by Giovanni Moschioni Researcher Politecnico di Milano, Facoltà di Ingegneria di Lecco, Italy

Transcription

1 Category: Maintenance/Field Test Prodotti utilizzati Labview FDS, Picture Toolkit PCI 6031E FP1001, RTD, AI PCI 485/2 Labview Monitors The Tallest Medioeval Bell Tower In Europe by Giovanni Moschioni Researcher Politecnico di Milano, Facoltà di Ingegneria di Lecco, Italy The Challenge: Setting-up a complex system for long term investigation about static and dynamic behavior of a tall masonry medioeval bell-tower The Solution: A hybrid system with high frequency measurement by means of a DAQ board for vibrations, and low frequency measurement with Fieldpoint network for cracks opening. The system, controlled with LabVIEW, is remotemonitored via modem. Paper A long term investigation is being carried out by the Politecnico di Milano on the Torrazzo, the Cathedral bell tower and symbol of Cremona, Italy, the city of Stradivarius. The date of construction can be traced up to the end of 13 th century. The Torrazzo with its 112 m (367 ft) is the tallest masonry bell tower in Europe (Fig. 1). Several signs of damage like some passing through cracks and surface deterioration appearing on the tower walls suggested to begin an on-site investigation in order to understand the tower dynamic and static behavior mainly under the action of wind and thermal gradients. A PC based system has been setup for continuous measurement, on-line analysis, and monitoring of slow, such as walls and air temperature and crack openings, and fast signals (vibrations, wind velocity and direction). Measurements And Transducers Static measurements are mainly related to the tower cracks opening, which are supposed to depend predominantly from thermal actions during the day (sun radiation) and during the year (seasons changing). Air and walls temperatures are measured by means of about 40 RTD s (Pt100), cracks opening with 16 LVDT s. Transducers and sensor are placed in various sections of the tower. The main purpose of slow measurements is to understand whether the cracks openings follow a natural closed cycle during the year or they get larger and larger as time goes by. Dynamic measurements (essentially tower vibrations and their correlation with wind and bells actions) are somehow a bit more complex. They re asked to give both direct and diagnostic Figure 1. The Torrazzo in the sunset seen from the Duomo Square, Cremona, Italy

2 information. Direct information can be helpful to understand if there s a risk of structure damages under the action of the wind, bells movements or any other dynamic stimulus. In the meantime, data can be used to calibrate finite elements models of the structure. Once the direct information is acquired it can be used for diagnostic purposes: if one notices changes, for example, in natural frequencies from a standard situation, this could be considered as an alarm signal that something has changed in tower structure. In a nutshell, after the first part of investigation, one should be able to know the tune of the tower, and, during the following years, listen to changes in the tune to understand if something s going wrong. Dynamic signals come from nine servoaccelerometers in different sections and axis, placed so that it s possible to measure flexural and torsional modes of the tower. Instantaneous speed and direction of the wind on the structure are measured by means of four cup anemometers and four wind vanes installed close to the top (85 m), on the four sides of the tower. Such a high number of anemometers has been chosen in order to minimize structure blockage effects on wind measurements. Cracks openings dynamic signals from LVDT s are acquired too. The lowest transducer is placed near the base of the tower (about 15 meters), the highest at about 100 meters. Acquisition And Analysis System One of the main challenges of the investigation was the project of the measurement and data acquisition system, due also to the following requirements: signal and supply cables are to be as short as possible, both for obvious economic reasons and for measurement reliability; a good mass reference has to be given to all signals to avoid ground loops; the system must be flexible and scalable, in case of failure of parts of it or in case new sensors are added during the campaign; data are supposed to be enough for a correct and detailed knowledge of tower behavior, but not redundant and memory (and operators patience) consuming; system can be controlled from anywhere by means of modem. The choice was the split of slow and dynamic parts (Fig. 2). For example signals from accelerometers are to be acquired at 100 S/s; for RTD s a ten minutes sampling time is enough. Moreover dynamic sensors are installed, close each other, in few sections, while temperature and crack openings sensors are spread all over the tower.

3 LVDT Fieldpoint Modules (75 m) Anemometers LVDT Servoacc. RTD RS485 ADC Board RS485 Fieldpoint Modules (15 m) LVDT Figure 2. Synoptic view of the system, with sensors, communication and signal paths, and computers. Dynamic Acquisition And Analysis When beginning a blind and unattended long measurement campaign (in cases like this you have just a little idea of what phenomena and signals amplitudes will be, or, better, you do know that sometime they ll be very high and sometime very low) it s very important to have, at the same time, a wide range, useful when signals are high, and a good resolution, useful when signals are low. That s the reason why a 16 bits NI PCI-6031 E was chosen for dynamic data acquisition. As already mentioned it s a fundamental goal to know in the clearest way the tower behavior, but it s also important to avoid overflowing data. So the software, written with LabVIEW, acquires data continuously on a ten minutes basis (continuos scan), and stores them to disk only if a situation like this has never happened. Otherwise just stores statistic data for all channels (mean, RMS, and so on). For examples the user can decide to store just ten files for wind coming from direction 0-10 and velocity 5-10 m/s. When these conditions occur, the program stores waveform data as long as ten files are on the disk. Afterwards only the statistic data are saved. This is repeated for all directions and for wind velocities (Fig. 3). RTD

4 Figure 3. Screen capture of data acquisition program. (Top-left display: stored signal of an accelreometer, bottom left:real-time signal, bottom right real-time wind measurement) In this way in few months data collected are enough to cover all 360 circle with wind up to 27 m/s, and allow a nearly complete dynamic analysis. At the same time there s a complete mean history of events. With the choice of continuous scan, which is quite time-critical with such a high number of channels, it was possibile, under certain wind conditions, to acquire up to 20 hours of consecutive waveforms, with an incredibly high frequency resolution for spectral analysis (10-3 Hz resolution with 0-50 Hz band). Data stored can be analyzed directly on-site or at home, with general purposes and dedicated programs written in LabVIEW (Fig. 3,4).

5 Figure 4. Screen capture of one of off-line programs. In figure the animation of displacement of the top of the tower under 20 m/s wind. (Double integration of acceleration signals)

6 Static acquisition and analysis In this case sampling rate, and, therefore, the amount of data is not a great problem. On the other hand the high channel count and the distribution of sensors along the tower suggested the installation of a FieldPoint network connected to a PC in a room at 35 m. The network is made of two FP1001 network modules in two locations at 25 and 75 meters (Fig. 5), which communicate with PC by means of a PCI 485/2 board. The two network modules control analog input modules for LVDT s and RTD signals. This solution allows signals to have short cables paths to data acquisition and conditioning points and to carry only two serial cables (from the top and the bottom of the tower) instead of hundreds from each sensor to a DAQ board. A program written with LabVIEW controls FieldPoint network and collects data onto a dedicated PC, which is, on its own, in a Ethernet network with the dynamic PC. Conclusions In a measurement campaign, like the one on the Torrazzo, in a hostile and noisy environment, with logistic problems, with many transducers of different kinds, the acquisition system is required to be reliable and easy to use, perfectly tailored to the necessities, and at the same time, as cheap as possible. With LabVIEW and National Instruments hardware we got the goal (although we are a group of mechanical and civil engineers with just a little porgramming background), saving a lot of time thanks Figure 5 FieldPoint Modules in the bell-cell. to software (a lot of code prepared and debugged for other campaigns was reused) and hardware scalability. It s quite difficult to estimate money savings compared to commercial systems. The point is that without LabVIEW it wouldn t have been possible at all to have a system exactly as we wanted.