Technology

information and updates on the impact of technology on structural engineering

Seoul ®

Daejeon

nd Figure 1: Twin Jindo Bridges (2 Jindo Bridge is on the left). Copyright Wireless Monitoring of Civil Infrastructure Comes of Age Jindo Jindo Bridges Island

edical personnel routinely per- Girardeau, Missouri, is approximately $1.3 mil- By B.F. Spencer, Jr., Soojin Cho, form health screenings for the lion for 86 accelerometers, which makes the and Sung-Han Sim early detection of disease so that average installed cost per sensor a little more magazineappropriate preventive measures than $15,000; this cost is not atypical of today’s S TMcan be Rtaken. HealthU is an importantC Tissue not Uwired SHMR systems. E only for our population but also for our nation’s This article discusses a low-cost, wireless means civil infrastructure, much of which is rapidly for continuous and reliable structural health approaching the end of its intended design life. monitoring recently developed by researchers at B.F. Spencer, Jr. ([email protected]), With age comes deterioration; the most recent the University of Illinois at Urbana-Champaign, is the Nathan M. and Anne M. evaluation by the American Society of Civil as well as its successful deployment at full scale Endowed Chair in Civil Engineering Engineers cites grimly poor “grades” for almost on the 2nd Jindo Bridge in South (Figure and the Director of the Newmark all infrastructure types. The need for routine 1). This SHM system is the first long-term, dense Structural Engineering Laboratory monitoring of civil infrastructure is now more deployment of a wireless sensor network to mon- at the University of Illinois at critical than ever. itor civil infrastructure and demonstrates the Urbana-Champaign. Bridges account for a large part of the capital tremendous potential of this technology. Soojin Cho ([email protected]), investment in the construction of road networks is a Postdoctoral Research Associate and represent a key element in terms of the safety Next Generation Wireless in the Smart Structures Technology and functionality of the entire highway system. Smart Monitoring System Laboratory at the University of Structural health monitoring (SHM) offers the Illinois at Urbana-Champaign. ability to continuously observe the integrity of While much of the technology associated with our nation’s bridges in real time, with the goal wireless smart sensors (WSS) has been avail- Sung-Han Sim ([email protected]), of enhanced safety and reliability, and reduced able for over a decade, only a limited number is an assistant professor at UNIST maintenance and inspection costs. Furthermore, of full-scale implementations have been realized, (Ulsan National Institute of Science such SHM systems allow for emergency facilities primarily due to the lack of critical hardware and and Technology) in Korea. and evacuation routes, including bridges and software elements. Using MEMSIC’s Imote2, highways, to be assessed for safety after cata- researchers at the University of Illinois at Urbana- strophic events, such as earthquakes, hurricanes, Champaign have developed a flexible WSS tornados, etc. framework for full-scale, autonomous SHM that Many recently constructed bridges have in- integrates the necessary software and hardware depth, yet costly, monitoring systems. For elements, while addressing key implementation example, the total cost of the monitoring system requirements. The hardware delivers the neces- on the Bill Emerson Memorial Bridge in Cape sary high-fidelity data, while the software allows

12 October 2011 STRUCTURE magazine 24 72 72 24

Jindo side  Haenam side (South) (North)

24 81 T 118 37 133 34 87 120 104 51 81 51 24 104 120 87 34 133 69 53 88 37 101 88 6953

East West East West Deck networks 113 WT 2 9 3 7 69 66 53 41 39 64 100 56 8 20 3 23 6 20 152 33 35 53 83 54 69 114 T S

West 101 118 70

71 32 East 104 80 25 88 136 145 121135 102 144133 86 98 96 116144 81 18135 150132 39 129 85 52 ST H T H H H 151 H H H H 19 149 146 W W Cable networks W 52 132 6 170 145 35 21 105 86 23 3 41 70 7 150 151 56 135 119 4 40 20 133 65 West

69 53 88 37 133 34 87 120 104 24 51 81 85 113 51 32 24 73 23 89 48 2 34 64 Jindo‐side East Haenam‐side Figure 2: Imote2/ISM400 wireless smart sensor node. Pylon Pylon

: Jindo Deck network (single hop) : Jindo Cable network (multi hop) engineers to more readily realize the potential : Haenam Deck network (single hop) : Haenam Cable network (single hop) of smart sensor technology. ® An example of one such critical issue W : Anemometer (SHM‐DAQ) WT : Wind Turbine power S : Strain sensor (SHM‐S1) T : Temperature H : High‐sensitivity (SHM‐H) ST : Strain + Temp correction (SHM‐S2) addressed by this framework is network scalability. A wireless sensor network imple- Figure 3: Sensor layout on the Jindo Bridge. mented on the Golden Gate Bridge in 2008 nd took approximately 10 hours to collect 80 The wireless monitoring system on the 2 seconds of data (sampled at 1000 Hz) from 56 Jindo BridgeCopyright was initially installed in the sensor nodes to a central location. To address summer of 2009. Seventy-one state-of-the- such problems, Illinois researchers leverage art wireless smart sensor nodes, with a total the on-board computational capacity of the of 427 sensing channels, were installed on WSSN to allow data processing to occur the girder, the pylons, and the cables. Each within the network, as opposed to processing node was comprised of the Imote2 (including at a central location. By implementing data on-board CPU, radio, power management processing techniques (e.g., modal analysis integrated circuit), the ISM400 sensor board, Figure 4: Example of wireless smart sensor nodes or damage detection algorithms) in such a and a battery (Figure 2). Measurements deployed on the Jindo Bridge. distributed manner, the amount of communi- included 3-axes acceleration, plus tempera- cation that occurs within the network can be ture, humidity, and light. Combined with This year, a wireless strain measurement reduced, while still providing usable informa- the magazineISHMP Services Toolsuite, these powerful became available with the newly developed tion on the structural condition.S The TIllinois Rnodes allowU for synchronized C dataT collection, U SHM-S R sensor Eboard. All 113 nodes are self- WSS framework provides a cost-effective and aggregation, synthesis, and decision-making powered using solar or wind harvesting. The scalable solution that is revolutionizing struc- in real time. sensors use both single-hop and multi-hop tural health monitoring. The smart sensor nodes for the deck, pylons communication protocols to interact with two and cables were attached using different meth- base-station computers, which are remotely Wireless Smart Monitoring ods. The nodes were mounted to the steel accessible via the Internet. Should any anoma- System on the Jindo Bridge deck and pylons using two magnets, each lies in the measured data be detected during the with a holding capacity of 10 kg, attached to autonomous system operation, the base-station A joint effort between the University of the bottom of the enclosure. The nodes were computers automatically email the research Illinois at Urbana-Champaign, KAIST and mounted to the cables using two U-bolts and team so that appropriate action can be taken. National University in Korea, and the an aluminum mounting plate. These methods The monitoring system can autonomously University of in was undertaken of mounting the sensors have proven to be estimate various physical states of the bridge. in 2009 to demonstrate the efficacy of this fast, inexpensive, and secure. For example, the modal properties (i.e., wireless smart sensor framework, resulting in Based on the success of the 2009 deploy- natural frequencies, modal damping, and the first autonomous, full-scale implementa- ment, the monitoring system was extended in mode shapes) can be obtained from the mea- tion of a wireless smart sensor network for 2010 to include 113 WSS nodes, measuring sured accelerations and utilized to refine the structural health monitoring on the Jindo a total of 659 channels of data, (Figures 3 numerical model, to determine structural per- Bridge in . and 4), resulting in the world’s largest wire- formance, and to find locations of possible Opening in 1984, the 1st Jindo Bridge is a less smart sensor network for SHM. The fatigue damage. Cable tension force, one of the cable-stayed bridge connecting the Korean ISM400 sensor board is used on 100 nodes. most important integrity measures for cable- peninsula and the Jindo Island; however, A new high-sensitivity accelerometer board stayed bridges, is estimated automatically using growing traffic demands quickly exceeded (SHM-H board) developed by the Illinois a vibration-based method. Deck-cable interac- the load carrying capacity of the first bridge. research team, which enables measurement tion, which may cause dynamic instability, also The nd2 Jindo Bridge, which opened in 2005, of accelerations as low as 0.05 mg, is used can be assessed. Aerodynamic and aeroelastic is a streamlined steel box girder with a center for 10 nodes. The remaining three nodes are properties of bridges are estimated based on the span of 344 meters (1129 feet) supported by connected to 3D ultrasonic anemometers to synchronized wind speed and response data. 60 parallel wire strand cables anchored to two measure and collect, wirelessly, the speed and This dense information enables comprehensive diamond-shaped pylons. direction of the wind on the bridge. monitoring of the bridge’s health. continued on next page STRUCTURE magazine13 October 2011 Fig 5

Typhoon Kompasu Hits Jindo 30 20 In September 2010, Typhoon Kompasu hit Haenam 10 the Korean Peninsula with sustained winds of (Inland)

54 m/s (195 km/h or 121 mph). The Korea (m/s) Speed Wind 0 Meteorological Administration (KMA) station 0 200 400 600 800 1000 at Jindo Island measured wind speeds of 15-20 250 m/s (or 49 to 6 ft/s) out of the SSE (Figure 5). 200 The measured wind by the wireless monitor- ing system was 18-26 m/s from the southeast, 150 slightly faster than reported by KMA station 100 Azimuth Dir. (Deg) Dir. Azimuth and likely due to the local topography. 0 200 400 600 800 1000 During Kompasu, the vertical acceleration 40 of the deck exceeded 20 mg, which is greater 20 Jindo than the level of acceleration generated by a 0 40-ton truck (Figure 6). Nonetheless, these -20 Fig 6® accelerations were below the limit of 50 mg (Deg) Elevation -40 proposed in the Korean Design Guidelines of 0 200 400 600 800 1000 Time (sec) Steel Cable-Supported Bridges. Figure 5: Wind measurements on the bridge during Typhoon Kompasu. The modal properties of the deck and pylons were obtained using the acceleration data. Utilizing an output-only modal analysis Copyright approach, nine vertical bending modes and one torsional mode were clearly identified 5 from the vibration induced responses mea- sured during the typhoon (Figure 7 and 8). Compared with the identified modes using the data from an existing wired high-precision Fig 7 accelerometer on the bridge, the current SHM system is found to provide highly accurate modal properties. The wind velocity fluctuations were mea- sured with ultrasonic anemometers installed magazine along the bridge. This collectedS data Tenables R U C T U R E the estimation of spatial as well as tempo- ral variation of the wind velocity along the Figure 6: Acceleration response due to 40-ton truck and Typhoon Kompasu. bridge. The steady wind load coefficients and flutter derivatives were also obtained from a 0.45 104 1.03 PSD (Haenam side deck) series of wind tunnel tests using a 1/36 scaled 0.66 1.87 z-axis 6 1.56 model of the deck section. The buffeting 2.26 1.36 1.65 2 2.81 analyses results considering aerodynamic 10102 2.41 admittance effects show good agreement to the measured responses (Figure 9). This effort 0 offers a significant enhancement in evaluating 10100 the aerodynamic safety and serviceability of bridges during strong winds by using synchro- -2-2 nized simultaneous measurement of wind and 1010 acceleration data along the bridge.

-4 The tensions in the bridge’s stay cables were 1010-4 obtained autonomously using vibration data, 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 freq(hz) 2.5 3 3.5 4 4.5 5 combined with the known cable properties (i.e., Frequency (Hz) length, weight, stiffness, etc.). In this approach, the vertical acceleration was used primarily to Figure 7: Power spectrum for bridge responses during Typhoon Kompasu. estimate the tension, while longitudinal and lateral accelerations helped to delineate the values and have changed little from the 2008 is assessed periodically, including the charging modes unique to the cable (as opposed to the inspection, which confirms the integrity of the current from the solar panel and the battery deck and pylon). Estimated tensions during bridge. Thus, continuous monitoring of this voltage (Figure 11). Should the battery voltage the typhoon were compared with the design important health indicator is facilitated. in a specific node become too low (e.g., due to tensions and those obtained in the previous In addition to continuously monitoring the too many cloudy days), the node is put into a routine inspection. As shown in Figure 10, the bridge structure, the WSS network monitors deep sleep until the batteries can be charged current cable tensions are close to their design itself. For example, the battery power system and the node brought safely back online. 7

STRUCTURE magazine14 October 2011 STRUCTURE magazine Fig 8 Fig 9

DV1: 0.4462Hz DV2: 0.6471Hz DV3: 1.0326Hz Fig 10

DV4: 1.3421Hz DV5: 1.5490Hz DT1: 1.8022Hz

DV7: 1.8704Hz DV8: 2.2609Hz DV9: 2.8133Hz

9 Figure 8: Identified mode shapes below 3 Hz. Figure 9: RMS acceleration from buffeting analyses and measurement. ® 8 Copyright

Fig 11 Figure 10: Comparison of estimated cable tensions. magazine Conclusions and Future Work East-side nodes S T R4.25 U C T U R E Smart sensing technology for structural 4.2 health monitoring is an important field that is coming of age. Combining civil engineer- 4.15 151 96 4.1 ing knowledge with developments in sensor 10116 technology and network/information man- 4.05 144 agement has provided a solution that is a Battery Voltage (V) 64 robust and significantly lower-cost (approxi- 4 135 mately $100 per channel) alternative to 150 traditional wired monitoring systems. Indeed, East-side nodes 132 39 200 WSS provides an important new tool to help 2 engineers address the many challenges of 150 129 managing civil infrastructure. Finally, a joint 85 100 effort between MEMSIC and the University 52 of Illinois is currently underway to develop 50 the next-generation wireless smart sensing 0 Charging current (mA) Charging nodes for structural health monitoring. -50 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 For more information on the develop- Time (o'clock) ment and implementation of the smart sensor network, visit the ISHMP site at Figure 11: Monitored battery voltage and charging current. http://shm.cs.illinois.edu.▪ Grant NRF-2008-220-D00117 in Korea, appreciation the following individuals for their Acknowledgements as well as the technical and administrative careful review of this manuscript: Gul Agha, 11 cooperation regarding Jindo Bridge testing Yozo Fujino, Shinae Jang, Hongki Jo, Hyung-Jo The authors gratefully acknowledge the sup- by the Iksan Construction and Management Jung, Jongwoong Park, Ho-Kyung Kim, Robin port of National Science Foundation Grants Administration, Korean Ministry of Land, Eunju Kim, Jian Li, Kirill Mechitov, Parya CMMI-0928886 and CNS-1035773 in the Transportation, and Maritime Affairs. The Moinzadeh, Tomonori Nagayama, Jennifer USA and National Research Foundation authors also would like to express their Rice, and Chung-Bang Yun.

October 2011 STRUCTURE magazine15 October 2011 References ASCE Infrastructure Report Card; www.infrastructurereportcard.org. Mehmet Çelebi (2006). “Real-Time Seismic Monitoring of the New Cape Girardeau Bridge and Preliminary Analyses of Recorded Data: An Overview.” Earthquake Spectra, Vol. 22, pp. 609-630. doi:10.1193/1.2219107. Pakzad, S.N., Fenves, G.L., Kim, S. and Culler, D.E. (2008), “Design and Implementation of Scalable Wireless Sensor Network for Structural Monitoring”, J. Infrastruct. Syst., 14(1), 89-101. S.A. Jang, H. Jo, S. Cho, K.A. Mechitov, J.A. Rice, S.H. Sim, H.J. Jung, C.B. Yun, B.F. Spencer Jr., and G. Agha (2010). “Structural Health Monitoring of a Cable-stayed Bridge using Smart Sensor Technology: Deployment and Evaluation,” Smart Structures and Systems, Vol. 6, No. 5-6, pp. 439-459. ® H. Jo, S-H. Sim, K. A. Mechitov, R. Kim, J. Li, P. Moinzadeh, B. F. Spencer, Jr., J. W. Park, S. Cho, H-J. Jung, C-B. Yun, J. A. Rice, and T. Nagayama (2011). “Hybrid Wireless Smart Sensor Network for Full-scale Structural Health Monitoring of a Cable-stayed Bridge.” Proceedings of SPIE 7981, 798105, San Diego, CA, USA. Copyright Korea Society of Civil Engineers (2006). “Korean Design Guidelines of Steel Cable-Supported Bridges,” Haklimsa, Seoul, Korea (in Korean). Park, Y. S., Choi, S. M., Yang, W. Y., Hong, H. J., & Kim, W. H. (2008).” A study on tension for cables of a cable-stayed bridge damper is attached,” Journal of KSSC. Vol. 20, No. 5, pp. 609-616 (in Korean).

S T Rmagazine U C T U R E

STRUCTURE magazine16 October 2011