STRUCTURAL MONITORING OF A LARGE DAM BY TERRESTRIAL LASER SCANNING M. Alba (*) , L. Fregonese (*) , F. Prandi (*) , M. Scaioni (*) , P. Valgoi (**) (*) Politecnico di Milano – Dept. I.I.A.R., p.za L. da Vinci 32, 20133 Milano, Italy e-mail: {mario.alba, luigi.fregonese, federico.prandi, marco.scaioni}@polimi.it (**) AEM S.p.A. – Production Dept., Grosio (SO), Italy, e-mail: paolo.valgoi@aem.it KEYWORDS: Accuracy, Dam Monitoring, Deformation Analysis, Georeferencing, Terrestrial Laser Scanning ABSTRACT: This paper presents some first results of a project aimed to access the feasibility of monitoring deformations of large concrete dams by terrestrial laser scanning. For this purpose a test field has been established on the dam of the Cancano Lake (Valtellina, Italy). This is made up of a geodetic network materializing a local datum for georeferencing all data acquired at different times. A large number of retro-reflecting targets have been positioned and measured by a total station in the local datum. Moreover, targets have been measured also in captured scans. Some of these have been placed on the dam structure to be used as independent check points, the others in the surroundings to play as ground control point. Three measurement campaigns have been accomplished so far by using two laser scanners: a long range Riegl LMS-Z420i and a medium range Leica HDS 3000. Reported analyses are focused on two main problems: the first one is the accuracy and the stability of georeferencing, which is fundamental to make comparisons between different multi-temporal scans; the second one is the computation of deformation based on the acquired point-clouds. 1. INTRODUCTION This paper presents some results of a research that, in the opinion of the authors, could be considered a fascinating challenge: can be used Terrestrial Laser Scanning (TLS) techniques for monitoring displacements of big structures, and in particular of large dams? Monitoring the static behavior of large dams has always been a topic of great importance, due to the impact these structures have on the whole landmark where they are built. Many instruments and surveying methods have been applied so far in order to continuously access the safety of this kind of big structures. However, the common characteristics of all approaches is the possibility of measuring with high accuracy displacements of a small number of points, if compared to the size of a big dam. The number of control points is even smaller if an automatic measurement system is applied (i.e when using opto-electronic collimators or robotic total stations). On the other hand, laser scanners are capable to acquire a very huge number of points, so that the control could be extended to the whole structure instead of being limited to a few points. At a first look, the answer to the question introduced at the beginning of the paper would be negative. The small number of papers dealing with this subjects (see e.g. Schäfer et al., 2004) confirms this finding, even though the interest of researcher on TLS in applications for deformation analysis is quickly increasing (see e.g. Teskey and Bijoy, 2005; Guarnieri et al., 2006; Tsakiri et al., 2006). Although long range instruments capable to measure 3D points also at 400-500 m of distance (greater ranges are not considered because the data quality sharply decreases) have been sold by vendors for about a half decade, in concrete dam monitoring an accuracy of few mm is always required. Unfortunately, current instruments cannot satisfy these requirement. For better exploring this possibility, a group of Italian researchers in the context of a national research program (COFIN 2004) have started an activity with the aim of checking the real possibility of using laser scanners for dam monitoring. This research has been planned in cooperation with AEM S.p.A. Production Dept. 1 , which has opened one of its dams (the dam of Cancano Lake, near the well known village of Bormio, Valtellina) to setup a test field. Along this project a complete laser scanning surveying of this dam has been also accomplished, as described in Alba et al. (2006). Two laser scanners have been used in tests along three measurement campaigns, in May 2005, October 2005 and May 2006. A long range instrument Riegl LMS-Z420i has been stationed on two different stand points respectively at a mean distance of 200 and 300 m from the dam surface; a Leica HDS 3000 at shorter distances. To access the measurements of displacements derived from laser scanning, a set of 68 signalized targets have been distributed on the whole front of the structure, so that their 3D positions could have been measured by multiple intersections using a total stations. Data acquisition is performed in order to acquire a dense point cloud at the highest achievable accuracy. The concept we would like to apply for deformation analysis is to use an area-based method (Schneider, 2006), i.e. to make a comparison between data acquired at different times by considering mathematical surfaces fitting measured points, so that noise could be reduced (Lindenbergh and Pfeifer, 2005). The surface of the dam is divided is small homogeneous areas and for each of these an interpolating surfaces estimated. Because different point clouds have been already georeferenced into the same reference system, the comparison between interpolated surfaces allows to evaluate displacements. Unfortunately, the georeferencing is affected by errors, so that an analysis to check which part of detected displacement is really due to a structure’s movement is needed. At the current stage of the project only some simpler methods have been applied to make comparisons and then to detect deformations, as presented at par. 5. More enphasis has been given so far to the analysis of georeferencing stability, which is fundamental to make comparisons between surfaces registered into the same reference system (par. 4). On the other 1 AEM S.p.A. (Azienda Elettrica Municipale) is currently the main facility management company of the city of Milan; it is the owner of several hidropower plants and dams in Valtellina Valley (Northern Lombardia).