An Internet observatory: remote monitoring of ...

Home

Search

Collections

Journals

About

Contact us

My IOPscience

An Internet observatory: remote monitoring of instrumented civil structures using the information superhighway

This content has been downloaded from IOPscience. Please scroll down to see the full text. 1995 Smart Mater. Struct. 4 14 (http://iopscience.iop.org/0964-1726/4/1/003) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 132.198.40.183 This content was downloaded on 03/08/2015 at 00:34

Please note that terms and conditions apply.

Sinaft Mater. Sbuct. 4 (1995) 14-19. Printed in the UK

An Internet observatory: remote monitoring of instrumented civil structures using the information sUperhighway ~~

~

~

Peter L Fuhr, Dryver R Huston, Timothy P Ambrose, and Euan F Mowat University of Vermont, Intelligent Structural Systems Research Institute, College of Engineering, Burlington, VT 05405,USA Received 26 April 1994, in final form 25 November 1994 Abstract. The requirements of sensor monitoring associated with instrumented civil structures pose potential logistical constraints on manpower, training and costs. The need for frequent or even continuous data monitoring places potentially severe constraints on overall system performance given real-world factors such as available manpower, geographic separation of the instrumented structures, and data archiving as well as the training and cost issues. While the pool of available low-wage, moderate-skill workers available to the authors is sizable (undergraduate engineering students), the level of performance of such workers is quite variable leading to data acquisition integrity and continuity issues-matters that are not acceptable in the practical field implementation of such developed systems. In the case of acquiring data from the numerous sensors within the civil structures which the authors have instrumented (e.g., a multistory building, roadway/railway bridges, and a hydroelectric dam), we have found that many of these concerns may be alleviated through the use of an automated data acquisition system which archives the acquired information in an electronic location remotely accessible through the Internet global computer network. It is therefore possible for the data monitoring to be pelformed at a remote location with the only requirements for data acquisition being Internet accessibility. A description of the developed scheme is presented as well as guiding philosophies.

1. Introduction

The recent advent of advanced materials with internal sensory and reaction capabilities, i.e. 'intelligence', has opened the door for many useful applications in civil Structures, e.g. bridges, buildings, dams, in contrast with aerospace structures. Structures built with intelligent materials will be able to sense external loads, assess the internal stresses and damage caused by the loads, and, if necessary, respond appropriately. Although relatively few structures are currently built with such materials, the potential market for the application of these materials can be quite large [I]. The most probable candidates for the initial application of these materials will be highperformance structures such as skyscrapers, including intelligent buildings, bridges, and those facilities where failure presents large safety concerns, such as nuclear waste containment vessels, bridge decks, dams, and structures under construction. Large-scale structures that have actually had fiber optic sensors embedded into them are still relatively few. Holst 0964-17'26/95/010014+06519.50

8 1995 IOP Publishing Ltd

and Lessing [Z] have installed fiber optic displacement gauges in dams to measure shifting between segments. Fuhr et a1 [3] report the installation of fiber optic sensors into the Stafford Building at the University of Vermont. Similar installations have been completed at the Winooski One dam in Winooski, VT [4], in the newly constructed physics building on the campus of the Dublin City University in Ireland [SI, and in a railway overpass bridge in Middlebury, VT [6]. Researchers at the University of Toronto, in conjunction with City of Calgary (Alberta, Canada) officials, have embedded fiber optic strain gauges into the prestressed concrete support girders on a highway overpass [7]. A major challenge to building and using a comprehensive sensor system is that a variety of physical parameters may be measured using sensors that arc often spatially distributed across the span of a large structure. The sensing and data processing requirements are complicated by most of the measurands exhibiting dynamic behaviors with timescales that vary over several orders of magnitude. In the case of an instrumented civil structure questions re-

An Internet obsewatoly garding the data acquisition and frequency of monitoring of the embedded sensors arise. An additional concern is that modem data acquisition hardware collects enormous amounts of data. only a small fraction of which may be useful. In parallel, labor concerns ranging from training of sensor ‘inspectors’ andlor ’interpreters’ who actually performs the processing and analysis of the acquired data are legitimate matters in the practical field implementation of such developed systems. Such questions have arisen in the routine monitoring of acquired information from the sensors embedded into the Winooski One hydroelectric dam. The dam spans the Winooski River in Winooski, VT. The hydroelectric powerplant consists of three parallel turbogenerators yielding a rated power output of 8.5 MW. The plant is a ‘run-of-the-river’ type in that only minimal levels of water are stored upstream of the dam for power generation. The water rate that is sent through the turbines is therefore highly variable, and depends upon river flow volumes and regulatory constraints. As a result, the turbogenerators are routinely being brought into and out of service. This creates a highly variable dynamic state of the structure, which is amenable to vibration and structural monitoring efforts. The physical placement of the sensors was based on a detailed structural analysis of the architectural drawings of the dam and powerhouse. Given the limited number of sensors to be used (limited ‘due to fiscal realities), key structural areas were chosen for structural health monitoring. In certain instances, it was determined that simple crack detection was of prime importance (e.g., near the wall anchorages) while in other areas vibration analysis was of principal importance (e.g., in the vicinity of the generator turbines) or intrusion detection was deemed most important (e.g., entrance walkway). In other instances the site and application specifics dictated the type of fiber sensor (and type of fiber) which was installed at those locations. In deciding the type of sensor to be embedded, robustness was heavily weighted since certain other attributes of the sensing system would be negated by that system’s high degree of structural failure or maintenance requirements. The resultant sensor ‘web’ embedded into the powerhouse is composed of 74 sensors with 17 associated access locations placed non-uniformly throughout the four story 30 000 ftz structure.

2. Data acquisition from instrumented structures The concern regarding acquisition of data from embedded and surface-attached sensors is magnified when the sensors are located throughout a potentially large instrumented civil structure. Direct cable interconnection between a centralized data-acquisition location and the sensor may quickly become impractical from logistical, aesthetic, and, in certain cases. costing perspectives. As such, two issues predominate our activities with respect to sensor information acquisition: (1) collection of the sensors’ data at a single central location within (or near) the instrumented structure: and (2) development of a method for remote monitoring of this sensor information.

With respect to issue 1, we have proceeded to use radio telemeuy for data acquisition and concentration at a single central location [SI. In the context of using existing cordless/cellular telephone components, their frequency passband is optimized for low-frequency audio transmission. While the frequency response of such systems may not be directly suitable for the very low-frequency resonances typically found in large civil structures, the sensed signal may be readily frequency up-shifted using standard subcarrier modulation (or other) techniques. We have demonstrated use of a radio telemetry system for the transmission of such frequency up-shifted fiber optic sensed vibration data. Error levels were found to be less than 0.5%. A range of over 300 feet was easily demonstrated with these off-the-shelf components. Range increases may be as easy to achieve as simply amplifying the signal and using better transmit and receive antennas. Using spread spectrum techniques, as well as low-power emissions, it is envisioned that reasonably low-cost commercially available transceiver devices will provide improved wireless data acquisition from embedded sensors into a central location. The requirements associated with issue 2: ‘development of a method for remote monitoring of this sensor information’ are currently being addressed through the use of system software resources associated with the Internet global computer network. The Internet is actually a collection of over 30000 computer networks forming one of the world’s most comprehensive information sources. Unlike commercial networks, the Internet has no central authority, administrators have only agreed that all participating computers use the Transmission Control Protocolfinternet Protocol (TCP/IP) communications language. This .vast network was originally intended to serve the academic and research communities, but has grown to link universities, research and commercial sites to form an enigmaiic web OF systems that can be accessed in a variety of ways. The lack of Internet central administration makes navigation of the network difficult, resulting in the development and accessibility of a vast array of network software utilities that manage and control network node-to-node communications. These resources present several methods by which instrumented structures’ sensor information may be acquired. We have used simple terminal emulators (e.g., Kermit, Versatenn, Novaterm) for retrieving the sensor information from archive accounts on an Internet mainframe computer. While successful, this is a cumbersome method. We have found that for simple data archiving and subsequent remote locale retrieval, the use of an ‘anonymous’ file transfer protocol (FTP)site is optimal. In order to use FTP we have established an FTP server, [email protected], which serves as the Internet location where the sensor data from the Winooski One hydroelectric dam is stored. In addition, we use the accessinghploading-downloadingsoftware application ‘Fetch’ for data storage and subsequent retrieval. Fetch is basically a graphical front-end application used to simplify Internet downloading and uploading of files/information using FTP (Fetch is free to universities and $25 to others. from [email protected]). Alternatively the Unix-based 15

PL Fuhr et al

Point#

Intensity (V)

0 1 2 3 4 5

Frequency (mHz)

6.1583e-05 3.3713e-08 8.6179e-08 3.6841e-08 1.6613e-07 3.1326e-08

0 12 24 36 48 60

•

••

Figure 1. FTP-accessed embedded sensor data obtained from the Winooski One dam (only a small subset is shown).

program FTP may be executed directly with the sensor data stored and retrieved using the FTP 'get' and 'put' commands. In either case, raw time history data from each embedded sensor is archived on an Internet networked computer in a file (or folder) that may be accessed using FTP. Information security may be obtained by requiring a password to access this data, or, complete world-wide Internet access may occur if the sensor data is archived in an anonymous FTP accessible file/folder, as we have done at Internet site Win [email protected]. In a manner somewhat similar to the FTP site, the use of the Internet network utility program 'Mosaic' has been chosen for the visual display of video/audio frames acquired from video cameras positioned at the Winooski One Dam. Mosaic is an Internet-based global hypermedia browser that allows the user to discover, retrieve, and display documents and data from all over the Internet. It is part of the World Wide Web project, a distributed hypermedia environment originated at CERN and collaborated upon by a large, informal, and international design and development team. Mosaic, which can be copied for free from many Internet software archives. organizes information into a so-called 'home page'. which serves as a table of contents about a particular topic or group of topics. It differs from FTP in that while FTP is used for the storage and retrieval of ASCII/binary data files, Mosaic provides a graphical front-end method of literally viewing scanned photographic and video images and hearing associated recorded sounds. Using Mosaic, the server is accessed in a manner similar to Fetch/FTP, but in addition the file you wish to view at the Mosaic server site is also accessible (in this particular case the Mosaic server site address is: issri.emba.uvm.edu).

3. Internet monitoring of the instrumented Winooski One Dam The Internet has been used for the monitoring of sensors installed at the Winooski One hydroelectric dam. Centralized data collection of the fiber optic sensors 111stalled within the superstructure of the dam occurs on site with the resultant data transferred to an FTP-accessible partition on an Internet accessible computer. This information may subsequently be accessed/downloaded 16

10 -5 ,-...

>

.._., >. ...... .....

10 -6



.....0Cd

rX

10 -7

10 -8

l l. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 l !~ Gear Mesh ! ·~

1~-

t ........... t....... Freq. (175 Hz)

...... i

111111111·1111111

cq

0

1000

~~~~ Frequency (Hz)

Figure 2. Power spectrum of the data shown in figure 1.

to an 'Observer's' home location using Fetch/FTP or a terminal emulator ASCII file transfer program. The observer may then locally process the information. As an example, a portion of the raw time-history intensity data obtained from fiber optic vibration sensor #16, located within turbogenerator #3's support mount, is shown in figure 1. The associated power frequency data is shown in figure 2. This data was sensed using embedded fiber optic vibration sensors, locally collected, and transferred to the Internet host computer (a Sun Microsystems computer at the University of Vermont) where it was placed in the appropriate partition (the '[email protected]' anonymous FTP data site). The data was then accessed remotely using a modem connected to a Macintosh LC2 microcomputer. Data processing and interpretation occurred using the LC2. In conjunction with the monitoring of the status of the embedded sensors is the need to concurrently monitor the electric power generation status as well as the water level and flow rates. Two methods are used for this data acquisition: (I) traditional analog-to-digital (AID) conversion of the electrical values associated with each turbogenerator's water flow rate, turbine rotation rate (RPM), generation efficiency and generator output electrical power; and (2) simple TV camera video capture of the operator's consoles associated with each turbogenerator. In the first

An Internet ObSeNatOly

case, the acquired individual voltage levels are interfaced to a unique channel on a microcomputer's ND board, sampled, scaled, time-tagged and stored in memory for transmission to the Internet host computer (similar to the fiber-optic-sensed data). This information is then formatted on the Internet host and placed in the same anonymous FTP account but under a different filename. In the second case, the acquired video frames ( I for each console display plus the outside camera) are transmitted to the Internet host computer's 'Winl-video' Mosaic file for worldwide Internet access and viewing using Mosaic or a comparable video display program. From the Internet access perspective, while it is unimportant which type of computer is used as the Mosaic server (only full bandwidth Ethernet connections are required), the Win] Mosaic Site (issri.emba.uvm.edu) is physically located on another Macintosh LC2 microcomputer. I,, the cOntext of instrumented civil StrUctUreS information, we have configured a multimedia (video/audio) equipped Power Macintosh 7100 microcomputer for acquiring and preprocessing of the video images obtained from the cameras located at the instrumented structure, in this case, the Winooski One hydroelectric dam. Near realtime video images are placed into the appropriate Mosaic source location home page. All video images are not archived, due to limited physical disk space, but rather the last 30 frames are scquenced through a storage buffer. Viewing of the video frames is achieved through the use of Mosaic on any Ethernet-configured Internet-accessible (IP-Internet protocol addressed) microcomputer. The images shown as figures 3 and 4 represent the Mosaic video viewing of the Winooski One dam on April 6, 1994 at 2 2 0 pm EST. Specifically, figure 3 shows the outside video view of the water flowing over the dam at this time. Figure 4 shows the associated video frame of the camera's view of the control monitor for turbogenerator number 2. From this frame the operational values and river level and flow rate data of this generator are readily apparent (please note that the full-screen-sized video frames have been reduced for this publication).

4. Summary The requirements of sensor monitoring associated with instrumented civil structures pose potential logistical constraints on manpower, training and costs. For example, the need for frequent or even continuous data monitoring places severe constraints on the mere monitoring process when faced with real-world factors such as available manpower, geographic separation of the instrumented structures, and data archiving. In many instances, the mere instrumentation of a structure does not address issues such as: 'Who will acquire the data?'; 'Who will interpret the data'?'; 'Is specialized training necessary for some onsite individual?'. Such questions directly raise the very real issues regarding the eventual acceptance and usage of instrumented structures. We believe that many of these concerns may be reduced or alleviated through the use of an automated data acquisition system which archives the acquired information

~i~~~~ 3. Internet.accessed, ~ ~ ~ ~ i ~ . d i s p realtime layed 'snapshots' from an outside deck video camera showing the water flow level at the Winooski One hydroelectric dam. The actual display has been reduced from its full screen size for this publication.

Figure 4. Internet-accessed, Mosaic-displayed realtime video camera 'snapshots' of the powerhouse control console associated with turbogenerator #2. The actual display has been reduced from its full screen size for this publication.

in an electronic location accessible through the Internet. It therefore becomes possible for an individual or service team to monitor the instrumented structures from a remote site. Two methods of sensor data storage and retrieval have been demonstrated are currently in use: ( I ) direct accessing of the time-tagged sensor readings is achieved using FetchATP file transfers from an 'anonymous' ETP site where the Winooski One Dam's embedded sensor values are archived; ( 2 ) Mosaic videolaudio accessing of stored near realtime video images, and associated recorded sounds, producing a playback method which allows any user anywhere on the Internet to obtain a current view of the dynamic status of the dam. While questions regarding the reliability of the Internet arise, if similar data acquisition and monitoring strategies are utilized at other instrumented structures, it becomes possible for a single remotely located observer to 'visit' each structure and obtain the requisite sensor readings. 17

P L Fuhr et a/

Figure 5. Diagramatic view of the information and connection paths used for remote monitoring of the Winooski One hydroelectric dam.

Appendix A. An introduction to the Internet-the World Wide Web and Mosaic As prcviously stated, the Internet represents a vast conglomeration of individual computer networks and svand-alone computers interconnected through a common communication protocol structure. As such, the Internet, with n o centralized structure, represents the best and worst worlds of freedom of aCceSS and eaSe of use, Groups withi,, the I~~~~~~~ community estah,ish ,,,orking groups with SOmctimeS specific software used withi" that particular group for adapting the loosely frameworked Internet for their particular use-typically for ease of information sharing. One such group, the World Wide Weh (WWW), was conceived and developed at CERN (the European Laboratory for Particle Physics) in the late 19x0s supporting the premise that the vast cadre of researchers interested in CERN research would he hctter served via a more organized informatinn exchange of CERN rescarch. The growth if the Internet usage of late has paralleled the growth in the use of a particular WWW application software program, Mosaic. Mosaic is known as a graphical 'browser' program-it ailows a combination of tCxt and graphics. such as scanned photographs, graphs or movies, to he viewed across the Internet. There are currently Over 10000individual ~~~~i~ sites providing information such as online encyclopedias, college catalogues and campus descriptions and telescope 18

images of celestial hodies. Mosaic was developed under federal government sponsorship and is therefore free. However, the usefulness of this program has spawned the development of numerOuS commercially available browser programs (most cli which have a 'feel' similar 10 Mosaic's). As such these programs are also able to access the multitude of Mosaic sites. such as the site we use to provide access to our smart civil structural information, with certain program specific advantages over the free version of Mosaic. Versions of Mosaic exist for computer platforms such as Macintosh, Windows. Unix. and X-Windows machines. The latest version of Mosaic can he found at any of [he popular gopher sites (such as oac.hsc.uth.tmc.edu or ncsa.uiuc.edu), and is exlremely easy tO download using the Macintosh application Turbo G ~ Once~the lateSt ~ version of Mosaic is found a~ one of the siteS. all you need do is click on Mosaic twice, and Turbo Gopher will download it to your disk. It will appear on your disk in archive form (either CompacPro, or Stuffit archives) and can be expanded using the appropriate application. (For further information, both general and specific, regarding Intcmet, its services and applications, please refer to the growing number of popular press books or the himonthly magazine Internet World.)

~

An Internet observatory

Appendix B. Specific use of the Internet for data retrieval While a general framework for the use of the Internet for data storage'and retrieval has been described, the user wishing to access the data mentioned in this article may use the Internet retrieval application FTP. FTP (file transfer protocol) runs on all computers with Internet access. A Help file resides within FTP, thus for detailed questions regarding FTP usage, merely launch the application on your host machine and type '?' in response to the FTP prompt. The Winooski One hydroelectric dam's sensor information is continuously updated into the storage computer disk folder (partition) entitled, [email protected]. Any Internet user may use FTP to access the [email protected] folder through logging on to the University of Vermont (UVM) host computer as a guest (FTP> rlogin WinlDam~emba.uvm.edu).The user may then wish to obtain a directory of the files available for their downloading via the FTP command (FTP> dir). Files are then transferred from this W M anonymous FTP site to the user's host computer via the GET command (for example, FTP> get Winldaily-vibe-data). After transfer the user may acquire additional data or leave the site by simply quitting the FTP application. The near-realtime video images (movies) and additional Mosaic information referred to in this article are Found at the Internet Mosaic WWW site: issri.emba.uvm.edu. (The associated electronic IP address of the server is 132.198.8.151.) The Mosaic user need only 'Open' that URL site for access to the site. An introduction is then shown to allow the user to choose between the services that are available. It is recommended that the user wishing to view the near realtime video images have full-bandwidth Internet access (video images have a lot of information that needs to be transferred from this Mosaic site to the user's host machine for subsequent display). A depiction of the computer connections used in this Internet application discussed here is presented as figure 5.

University (CU), represents an example of video conferencing software which allows many Internet users to place video and audio information into a single reflector (or satellite) location thereby making all of it available to each respective user. We have found that CUSeeMe provides an excellent mechanism for the viewing of instrumented structures. As with Mosaic, the Internet connection requirements are full bandwidth Ethernet, thereby allowing video frame refresh rates to range from I to 20 frames per second; this variation depends on local host computer parameters as well as the number of CUSeeMe participants that you are viewing. The local computer must be multimedia (audiohideo) equipped. While the current major limitation is that the video frames are not stored, and hence, as in traditional over-the-air television transmission, once something interesting is shown it is gone, it is anticipated that there will be some sort of recording option added to CUSeeMe. When used in conjunction with FTP data storage and Mosaic graphical viewing, CUSeeMe provides an excellent realtime information viewing mechanisim. Undoubtedly, there will be other similar programs available in the future most probably also allowing the use of the Internet for remote monitoring of instrumented stcuctures.

References I11 Huston D R 1991 Smart civil structures-an overview Proc. SPIE Fibers '91 Symp. (Bosron MA. 1991) pellingham, WA: SPIE) paper No 1588-21 [21 Holst A and Lessing R 1992 Fiber-optic intensity-modulatedsensors for continuous observation of concrete and rock-fill dams Proc. 1st European ConJ onSmart Strucrures and Marerials (Glasgow, 1992) ed B Culshaw, P T Gardiner and A McDonach (Bristol: Insitute of Physics Publishing) p 223 131 Fuhr P L. Huston D R, Kajenski P I and Ambrose T P

1992 Performance and health monitoring of the Stafford Medical Building using embedded sensors J. Smnrt Marer. Struct. 1 6 3 - 8 [41 Fuhr P L and Huston D R 1993 Multiplexed fiber optic pressure and vibration sensors for hydroelectric dam monitoring J. Smart Mater. Srruct. 2 320-25 McCraith B (Dublin Citv Universitvl 1993 oersonal communication to P L Fuhr Huston D R. Fuhr P L and Ambrose T P 1994 Intellizent civil structures-activities in Vermont J. Smart Mater. srrucr. 171 Measures R er a1 Strain gauge measuremen&obtained from an instrumented bridge girder Smart Malerials and ,I

Appendix C. Late breaking news With the Internet usage currently increasing at an exponential rate, there are many software applications that are being continually improved or newly developed. In addition to the FTP and Mosaic data observing methods which we use, and are described in this paper, we are also using another free program for the graphical viewing of sensor measurements, namely CUSeeMe. CUSeeMe, developed at Cornell

Structures Canj 2 (Orlando, FL, 1994)

Fuhr P L. Huston D R and Ambrose T P 1993 Interrogation of multiple fiber sensors in civil Structures using radio telemetry J. Smarr Marer. Slruct 2 264-69

19