MEASURING THE BOWING OF MARBLE PANELS IN BUILDING FACADES USING TERRESTRIAL LASER SCANNING TECHNOLOGY

Transcription

1 - Journal of Information Technolog in Construction - ISSN MEASURING THE BOWING OF MARBLE PANELS IN BUILDING FACADES USING TERRESTRIAL LASER SCANNING TECHNOLOGY PUBLISHED: Januar 00 at EDITOR: B-C Björk Fahim Al-Neshaw, Lic.Sc. (Tech.), Jukka Piironen, M.Sc. (Tech.), Susanna Peltola, M.Sc. (Tech.), Department of Structural Engineering and Building Technolog, Helsinki Universit of Technolog, Finland; firstname.lastname@tkk.fi Anna Erving, M.Sc. (Tech.), Nina Heiska, MA (Archaeolog), Milka Nuikka, M.Sc. (Tech.), Institute of Photogrammetr and Remote Sensing, Helsinki Universit of Technolog, Finland; firstname.lastname@tkk.fi Jari Puttonen, Professor of Structural Engineering and Building Technolog, Helsinki Universit of Technolog, Finland; firstname.lastname@tkk.fi SUMMARY: Natural stones are widel used as uilding facade materials. Numerous cases of damage indicate that their deterioration depends mainl on amient climate. The marle claddings in some cases are known to ow, epand and lose their strength when eposed to weathering. Depending on the owing, panels ma show cracks mostl initiated at the dowels. The percentage of visile cracks and reakouts increases with the amplitude of owing. The ojective of this paper is to find a potential use for the terrestrial laser scanning technique in detecting the deterioration of the uilding facades and in measuring quantitativel the dimensions of the damaged areas. This paper focuses on detecting the owing of marle cladding of uilding facades. Field measurements were carried out using a terrestrial laser scanner. The measurements of the owing of the marle panels were also carried out manuall with a so-called ow meter. The results show that the terrestrial laser scanning technique gives an accurate and reasonale method for measuring the owing of marle panels of the uilding facades. The terrestrial laser scanning is not a replacement for eisting condition surve techniques, ut an additional method. The terrestrial laser scanning can e emploed to complete man surveing tasks on large surfaces ecause of the spatial coverage of the point cloud, and the non-touching measurement principle. KEYWORDS: Laser scanning, Deterioration, Inspection, Marle panels, Building Facades. REFERENCE: Al-Neshaw F, Piironen J, Peltola S, Erving A, Heiska N, Nuikka M, Puttonen J (00) Measuring the owing of marle panels in uilding facades using terrestrial laser scanning technolog, Journal of Information Technolog in Construction (ITcon), Vol. 5, pg , COPYRIGHT: 00 The authors. This is an open access article distriuted under the terms of the Creative Commons Attriution 3.0 unported ( which permits unrestricted use, distriution, and reproduction in an medium, provided the original work is properl cited.. INTRODUCTION Durailit is an important issue to consider when choosing stones as cladding material for eterior eposure. The use of stone panels as cladding materials for facades has undergone a considerale increase in the last decades. Durailit prolems of marle claddings focus on the most significant deterioration feature known as the owing ehaviour. An eample of the owing of marle panels on a uilding facade is shown in FIG.. The reasons for ITcon Vol. 5 (00), Al-Neshaw et al., pg. 64

2 the deformation of the marle panels could e the thermal micro-fracturing of marle, the freezing and thawing ccles, the variation of moisture contents and the comined effect of temperature ccles under the presence of moisture (Siegesmund et al. 008). FIG. : Eample of the clearl developed conve owing of marle panels (Siegesmund et al. 008). Structural condition monitoring is an important methodolog in evaluating the condition of a structure assessing the level of deterioration and the remaining service life of the structures. Toda terrestrial laser scanning has een applied to structure condition monitoring. This technique allows the full remote acquisition of a surface with the sampling step of one centimetre step and is, therefore, ale to detect deformations. The ke of the proposed method is ased on curvature analsis of Terrestrial laser scanning ased data and is aimed to recognize the surface defects of the structure. (Teza et al.009) The Terrestrial laser scanning technolog has een developed a lot in the recent ears. The scan speed eceeds tens of thousands points per second. High performance scanners are used widel to record defect information and structural damage. The detailed information of the defects recorded the 3-D laser scanners can e used in digital format. The digitized image can e manipulated further using colouring schemes to magnif the defects. (Chang et al. 008) As a result, the terrestrial laser scanning is ecoming more feasile as a data collection method for applications in different industrial and construction fields. The data received from laser scanning have een widel used in surve applications, gloal positioning, maintenance of historical sites and structural monitoring. The last two decades have seen the development of various scanning technologies and various defect inspection methods and algorithms have een developed using terrestrial laser scanning. (Lia 006) & (Van Gosliga et al. 006) FIG. : The principle of terrestrial laser scanning of uilding facades. ITcon Vol. 5 (00), Al-Neshaw et al., pg. 65

3 Terrestrial laser scanning is a special technique in which the uilding is scanned as shown in Fig.. The result is a high-resolution cloud of points. The terrestrial laser scanning sstems use a directed laser eam for distance measurement to collect spatial information. A laser scanner creates a model of a large numer of points with z coordinates. The point cloud is a regularl sampled spatial representation of the real world. The z coordinates relate the points measured on real world ojects to the origin of the scanner, or more often to a project coordinate sstem used to tie several scans together. The ailit of a laser scanner to capture large amounts of data quickl and with a fine resolution means that the real world can e accuratel modelled. [Bornaz et al 004]. FIELD MEASUREMENTS AND DATA ANALYSIS Field measurements were carried out using a terrestrial laser scanner. A flow chart for the eecution sequence of the terrestrial laser scanner technique is shown in Fig. 3. FIG. 3: Flow chart of the laser scanning technique. The owing of marle panels was calculated fitting a polnomial curve to the laser scanning point cloud data in the vertical and horizontal directions. Geomagic Qualif software was used for further analsing and determining the surface delamination and the joints decaing of the masonr facade. Measurements of the owing of the marle panels were also carried out manuall with a so-called ow meter.. Field measurements For the field measurements, the marle panels of a wall at Helsinki Universit of Technolog (TKK) were selected, shown in Fig. 4. FIG. 4: The marle wall of the department of Architecture at TKK ITcon Vol. 5 (00), Al-Neshaw et al., pg. 66

4 The field measurements were carried out with FARO LS 880HE80 terrestrial laser scanner and manuall using ow-meter. The distance measurement of the terrestrial laser scanner is ased on phase difference technique. The marle wall was scanned from one position without overlapping measurements. The scanner was located 4.36 m from the centre of the marle wall. The sstematic distance error is +/- 3 mm. The achieved piel resolution for the terrestrial laser scanner was approimatel a mm point spacing (grid), when the scanner was approimatel four meters awa from the oject. In terms of rangefinder repeatailit, the manufacturer of the FARO LS 880HE80 terrestrial laser scanner specifies ±.6 mm and ±5. mm for 90% and 0% reflectivit respectivel, at a 0 m range. At a 5 m range, the repeatailit for 90% is ±4. mm and for0 % it is ±0 mm The scan angle accurac for the FARO LS 880HE80 scanner is ± Filtering noise and deleting additional and unnecessar points from the point cloud were performed with FARO Scene software; Geomagic Qualif and Realworks Surve software were used for further analses of the data. For the purpose of a more in-depth analsis with other mathematical software, the coordinate sstem of the point cloud was transformed. The origin of the sstem was transferred to the lower left corner of the fitted plane of the facade. Likewise, the point cloud was thinned out in order to manage the measurement data etter, resulting in 4, 5 and 7 mm grids to the data. The measurements of the marle panels owing were also carried out with the so-called ow meter. shown in Fig. 5.a. The ow meter is a 400 mm straight ar with a digital calliper which allows the distance from the edge of the ar to the panel surface to e measured accuratel. The measurement accurac of the ow-meter is ± 0. mm. The owing of the marle panel was measured in the vertical and the horizontal direction as shown in Fig. 5.. (a) () FIG. 5: Manual measurement of the owing of marle panels: a) Sketch of the ow-meter and ) the location of the measuring points on the marle panel.. Calculation of the owing of marle panels Deterioration of marle panels involves several parameters and properties. Shape deformation is the most ovious phenomenon. where the panels ow either convel or concavel out of their original plane. In addition to the owing there also appears some permanent volume changes i.e.. the marle epands. [Grelk et al. 007] The owing magnitude of marle panels was calculated using equation. d B = *000 () L where, B is the owing magnitude epressed in (mm/m); d is the measured value of owing in (mm); and L is the measuring distance etween the supports of the marle panel in (mm). According to the results of the manual measurements and the visual inspection, the owing of the marle panels was calculated fitting a second order curve to the laser scanning point cloud data from the centre line of the panel oth in the vertical and the horizontal direction. The quadratic polnomial curve for the vertical and the horizontal owing are shown in Fig. 6. ITcon Vol. 5 (00), Al-Neshaw et al., pg. 67

5 FIG. 6: The quadratic polnomial curves for the vertical and the horizontal owing of the marle panel. The polnomial equation of the vertical owing is: f ( z) a z + a = () * z + a * 0 where: z is a variale of the panel height and the unknown coefficients a, a, and a 0 are calculated using equation 3. a 0 a... a z.. z n z... =... z n n (3) The polnomial equation of the horizontal owing is: f ( ) + = (4) * + * 0 where: is a variale of the panel width and the unknown coefficients,, and 0 are calculated using equation =... n n n (5) The horizontal and vertical owing of marle panels was calculating using mathematical equations which descrie the curve fitting of the terrestrial laser scanning point cloud data. The following eample shows how to calculate the horizontal owing magnitude of the centre line of the marle panel. The owing magnitude of the vertical centre line of the marle panel is calculated using a similar calculation. An eample of the polnomial curve fitting of terrestrial laser scanning point cloud data of the horizontal centre line of the marle panel is shown in Fig. 7. ITcon Vol. 5 (00), Al-Neshaw et al., pg. 68

6 ITcon Vol. 5 (00), Al-Neshaw et al., pg. 69 FIG. 7: Eample of the polnomial curve fitting of the terrestrial laser scanning point cloud data. The value of the measuring distance etween the supports of the marle panel was calculated with equation 6. ) ( ) ( L + =, (mm) (6) where : (, ) = the first point in the fitted quadratic polnomial curve and 0 * * + + = (, ) = the last point in the fitted quadratic polnomial curve and 0 * * + + =, and 0 are calculated with equation 5 The maimum owing value of the marle panel was measured from the verte of the quadratic polnomial curve. The coordinates of the verte of the curve are: verte = (7) * * f verte + = + + = = (8) The measured value of owing of the marle panel in (mm) is calculated with equation 9. * ) * ( + = ) ( ) ( ) ( d verte verte, (mm) (9) The owing magnitude of marle panels calculated with equation.

7 3. RESULTS AND DISCUSSION The field measurements of the marle panels mostl show a remarkale conve owing, as shown in Fig. 8. One panel on the fourth row shows a clear concave owing. One reason for the panel owing could e water penetrating ehind the marle panels, which increases and accelerates the deterioration of the panels. The failure of the lateral fiing of the panels could also e a reason for the marle panel owing. FIG. 8: Terrestrial laser scanning data, colouring the magnitude of the deformation from the planarit. The results of the manual measurement of two marle panels using the ow-meter are shown in Figures 9 0. The values of the owing of the panel Mar-R4-C are 5. and 6.9 mm/m for the horizontal and the vertical directions respectivel. The coefficient of determination (R-square) for the curve fitting of the manual measurements of the panel Mar-R4-C is aove FIG. 9: The manual measurements of the owing of the marle panel Mar-R4-C. The values of the owing of the panel Mar-R4-C4 are -. and -4.3 mm/m for the horizontal and the vertical directions respectivel, which indicates a concave owing of the panel in oth directions. The coefficients of ITcon Vol. 5 (00), Al-Neshaw et al., pg. 70

8 determination (R-square) for the curve fitting of the manual measurements of the panel Mar-R4-C4 are 0.99 and 0.89 for the horizontal and the vertical directions respectivel. FIG. 0: The manual measurements of the owing of the marle panel Mar-R4-C4. The results of the manual measurements and the visual inspection of the owing of the marle panels show that the quadratic polnomial curve fitting was the est alternative for measuring the owing magnitude using terrestrial laser scanning point cloud data. Some eamples of the owing in the marle panels are shown in tale and Figures 4. The polnomial curve fitting of the laser scanning point cloud data for the centre line of the panel Mar-R-C3 shows a slight owing of the panel. The vertical direction owing value is. mm and the horizontal direction value is 3.6 mm. The standard deviations of the cloud point fitting to the curve are.8 mm on the vertical direction and 3.8 mm on the horizontal direction. The second order curve fitting to the laser scanning point cloud data in the vertical and the horizontal directions shows low (0.5 and 0.3) coefficients of determination due to the scattering of the point cloud data. A slight owing is measured in the horizontal direction of the panel Mar-R-C. The maimum conve owing magnitude was measured on the horizontal direction of the panel Mar-R-C. The magnitude (of the owing) is 8.5 mm (L/00) and the standard deviation of the distance of the measured laser scanning points from the fitting curve is. mm. The R-square coefficients of the curve fitting to the laser scanning point cloud data in the vertical and the horizontal direction are 0.6 and 0.4. The concave owing of the panel Mar-R4-C4 is distinct in the visual inspection, manual measurement and the curve fitting of the laser scanning point cloud data. The horizontal and vertical owing values of the panel Mar- R-C are. mm (L/85) and 4.3 mm (L/65) respectivel. The standard deviation of the distance of the measured laser scanning points from the fitting curve is.4 mm. The R-square coefficients of the curve fitting in the vertical and the horizontal directions are 0.76 and 0.8. TABLE : The owing magnitude of the marle panels using terrestrial laser scanning sstem. Marle panel Standard Tpe of Bowing Bowing L d deviation owing magnitude direction (mm) (mm) (mm) (mm/m) Mar R C3 Horizontal Conve 4 Vertical Conve Mar R C Horizontal Conve 4 Vertical Conve 6 Mar R4 C Horizontal Conve 6 Vertical Conve 9 Mar R4 C4 Horizontal Concave - Vertical Concave -5 ) Standard deviation of the distance of the measured laser scanning points from the fitting curve ITcon Vol. 5 (00), Al-Neshaw et al., pg. 7

9 FIG. : The horizontal and vertical owing in marle panel Mar-R-C3. FIG. : The horizontal and vertical owing in marle panel Mar-R-C. FIG. 3: The horizontal and vertical owing in marle panel Mar-R4-C. ITcon Vol. 5 (00), Al-Neshaw et al., pg. 7

10 FIG. 4: The horizontal and vertical owing in marle panel Mar-R4-C. Tale presents the results of the onsite owing measurements using the terrestrial laser scanner and the owmeter. The comparison of these results showed a difference of mm/m for the conve owing magnitude and 6 7 mm/m for the concave owing magnitude. TABLE : Comparing the owing magnitudes of two marle panels using laser scanning and ow-meter. Marle panel Mar R4 C Mar R4 C4 Tpe of owing B/TLS (mm/m) B/ow-meter (mm/m) B TLS B ow-meter (mm/m) Horizontal conve Vertical conve Horizontal concave Vertical concave The magnitude of conve owing (B = mm/m) was measured on the panel Mar R4 C. The magnitude of concave owing (B = - -5 mm/m) was measured on the panel Mar R4 C4. The high magnitude of the concave owing could e due to the failure of the panel anchoring to the wall. 4. CONCLUSIONS In this paper, the owing of marle cladding panels of a wall at Helsinki Universit of Technolog using laser scanning technolog was investigated. The field work was carried out with manual ow-meter and FARO LS 880HE80 terrestrial laser scanner. The terrestrial laser scanning technique is an alternative method for measuring the owing of cladding panels or elements of the uilding facades. The owing of the marle panels could e calculated fitting a curve to the laser scanning point cloud data oth in the vertical and the horizontal directions. The tpe of the curve fitting could e determined according to the results of the manual measurements, the visual inspection or the shape of the point cloud data itself. The comparison etween the results from the laser scanning and the ow-meter showed onl a slight difference in measuring the conve owing magnitude and quite a high difference in measuring the concave owing magnitude. Laser scanning is not a replacement for eisting condition surve techniques, ut an additional method, which provides location ased information on the uilding defects and deterioration. 5. REFERENCES Bornaz. L. and Rinaudo. F. (004). Terrestrial Laser Scanner Data Processing. International archives of photogrammetr remote sensing and spatial information sciences. Vol. 35. Part 5. pp ITcon Vol. 5 (00), Al-Neshaw et al., pg. 73

11 Chang K-T.. Wang E.. Chang Y-M. and Cheng H-K. (008). Post-Disaster Structural Evaluation Using a Terrestrial Laser Scanner. FIG Working Week June 008. Stockholm. Sweden. Availale online at: [Accessed ]. Grelk. B.. Christiansen. C.. Schouenorg. B. and Malaga K. (007). Durailit of Marle Cladding A Comprehensive Literature Review. Journal of ASTM International. Vol. 4. No. 4. Availale online at: [Accessed ]. Hall D. (008). Building defects. Inspections and Reports. Availale online at: [Accessed ]. Lia T. (006). Automated Detection of Surface Defects on Barked Hardwood Logs and Stems Using 3-D Laser Scanned Data. Doctoral Dissertation. Facult of the Virginia Poltechnic Institute and State Universit. Availale online at: [Accessed ]. Newman A. (00). Structural Renovation of Buildings: Methods. Details. and Design Eamples. McGraw-Hill. New York. Pp Siegesmund. S.. Ruedrich. J. and Koch. A. (008). Marle owing: comparative studies of three different pulic uilding facades. Environmental Geolog Journal. Springer Berlin / Heidelerg. Vol. 56. Numers 3-4 / Decemer Pp Teza. G.. Galgaro A.. Moro F. (009). Contactless recognition of concrete surface damage from laser scanning and curvature computation. NDT&E International Volume 4. Issue 4. June 009. Pp Van Gosliga. R.. Lindenergh. R.. Pfeifer. N. (006) Deformation analsis of a ored tunnel means of terrestrial laser scanning. ISPRS Technical Commission Smposium. Septemer 006. Dresden. German. International Archives of Photogrammetr. Remote Sensing and Spatial Information Sstems. Volume XXXVI. Part 5. Pp ITcon Vol. 5 (00), Al-Neshaw et al., pg. 74