Research Article Measurement Model for the Maximum Strain in Beam Structures Using Multiplexed Fiber Bragg Grating Sensors

Transcription

1 Distributed Sensor Networks Volume 2013, Article ID , 9 pages Research Article Measurement Model for the Maimum Strain in Beam Structures Using Multipleed Fiber Bragg Grating Sensors Se Woon Choi, 1 Jihoon ee, 1 Bo Hwan Oh, 2 and Hyo Seon Park 1 1 Department of Architectural Engineering, Yonsei University, 134 Shincshon-dong, Seoul , Republic of Korea 2 Daewoo Institute of Construction Technology, 60 Songjuk-dong, Suwon , Republic of Korea Correspondence should be addressed to Hyo Seon Park; hspark@yonseiackr Received 12 July 2013; Revised 6 September 2013; Accepted 8 September 2013 Academic Editor: Hong-Nan i Copyright 2013 Se Woon Choi et al This is an open access article distributed under the Creative Commons Attribution icense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited This study develops a strain measurement model for beam structures subjected to multiloading conditions by defining the strainshape function and participation factors to overcome the limitations of strain measurements using fiber Bragg grating (FBG) strain sensors Using the proposed model, the maimum strain in a beam is obtained by the sum of the strains caused by the different loadings acting separately In this paper, the strain-shape functions for various loading and support conditions are provided, and a system of equations is defined to calculate the participation factors Furthermore, the influence ratio is defined to identify the influence of each loading on the value of the total strain The measurement model is applied to the monitoring of the maimum strain in a 4 m long steel beam subjected to two concentrated loads For measurements during the test, seven s and nine electric strain gauges (ESGs) were attached on the surface of the bottom flange The eperimental results indicate a good agreement between the estimated strains based on the model and the measured strains from ESGs Furthermore, the dependency of the locations for the s installed at the beam structure on the selection can be avoided using the measurement model 1 Introduction Structural members in buildings or infrastructures eperience various loads, such as gravity-induced loads, earthquake,wind,orunepectedloadstosecurethesafetyofa structure, the maimum stresses in the members due to various loads must not eceed the allowable stress of a member [1, 2] For this reason, strain-based structural health monitoring (SHM) has been widely used to assess the structural states of the members by sensing the maimum stresses [3 8] Various types of point sensors, including fiber optic sensors and electric strain gauges (ESGs), are applied to measure the maimum stresses However, point sensors used to measure strains can cover only a relatively small range of structural members because they can measure the strain only at a local point of a member Thus, many difficulties eist when determining the maimum stress in a member with point sensors because the actual strain distribution of a member is nonuniform In this case, the reliability of the safety of a member based on the measured maimum strain depends on the number and location of the sensors However, when attempting to overcome these limitations, the number of sensors may not be increased to cover the entire length of a member because of practical problems related to maintenance and installation Varioustechniquestoestimatethemaimumstrainbased on the measured strains have been reported [5, 7]becausethe measured strains from point sensors cannot be directly considered when evaluating the safety of a member To cover a relatively long length of a member, the average strains from long-gauge fiber optic sensors (GFOSs) or vibrating wire strain sensors (VWSGs) have been used to estimate the maimum strains of a member subjected to a single loading condition However, the techniques are rather impractical because structural members in real building structures or infrastructures are subjected to multiple loading conditions Considering the limits of point sensors and long-gauge sensors, fiber Bragg grating (FBG) strain sensors are suitable for measuring the maimum strain of a beam Although an is a point sensor with a relatively short-gauge length, it can measure the maimum strains with a minimum

2 2 Distributed Sensor Networks Downward fleural deformation Compression Interrogator Bragg grating Neutral ais Tension Elongation Interrogator Cladding Fiber core Broadband light source Bragg grating ight with λ B λ B =2n e Λ Λ Figure 1: Measurement principle of the strain using an number of sensors because of the multipleing technology in FBG [9] Multipleing is a method by which multiple signals are combined into one signal, and only a single cable suffices for a certain number of sensors In addition, FBG strain sensors have high resolution and are convenient to install compared to long-gauge sensors However, to date, no research regarding point sensors, including s, has been reported on the measurement or sensing method for the maimum strains in a structural member subjected to real multiple loading conditions Therefore, in this paper, a maimum strain measurement model based on point sensors is developed for the to be applied to find the maimum strain in a steel beam structure subjected to multiple loading conditions In addition to the maimum strain, a distribution of the strains along the length of a member is provided by the measured local strains from the minimum number of s To evaluate the performance of the model, an eperiment has been conductedona4mlongsteelbeamtocomparethemaimum strain directly obtained from ESGs and the maimum strain from the model with the strains measured from s 2 FBG Sensors In s, a Bragg grating is introduced to the core of a fiber[10] If the broadband light source collides against the Bragg grating, a light with a particular wavelength, called the Bragg wavelength λ B, satisfying the relationship shown in (1) is reflected: λ B =2n e Λ, (1) where Λ is the period of Bragg grating and n e is the effective refractive inde Because the Bragg grating period varies according to the elongation, a strain can be calculated by detecting the change of the Bragg wavelength that reflects an elongation As shown in Figure 1, the longitudinal strain of a beam member subjected to multiple loadings can be measured by FBG Figure 2: Simply supported beam subjected to a linearly varying load typical s attached to the face of the member Especially from the aspect of SHM, the multipleing technology in an makes strain-based SHM more reliable and practical because of the convenience with which such sensors can be managed and installed A variety of applications can be found in the field of SHM, including building, civil, and infrastructure [11 15] 3 Maimum Strain Measurement Model In this paper, the strain measurement model is presented to estimate the maimum strain of a beam structure using local strains measured from s For a beam structure subjected to multiple loading conditions simultaneously, the total strain at a specific point can be found by superimposing the strains due to separate loadings However, to find the maimum strain of a beam structure instead of the strain valueatapoint,thedeformedshapecausedbythemultiple loadings must be defined by superimposing the distribution shape of the strains along the length of a beam for each loading separately The total strain at an arbitrary point in a beam structure can be estimated using the deformed shape 31 Shape Function for the Distribution of Strains Caused by a Single oad Based on general concepts in engineering mechanics, as shown in Figure 2, the longitudinal strain ε() of a beam subjected to a linearly varying distributed load of q

3 Distributed Sensor Networks 3 φ k j (): individual strain distribution to the kth FBG and jth load P 1 P2 1, ε 1 2, ε 2 k, ε k n, ε n P j Total strain distribution, ε t () n ε t () = λ k j φk j () P n 1 P n s P 1 φ k 1 () φ k 2 () φ k j () φ k n () P 2 P j + = + + kth k, ε k k, ε k k, ε k P n k, ε k λ k 1 λ k 2 λ k j λ k 4 Figure 3: Decomposition of the strain distribution caused by multiple loadings into individual effects q can be epressed as a function of the distance from the left-hand support: M () ε () = EZ, (2) where M() is the bending moment, E is the modulus of elasticity, Z is the elastic section modulus, and is the length of the beam For a FGB sensor installed at a length of FBG in Figure 2, the strain measured from the FBG strain sensor ε FBG is epressed as ε FBG = 1 EZ (q 6 FBG q 6 FBG 3) (3) From (3), the intensity of the distributed load q can be epressed as q= 2EZε FBG FBG ( 2 2 FBG) (4) Then, the general form for the longitudinal ε() at an arbitrary point can be defined by ( 2 2 ) ε () =ε FBG FBG ( 2 2 FBG) =ε FBGφ (), (5) where φ() is defined as a shape function for the distribution of strains along the length of the beam The strain-shape function φ() for the distribution of strains in (5) depends on the loading and support conditions of a beam structure For convenience, nine typical shape functions for the distribution of strains in beam structures subjected to various loading and support conditions are derived in Tables 1 and 2 Asgivenin(5), the general form of ε() for a beam subjected to a loading can be defined by the strain measured from anfbgsensoratanarbitrarylocationof FBG multiplied by the shape function φ() for the distribution of strains in Table 1 32 Superposition of Shape Functions Caused by Multiple oadings For a beam structure subjected to n different point loadings of P j (to n) infigure 3, the general form for the total strain ε t () of a beam at a distance from the lefthand support can be epressed as the sum of the strains due to n different point loadings acting separately: ε t () = n λ k j φk j (), (6) where λ k j is the participation factor for the jth strain-shape function φ k j () Thesuperscriptk in (6) isthefbgsensor number selected to calculate the strain-shape functions and the participation factors The calculation of the participation factors is not dependent on the selection of an, as with other types of sensors Thus, the value of the total strain in (6) does not vary with the selection of the used to calculate the factors For the kth selected for the calculation of the factors, a system of n simultaneous equations in (7) canbe epressed in matri notation as follows: ε φ k 1 11 φk 12 φk 1j φ k 1n λ k ε 2 φ k 21 φk 22 φk 2j φ k 1 2n λ { } = d k 2 d { ε j φ k j1 φ k j2 } φk jj φ k jn { } [ d λ k, (7) j d ] { ε n } [ φ k n1 φk n2 { } φk nj φ k nn] { λ k n} where φ k ij =φk j ( i) is the value of the jth strain-shape function at the position of i Then, the participation factors can be found as the solution to (7): {λ k }=[φ k ] 1 {ε FBG }, (8) where {λ k } is a column vector of the participation factors, [φ k ] is a square matri of φ k ij,and{ε FBG} is a column vector ofthemeasuredstrainsthen,usingthetotalstrainobtained by the superposition of n shape functions for strains due to the multiple loadings in (6), the location for the maimum

4 4 Distributed Sensor Networks Table 1: Distributional strain-shape function φ(): simply supported beam oading condition P M FBG FBG ω q Condition of location 0< FBG < 0< FBG < P 0< FBG a FBG a b FBG FBG FBG M a FBG < 0 FBG < 0< FBG 0< FBG φ() ( ) FBG ( FBG ) ( 2 2 ) FBG ( 2 2 FBG), { FBG (0 < a), { a ( ), { b FBG (a ) b, { a( FBG ) (0 < a), {, { FBG (a ) FBG FBG 1 strain caused by multiple loadings can be found by setting the derivative of (6)equal to zero: n d d ( λ k j φk j ()) =0 (9) 33 Eample Case: Simply Supported Beam Subjected to Three Types of oads To measure the strains in the beam structure subjectedtoauniformlydistributedloadf 1 and two end moments F 2 and F 3, three s are attached at three different locations, as shown in Figure Strain-Shape Functions and Total Strain The total strain distribution of the beams in Figure 4 is found by superimposing the strains caused by the three different loadings acting separately If the first is selected to calculate the strain-shape functions and participation factors, the strainshape functions for the three different loadings of F 1, F 2,and F 3 can be found in the first, fourth, and fifth rows of Table 1, respectively The strain-shape functions are given by φ 1 1 () = ( ) 1 ( 1 ), φ 1 2 () =, 1 φ 1 3 () = 1 (10) From (8), the participation factors for the three loadings λ 1 1, λ 1 2,andλ1 3 can be found by { { { λ 1 1 λ 1 2 λ ( 1 ) 1 ( 1 ) } = 2 ( 2 ) } 1 ( 1 ) } [ 3 ( 3 ) 1 ( 1 ) [ ] 1 1 ] 1 ε 1 { } { ε 2 } (11) { ε 3 }

5 Distributed Sensor Networks 5 F 1 F2 F 3 Strain Strain Strain Strain 1, ε 1 2, ε 2 3, ε 3 Strain distribution by F 1 : ε F1 () = λ 1 1 φ1 1 () Strain distribution by F 2 : ε F2 () = λ 1 2 φ1 2 () Strain distribution by F 3 : ε F3 () = λ 1 3 φ1 3 () 3 Total strain distribution: ε t () = λ 1 j φ1 j () 3 FBG strain sensors Figure 4: Distribution of the strain-shape functions for a beam subjected to three different loadings Then, the total strain ε t () at an arbitrary location is found by substituting λ 1 1, λ1 2, λ1 3, φ1 1 (), φ1 2 (),andφ1 3 () into (6): ε t () = 3 λ j φ 1 j () =λ ( ) 1 1 ( 1 ) +λ 2 +λ (12) 332 Maimum Strain Using the total strain distribution along the length of a beam structure in (12), the location for the maimum strain caused by the three loadings can be found by setting the derivative of (12)equaltozero: d d ε t () =λ ( 1 ) λ1 2 +λ =0 (13) 1 1 Then,thesolutionof(13),,isfoundby = 2 λ1 2 1 ( 1 ) λ 1 1 2( 1 ) + λ1 3 1 ( 1 ) λ 1 (14) Therefore, the maimum strain can be obtained by comparing (1) thestraincalculatedbysubstituting into in (12) and (2) the strains at both ends of the beam 333 Simulation of the Eample Case To test the performance of the measurement model, the intensities of the three loadings F 1, F 2,andF 3 acting on the steel beam in Figure 4 are set to 371, 6360, and 6360 knm, respectively For the simulation, H was used for the section of the beam structure, and three s were attached 3, 6, and 9 m from the left-hand end of the beam From the structural analysis for the beam structure, the strains for FBG #1, #2, and #3 were found to be 232, 2666, and 2666 με,respectively Strain (με) Etracted strains ε t () λ 1 1 φ1 1 () ocation (m) λ 1 2 φ1 2 () λ 1 3 φ1 3 () Figure 5: Composition of the total strain distribution Using the values of the strains for FBG #1, #2, and #3, the participation factors of λ 1 1, λ1 2,andλ1 3 in (11)werecalculated as 4866, 3707, and 927,respectivelyThus,thetotalstrain of the beam ε t () in (12) is epressed as ε t () = 3 λ 1 j φ1 j () = (4866) + 15 ( 2 ) 36 + ( 3707) ( )+( 927) ( 3 ) (15) The distributions of the total strains and the strains caused by the three loadings are plotted in Figure 5,whichillustrates that the strain values for FBG #1, #2, and #3 obtained from the structural analyses are identical to those estimated by the measurement model The influence ratio R j is defined in (16) to identify the influence of each loading on the value of the total strain: R j = 0 λk j φk j () d n i=1 ( 0 λk i φk i () d) 2 (16) The influence ratio R j indicates a ratio of the area of the strain distribution caused by the jth loading to the sum of the areas of the strain distributions caused by n different loadings If theinfluenceratior j has a positive sign, then R j has the same sign as the total strain distribution Furthermore, the sum of the absolute values of the influence ratios must be one Thus, the influence of each loading on the total strain distribution can be checked by the influence ratio proposed heretable 3 presents the strain-shape functions, participation factors, and influence ratios for this simulation for three different FBG sensor numbers (k =1,2,and3) For this eample case, the most active loading is the uniformly distributed loading of F 1

6 6 Distributed Sensor Networks Table 2: Distributional strain-shape function φ():cantilever beam oading condition ω P M FBG FBG FBG ConditionofFBGsensor location 0< FBG 0< FBG φ() 2 FBG 2 0< FBG 1 FBG Table 3: The function of the total strain distribution and its components oad λ k j k=1 k=2 k=3 φ k j () λk j φ k j () λk j φ k j () λ k j φk j () R j F ( 2 15) / ( 2 15) / ( 2 15) / ( 2 15) 0522 F ( 15)/ ( 15)/ ( 15)/ ( 15) 0239 F / / / () 0239 ε t = 3 λ k j φk j () = 1352 (2 15) 3089 ( 15) () Actuator FBG 1 FBG 2 FBG 3 FBG 4 FBG 5 FBG 6 FBG 7 ESG 1 ESG 2 ESG 3 ESG 4 ESG 5 ESG 6 ESG 7 ESG 8 ESG 9 FBGs ESGs (a) Side and bottom views for the test model subjected to two point loadings 500 Hydraulic jack oading points Pin support (b) Overall view of the test model Figure 6: Test model setup

7 Distributed Sensor Networks 7 4 Application to Measurements of the Maimum Strain of a Steel Beam The proposed measurement model could evaluate the maimum strain in beam structures subjected to multiple loading conditions using the strain-shape functions for various loadings, including point loading, distributed loading, and bending moment loading, as shown in Tables 1 and 2 The proposed measurement model was demonstrated for distributed loading and bending moment loading by a numerical simulation in Section 33 In this chapter, the eperimental test is conducted to evaluate the performance of the proposed measurement model for the point loading condition 41 Test Setup In the bending test of the simply supported steel beam in Figure 6, a concentrated load was applied on the upper steel beam by a hydraulic jack The load was increased in two steps, 74 and 129 kn The concentrated load was split into two concentrated loads applied 125 and 325 m from the left end of the beam The section of the beams was H- 6 8with a length of 4m Figure 6(a) shows that the measurements during the bending test were performed with both seven s with a 10 mm grating length and nine ESGs with a 5 mm gauge length attached to the surface of the bottom flange IS 7000, which is manufactured by Fiberpro, was used for the interrogation system The major specifications of IS 7000 are listed in Table 4The grating lengths of the seven s are 1,540 1,560 nm, in accordance with the wavelength tuning range of IS 7000 The sampling rate for the strain measurement was set as 01 per second 42 Results During the test, the beam deflected downward and tensile longitudinal strains occurred at the outer surface of the bottom flange Before testing the measurement model, the strains obtained from the FBGs and ESGs were compared to verify the quality of the value by the FBGs Table 5 indicates that the strain measured by FBGs could be assessed as reliable values for sensing beam members The maimum difference between the two measurements was found to be less than 134% for each load step To measure the maimum strain of the test model subjected to two different point loads, two s were required to estimate the maimum strain of the beam based on the model proposed in(6) Among the 21 possible combinations when choosing two of the seven s without repetition, the combination of the first and second s attached 05 and 10 m from the left end of the beam is not valid for the estimation because the matri [φ k ] in (8) based on the combination is singular To test the dependency on the selection of the locations for s in the measurement of the maimum strain, all 20 combinations in Table 6 were used in the measurements The numbers for the s are shown in Figure 6(a) Figure 7 compares the estimated distributions of strains from the 20 combinations with the strains measured directly from the nine ESGs for each loading step Furthermore, the dependency of the locations for Table 4: Major specifications of the interrogator (IS 7000) aser module (wavelength swept laser) Wavelength tuning range 1,5337 1,5695 nm Sensor module Repeatability ±2pm Wavelength accuracy <±10 pm Sampling frequency 200 Hz Strain (με) estimated strain distributions (loading step 2) 20 estimated strain distributions (loading step 1) ocation (mm) Strains measured by ESGs (loading step 1: 74 kn) Strains measured by ESGs (loading step 2: 129 kn) Figure 7: Estimated strain distributions from different combinations of s for two the s installed at the beam structure on the selectioncanbeavoidedusingthemeasurementmodelin(5) To evaluate the performance of the measurement model, the relative error in the estimated strain distributions was calculated by relative error of strain distribution (%) = ε est ε ESG ε ESG, (17) where ε ESG is a vector of the strains measured directly from nine ESGs and ε est is a vector of the estimated strains from the model Figure 8 indicates that the average relative errors for the 20 possible combinations of strain distributions are 237% and 184% for loading steps 1 and 2, respectively The maimum of the relative errors for all combinations was approimately 6% for the first combination in load step 2 The estimated maimum strains based on the 20 combinations are compared with the maimum directly measured from ESG #3 in Figure 9 The average relative errors for the 20 possible combinations in the maimum strain are 254% and 274% for loading steps 1 and 2, respectively The eperimental results indicate good agreement between the estimated strains based on the model and the strains measured from the ESGs

8 8 Distributed Sensor Networks Table 5: Measured strain by the s and ESGs Sensing location (mm) First step (74 kn) Second step (129 kn) FBG (με) ESG (με) Relative error (%) FBG (με) ESG (με) Relative error (%) , , , , , , , , Table 6: Number alignment of the combinations Combination number Combination of sensors (1,3) (1,4) (1,5) (1,6) (1,7) (2,3) (2,4) (2,5) (2,6) (2,7) Combination number Combination of sensors (3,4) (3,5) (3,6) (3,7) (4,5) (4,6) (4,7) (5,6) (5,7) (6,7) Relative error (%) Combination number oading step 1 (74 kn) oading step 2 (129 kn) Figure 8: Relative errors of the estimated strain distributions 5 Conclusion In this paper, a maimum strain measurement model for beam structures subjected to multiple loading conditions is developed by defining the strain-shape function and participation factors The maimum strain in a beam is found by the sum of the strains caused by the different loadings acting separately The strain-shape functions for various loading and support conditions are derived, and a system of equations is defined to calculate the participation factors The influence of each loading on the value of the total strain can be identified using the influence ratio defined here The measurement model is applied to monitor the maimum strain in a 4 m long steel beam subjected to two Strain (με) Combination number Measured strain (ESG): loading step 2 Estimated value: loading step 2 Measured strain (ESG): loading step 1 Estimated value: loading step 1 Figure 9: Maimum strains estimated from the estimated strain distributions concentrated loads Seven s and nine ESGs were attached on the surface of the bottom flange to obtain measurements during the test The estimated maimum strain and distribution of strains along the length of the beam were compared with those obtained directly from the ESGs The eperimental results indicate good agreement between the strains that were estimated based on the model and the strains measured by the ESGs Furthermore, the dependency of the locations for the s installed on the beam structure on the selection can be avoided using the measurement model

9 Distributed Sensor Networks 9 Acknowledgments This research is supported by Grant (code #09 R&D A01) from High-Tech Urban Development Program funded by the Ministry of and, Infrastructure and Transport in Canada with fiber optic sensors, Smart Materials and Structures,vol10,no3,pp ,2001 [15] W Chung and D Kang, Full-scale test of a concrete bo girder using FBG sensing system, Engineering Structures, vol30,no 3,pp ,2008 References [1] AISC, Manual of Steel Construction, Allowable Stress Design, American Institute of Steel Construction, Chicago, Ill, USA, 1989 [2] H S Park, H M ee, H Adeli, and I ee, A new approach for health monitoring of structures: terrestrial laser scanning, Computer-Aided Civil and Infrastructure Engineering, vol22, no 1, pp 19 30, 2007 [3]YWang,YWang,Yi,JRan,andMCao, Eperimental investigation of a self-sensing hybrid GFRP-concrete bridge superstructure with embedded s, International Journal of Distributed Sensor Networks, vol2012,articleid ,10pages,2012 [4]THYiandHNi, Methodologydevelopmentsinsensor placement for health monitoring of civil infrastructures, Distributed Sensor Networks, vol2012, Article ID , 11 pages, 2012 [5] HSPark,HSJung,YHKwon,andJHSeo, Mathematical models for assessment of the safety of steel beams based on average strains from long gage optic sensors, Sensors and Actuators A,vol125,no2,pp ,2006 [6] T A Hampshire and H Adeli, Monitoring the behavior of steel structures using distributed optical fiber sensors, Journal of Constructional Steel Research,vol53,no3,pp ,2000 [7] HSPark,SMJung,HMee,YHKwon,andJHSeo, Analytical models for assessment of the safety of multi-span steel beams based on average strains from long gage optic sensors, Sensors and Actuators A,vol137,no1,pp6 12,2007 [8]HMeeandHSPark, Gage-freestressestimationofa beam-like structure based on terrestrial laser scanning, Computer-Aided Civil and Infrastructure Engineering, vol26,no8, pp , 2011 [9] KTVGrattanandTSun, Fiberopticsensortechnology:an overview, Sensors and Actuators A, vol 82, no 1, pp 40 61, 2000 [10] S W James, M Dockney, and R P Tatam, Simultaneous independent temperature and strain measurement using infibre Bragg grating sensors, Electronics etters, vol 32, no 12, pp , 1996 [11] H i, D i, and G Song, Recent applications of fiber optic sensors to health monitoring in civil engineering, Engineering Structures, vol 26, no 11, pp , 2004 [12] D uo, Z Ibrahim, Z Ismail, and B Xu, Optimization of the geometries of biconical tapered fiber sensors for monitoring the early-age curing temperature of concrete specimens, Computer-Aided Civil and Infrastructure Engineering, vol28,no7, pp , 2013 [13] M Mieloszyk, Skarbek, M Krawczuk, W Ostachowicz, and A Zak, Application of fibre Bragg grating sensors for structural health monitoring of an adaptive wing, Smart Materials and Structures, vol 20, Article ID , 12 pages, 2011 [14] RCTennyson,AAMufti,SRizkalla,GTadros,andBBenmokrane, Structural health monitoring of innovative bridges

10 Rotating Machinery Engineering Journal of The Scientific World Journal Distributed Sensor Networks Journal of Sensors Journal of Control Science and Engineering Advances in Civil Engineering Submit your manuscripts at Journal of Journal of Electrical and Computer Engineering Robotics VSI Design Advances in OptoElectronics Navigation and Observation Chemical Engineering Active and Passive Electronic Components Antennas and Propagation Aerospace Engineering Modelling & Simulation in Engineering Shock and Vibration Advances in Acoustics and Vibration