Hydrostatic Levelling System: Monitoring of Historical Structures

The hydrostatic levelling system (HLS) is a highly accurate technique to monitor differential vertical settlements. The monitoring system is based on the principle of communicating vessels. The working principle is outlined and the results of two case studies are presented. These case studies not only focus on the use of the HLS as an accurate way of measuring differential settlements, but emphasize the relation with the structural behaviour of the building itself. Comparisons with more conventional geodetic systems are made. This provides objective information on which further options on strengthening and consolidation of the structure are based. d1 d2 d’1 d’2 d’12 Vessel 1 Vessel 2 Thermometer Data-acquisition Air water After vertical displacement Before vertical displacement Average water level Vessel 1 Vessel 2 530 Structural Analysis of Historical Constructions The readings range from 5000μm to 10000μm. Measurements are taken with the system frequency (about 33 Hz) and the results are directly visible on the computer screen. The data are stored with an adjustable frequency. The values stored are the means of 100 values taken in 3s. The temperature is measured in every vessel and the rough displacement measurement is directly corrected accordingly. The tubes of the water circuit are placed as horizontally as possible in order to remove the effects of a temperature gradient. The tubes of the air circuit lead upwards from the vessels to prevent the condensation water from staying in the air circuit. The analogical response of the sensor is digitalised and linearised by a 12Bit card. Since no vessel can be considered steady, it is impossible to know the absolute values of the settlements. In practice, all the displacements are recalculated in relation to the mean level of the water in the circuit. 2 CASE STUDIES In both case studies the measuring period extended sufficiently long – 5 months to filter out the influences of temperature and relative humidity enabling the real influencing parameters to be identified, such as the water level of the river Demer in the first case study. Additionally, the system of hydrostatic levelling has been placed in parallel with a more conventional geodetic system. The latter has been measured at discrete moments in time. These allow to compare both measurement techniques and to draw some conclusions on the accuracy of the geodetic survey and its ability to complement the findings from a long term measurement campaign. In both case studies, the causes of the measured differential settlements could be retraced, delivering crucial information related to the structural behaviour of the historical building and its loadbearing capacity. 2.1 Duke’s Mill Aarschot (Belgium) The first case study deals with the largest water mill in Western Europe, named as Hertogenmolens (E: Duke’s Mill) in Aarschot (B). Notes of the existence of a water mill at the site were retraced and date back from the beginning of the 16th Century. During the centuries, the mill underwent many changes. Moreover, in 1970 it suffered from a fire and in 1986 a part of the western facade collapsed. In 1997 it was decided to set up a global restoration plan. In 1998 and 1999 the building pathology and stability were studied and an extensive monitoring program was executed to evaluate possible differential settlements of the foundations under scientific supervision of the KULeuven and in collaboration with the Ministry of the Flemish Government and Studiegroep Omgeving, an engineering office with a department specialized in topography. The central buildings that cross the river Demer, are supported by wooden piles, as constructed in 1582. Based on this measurement campaign, the global movements of the load-bearing structure could be understood. The dependency of temperature and water level of the river Demer as well as the mutual dependency between the different load-bearing elements was quantified (Van Balen et al. 1999). During a period of 6 months the differential settlements of the central buildings were monitored. Table 1 gives an overview of the measured quantities, measurement period, frequency and institute that took care of the measurements. Table 1 : Overview of the executed measuring campaign ________________________________________________________________________________________________________ Measurement quantity measurement period measurement frequency Institute ________________________________________________________________________________________________________ Temperature 23/07/1998-04/02/1999 0.5Hz a,b Water level of the river Demer permanent twice a day c HLS-measurement – 6 vessels 23/07/1998-04/02/1999 0.5Hz a,b Geodetic Survey 29/04/1998-25/01/1999 twice a month d ________________________________________________________________________________________________________ Legend: a: Studiegroep Omgeving cvba; b: KULeuven, Department of Civil Engineering; c: DIHO Ministry of the Flemish Community, department of Hydrological Research; d: MVG-ATO Ministry of the Flemish Community, department of overall technical support ________________________________________________________________________________________________________ Luc Schueremans, K. Van Balen, P. Smars, V. Peeters and D. Van Gemert 531 The six measuring devices (HLS-1 to HLS-6) were placed on the main floor, and attached on the transversal load-bearing walls of the central building, Fig. 2. The temperature is measured at each of the six HLS measurement devices, at the same frequency as the data acquisition of the water level in the devices. Fig. 3 presents the water level of the 6 HLS-measurement devices and the water level of the river Demer. The values shown in Fig. 3 represent daily mean values. Figure 2 : Hertogen’s models (E: Duke’s Mill), ground plan, HLS-network and points for geodetic survey. Figure 3 : Water level of the 6 HLS-measuring devices and the river Demer For the HLS-measurements, a distinction was made between the measurements for the period 23/07/1998-18/08/1998 and the period 19/08/1998-04/02/1999. During the first period the end of the tube which interconnects the different HLS devices, was not closed airtight. The water in the tube evaporated slowly. Although this did not influence the mean value, the spread was higher, demonstrating a lower accuracy, Fig. 4. These data should not be considered lost. On the -1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4 0,6 0,8 1 1,2 14/07/1998 2/09/1998 22/10/1998 11/12/1998 30/01/1999 Date Le ve l H LS [m m ]