.0913.63

Aug-09Max−2624−2524−2750−2542−2882−2592−2492−1872−1634−1715−1742−2107−1446
Tem34.9334.2235.3734.7935.3735.2234.8336.4435.3734.7735.9635.0336.1
Min−2277−2071−1375−2002−1724−2033−2160−1380−1381−1073−1347−980−985
Tem12.913.7313.181312.9212.631312.8114.0912.0912.6311.9812.36

Sep-09Max−2579−2435−2654−2442−2790−2510−2447−1800−1614−1686−1687−2081−1390
Tem29.4528.9629.729.630.0830.1829.9530.5929.729.230.5429.4531.23
Min−2249−2025−1332−1957−1683−1988−2143−1330−1327−1044−1319−947−941
Tem8.018.897.758.288.097.938.297.938.87.497.847.47.49

Oct-09Max−2424−2218−2406−2199−2525−2245−2289−1647−1451−1550−1555−1952−1242
Tem18.2617.518.8718.3818.9718.7818.7118.7817.3118.6518.5718.7718.97
Min−2167−1898−1244−1834−1565−1835−2043−1252−1210−970−1225−872−859
Tem2.582.752.492.42.492.232.752.42.841.992.152.152.66

Nov-09Max−2364−2170−2342−2136−2467−2185−2244−1603−1405−1457−1523−1899−1203
Tem16.6316.1617.716.9318.0917.7917.717.2116.0617.2117.417.8517.7
Min−2083−1801−1184−1722−1448−1708−1954−1170−1132−885−1146−769−766
Tem−1.12−0.95−1.39−1.72−1.65−2.09−1.91−1.65−0.61−2.08−1.83−2−1.83

*1 psi = 6.89474 kPa.

South
MonthLongitudinalTransverse
S1P1S2P2S3P3S4S1P1S2P2S3P3S4

Dec-08Max−1627−1703−1012−1616−1582−1778−1799−869−1005−1112−914−1571−1119−938
Tem12.0811.112.1812.7312.2713.081312.6310.313.081313.2213.1813
Min−1195−1061−147−925−617−1058−1318−269−619−176−378−20−374−306
Tem−13.45−13.45−14.28−14.3−14.5−14.79−14.7−14.66−13.39−14.81−15.02−14.81−15.04−15.02

Jan-09Max−1416−1406−807−1288−1302−1458−1578−624−815−924−706−1379−907−706
Tem−0.17−0.6−0.6−0.27−0.5−0.11−0.09−0.5−1.2−0.08−0.430.1−0.260.09
Min−992−83422−693−405−849−1109−66−45726−179202−185−102
Tem−19.28−19.1−19.76−19.9−19.76−20.25−20.5−20.58−18.93−20.35−20.74−20.74−20.86−20.76

Feb-09Max−1614−1739−975−1664−1540−1827−1813−808−968−1088−878−1545−1100−919.5
Tem1312.211.313.9112.27141412.0910.3913.912.9214.1413.1213.91
Min−1107−1024−153−865−651−1040−1275−227−613−126−35373−355−277
Tem−13.5−11.84−12.54−12.85−12.44−12.94−12.74−13.87−11.35−13.66−13.15−13.76−13.05−13.66

Mar-09Max−1771−1907−1129−1834−1766−2011−1928−969−1113−1209−999−1657−1221−1057
Tem20.3819.3719.2720.7719.8620.9620.6720.3818.4821.121.220.7921.220.69
Min−1110−1049−18.8−905−708−1071−1299−231−647−123−38148−396−296
Tem−13.25−11.35−11.84−12.14−11.55−12.24−11.74−13.56−10.57−12.84−12.42−12.94−12.14−12.54

Apr-09Max−1980−2131−1246−2097−1957−2291−2136−1109−1197−1372−1121−1774−1340−1250
Tem27.9326.3325.227.8626.3328.2128.0427.224.6328.6728.0527.8228.2228.22
Min−1427−1401−441−1290−1047−1469−1595−544−824−440−638−219−869−578
Tem−0.73−0.01−0.35−0.35−0.35−0.62−0.69−0.940.34−0.87−0.69−0.82−0.61−1.04

May-09Max−2059−2220−1338−2170−2088−2345−2202−1237−1350−1487−1230−1889−1431−1364
Tem29.572927.1528.8628.128.7228.5333.3132.6233.833.6432.2233.3833.1
Min−1614−1657−635−1529−1282−1723−1774−726−1030−611−802−386−804−771
Tem9.510.399.178.899.158.288.18.889.867.848.277.8368.187.49

Jun-09Max−2329−2574−1640−2558−2455−2741−2506−1396−1519−1612−1362−1810−1573−1503
Tem41.0641.4139.7140.8739.8840.8940.0540.3840.3840.7240.7439.0840.8940.38
Min−1812−1855−891−1751−1504−1963−2019−883−1148−774−974−540−974−955
Tem13.2714.714.3114.251413.8514.2813.0914.6513.2713.8213.0913.7313.36

Jul-09Max−2312−2459−1554−2409−2386−2608−2466−1315−1477−1587−1313−1820−1545−1464
Tem34.3634.533.133.6633.133.3833.833.8434.3533.8734.0832.0233.8333.38
Min−1930−1940−958−1823−1596−2045−2091−935−1200−860−1020−600−1027−988
Tem14.1115.3514.231414.1914.1141415.4913.4513.8213.441413.45

Aug-09Max−2363−2528−1605−2497−2416−2687−2515−1378−1504−1626−1353−2022−1590−1519
Tem36.0935.8134.535.5134.6435.2235.0735.5535.2235.3735.9234.0735.5135.07
Min−2000−2000−1004−1892−1678−2108−2171−969−1255−904−1085−675−1100−1073
Tem12.7213.6413.0912.8112.8212.6312.7212.4513.6412.1812.6312.0812.6312

Sep-09Max−2315−2438−1523−2282−2339−2600−2491−1321−1466−1597−1320−1826−1566−1478
Tem30.5730.4629.4730.0830.2130.5330.5530.7230.4530.5931.0629.8331.430.51
Min−1980−1953−957−1935−1620−2067−2150−925−1213−884−1058−644−1078−1022
Tem7.938.887.838.17.678.028.17.668.457.487.947.148.016.97

Oct-09Max−2131−2186−1325−2130−2029−2318−2306−1156−1297−1460−1183−1901−1439−1325
Tem18.5818.0917.518.7317.9718.9418.9718.3917.0819.0718.4218.7818.6818.76
Min−1884−1812−881−1867−1516−1930−2048−842−1119−814−962−588−985−958
Tem2.242.752.582.232.582.322.752.152.662.322.152.232.412.34

Nov-09Max−2066−2111−1321−2066−2006−2249−2242−1149−1283−1415−1168−1867−1443−1283
Tem16.8716.4417.1117.2117.2117.717.6317.9916.6317.8917.7917.9818.2818.08
Min−1794−1699−810−1784−1392−1801−1946−743−1073−732−880−537−904−855
Tem−1.48−1.12−1.22−1.79−1.48−2.08−1.96−1.84−0.61−1.83−1.92−1.91−2.09−2

*1 psi = 6.89474 kPa.

Once a three-year data set has been collected and the bridge deck behavior analyzed, stress envelopes can be developed to provide a baseline for normal maximum and minimum stress values. We feel that three years of stress data collection and analysis in the early stages of the bridge life-cycle are necessary to experience all stress scenarios from traffic, environmental, and bridge weight (creep) loads to enable the development of a representative set of stress envelopes (baseline). If stresses fall outside these envelopes, this would trigger further investigation to determine the cause(s) for the deviation and to recommend the course(s) of action. While a longer period is more desirable, we are limited in this study by the timeline as desired by the Michigan Department of Transportation (MDOT), where Phase I and II of this project will result in the design and construction of the system along with three years of sensor stress data collection to develop the baseline stress envelops and a deterioration prediction model based on finite element simulations. After phase II is completed, we hope to begin actual monitoring of the bridge for some period of time until we turn the bridge over to MDOT for life-cycle monitoring. Certainly, the three-year baseline can always be reassessed after many years of data have been collected beyond phase II to determine if additional years in the baseline may lead to better prediction results. At this time, a three-year period seems to be a reasonable predictor of the weather patterns in our area.

6. Concluding Remarks

A health monitoring system was designed and deployed for the Parkview Bridge in Kalamazoo, Michigan. It is anticipated that this sensor-based health monitoring system would be capable of providing continuous monitoring of the bridge deck to determine its condition, assess the impacts from environmental factors such as temperature and from traffic loads, evaluate its deterioration rate, initiate maintenance and repairs when needed, and predict the remaining service life. Even with limited preliminary data, meaningful observations regarding the bridge performance and the relationship between temperature and stress can be obtained. It was found that recorded stresses vary widely due to the combined effect from loading and temperature variations. However, it was concluded that temperature is the controlling factor in stress variations that are measured by the static sensors in this study. Variations in temperature cause the bridge’s behavior to vary from season to season.

A few lessons were learned in this study. The first lesson is related to the installation of sensors which is found to need a formal quality control procedure. The method used for securing sensors worked fine, but could be further improved. For example, the foam spacers, which are tightened by zip ties, would occasionally fall out if disturbed before casting, especially when the workers were in contact with the reinforcement or stepping on the zip ties securing the sensors. To improve on this problem, workers should be prevented from standing over the rebar mesh when pouring concrete. A second lesson deals with sensor wire connections which present another challenge. Due to the large volume of sensors and wires being connected and spliced, strict supervision needs to be provided. The integrity of the project relies on proper sensor readings from known locations and orientations. The labeling and splicing processes must be carefully supervised to avoid errors in labeling sensors. A third lesson deals with the location of the cabinets that house the data loggers. While the cabinets are installed at the top of the pier, 16 feet (4.8 meters) high, to prevent unwanted access to the expensive equipment, this has posed an access challenge for maintenance when needed, particularly that the pier is very close to live highway traffic, requiring extra safety measures. A better approach would be to provide secured cabinets at the ground level for easy access.

Since the sensors use analog signals via telephone lines to communicate data, electrical noise interference can significantly affect the validity of the readings, degrading the signal. Strain data recorded by the sensors may show out-of-range values that are caused by signal interference. Long cable lengths have been found to weaken and degrade the analog signal as well. When using a large volume of vibrating wire strain gages, it is recommended to use a minimum time of ten-minute intervals for continuous monitoring. Using shorter increments were found to cause several erroneous readings. Furthermore, depending on the number of arrays and frequency of data collection, capacity limitations of the data logger must also be determined. Prior estimates of capacity were determined to be two weeks when the actual capacity was closer to one. This has caused data to be overwritten and lost for that week.

We recognize that at this time the bridge is still new and is not expected to have problems. However, as time goes by, data covering a relatively long period of time will be collected and, when combined with a bridge deterioration model, can help predict bridge performance and call for timely preventative maintenance. We believe that three years worth of monitoring data would provide sufficient baseline information to create behavior envelopes that can be used for future prediction of bridge condition and can be the basis of a bridge deck deterioration prediction model. The following is a discussion of our future tasks that will need to be achieved to fully reap the benefit from the sensor network discussed in this paper.(i)Stress Envelopes. Once large amounts of data have been collected and processed for a minimum recommended period of three years, envelopes can be developed to determine normal performance patterns and condition. This can provide fast and efficient assessment means for periodically evaluating the condition of the bridge deck in comparison to the design and behavioral limits.(ii)Full-Depth Deck Performance and Health Condition Prediction Modeling. Since the bridge is new, it is too early to begin to understand how loads will impact the bridge’s deck over time. Further analyses will be performed by finite element modeling, particularly at panel joints and at the closure grout, using actual strain data from the sensor network to shed additional insights on the health of the bridge deck. As more and more data is collected each year, these models could be adjusted, calibrated, and validated using the known behavioral strain data to simulate and develop defect signatures that would then be used for predicting the future behavior and condition of the bridge deck.

Acknowledgments

The authors are grateful to the Michigan Department of Transportation (MDOT) for funding this study through MDOT Contract no. 04-0090/Z3. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of MDOT or Western Michigan University.

References

  1. Federal Highway Administration (FWHA), “Deficient bridges by state and highway system,” U.S. Department of Transportation, 2007, http://www.fhwa.dot.gov/bridge/defbr07.cfm. View at: Google Scholar
  2. B. Phares, T. Wipf, L. Greimann, and Y. Lee, “Health monitoring of bridge structures and components using smart-structure technology,” Tech. Rep. 0092-04-14, Wisconsin Highway Research Program, WHRP 05-03, 2005. View at: Google Scholar
  3. NCHRP-234, Manual for Bridge Rating through Load Testing, Research results digest, no. 234, National Cooperative Highway Research Program, Transportation Research Board, National Research Council, Washington, DC, USA, 1998.
  4. D. A. Howell and H. W. Shenton, “System for in-service strain monitoring of ordinary bridges,” Journal of Bridge Engineering, vol. 11, no. 6, pp. 673–680, 2006. View at: Publisher Site | Google Scholar
  5. J. R. Casas and P. J. S. Cruz, “Fiber optic sensors for bridge monitoring,” Journal of Bridge Engineering, vol. 8, no. 6, pp. 362–373, 2003. View at: Publisher Site | Google Scholar
  6. J. M. Ko and Y. Q. Ni, “Technology developments in structural health monitoring of large-scale bridges,” Engineering Structures, vol. 27, no. 12, pp. 1715–1725, 2005. View at: Publisher Site | Google Scholar
  7. J. Olund and P. E. DeWolf, “Passive structural health monitoring of Connecticut's bridge infrastructure,” Journal of Infrastructure Systems, vol. 13, no. 4, pp. 330–339, 2007. View at: Publisher Site | Google Scholar
  8. M. J. Chajes, H. W. Shenton, and D. O'Shea, “Bridge-condition assessment and load rating using nondestructive evaluation methods,” Transportation Research Record, vol. 2, no. 1696, pp. 83–91, 2000. View at: Google Scholar
  9. M. Cheung, W. Li, and B. Noruziaan, “Data acquisition, processing and management systems for a Canadian bridge monitoring project,” in International Conference on Computing in Civil and Building Engineering, Bauhaus-Universitat, Weimar, Germany, June 2004, http://e-pub.uni-weimar.de/volltexte/2004/128/pdf/icccbe-x_012.pdf. View at: Google Scholar
  10. I. N. Robertson, “Prediction of vertical deflections for a long-span prestressed concrete bridge structure,” Engineering Structures, vol. 27, no. 12, pp. 1820–1827, 2005. View at: Publisher Site | Google Scholar
  11. Y. Yang and J. J. Myers, “Live-load test results of Missouri's first high-performance concrete superstructure bridge,” Transportation Research Record, no. 1845, pp. 96–103, 2003. View at: Google Scholar

Copyright © 2010 Osama Y. Abudayyeh et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder
Views2400
Downloads764
Citations

Related articles

We are committed to sharing findings related to COVID-19 as quickly as possible. We will be providing unlimited waivers of publication charges for accepted research articles as well as case reports and case series related to COVID-19. Review articles are excluded from this waiver policy. Sign up here as a reviewer to help fast-track new submissions.