Field Measurements of Wind Loads on Flat Roofs Under Hurricane Conditions
AMAURY A. CABALLERO, KANG K. YEN, ERNESTO INOA
Department of construction Management and Department of Electrical and Computer Engineering
Florida International University
10555 W. Flagler Street, Miami, Florida 33174
UNITED STATES
Abstract: The research work is devoted to the design of a data collection system to gather empirical data on stress induced by hurricane conditions on the roof of flat roof buildings. The system includes a prototype instrument developed at Florida International, based on National Instruments Lab View. The objective of the system is to calibrate via empirical methods the data collected from model tests. This may eventually lead to recommendations for improved building design or construction methods.
Key words: Instrumentation, hurricane, flat roof measurements, data collection system
1 Background and Objectives
There have been a number of field investigations to measure wind pressure on low-rise buildings, as reported by Michot [1]. In the early seventies, a full-scale two-story house was constructed at Aylesbury, England to conduct measurements of wind induced pressures. Differential pressure transducers housed inside the building were attached via tubing from the transducer sensing port to predetermined sites on the roof and walls of the building [2]. Other studies have been carried out in Montana [3], where a house was instrumented, two houses in the Philippines [4], and a mobile home at the Gaithersburg Campus of the National Bureau of Standards [5].
For the past 5 years, the State of Florida has supported the Florida Coastal Monitoring Program to instrument houses along the coast of Florida.
Because of the design of the sensor package, that project has been limited to monitoring wind pressures on houses with sloped roofs. But there is also a need for field measurements of wind pressures on commercial and light industrial buildings in hurricanes. Existing field data for these types of buildings has been limited to relatively low wind speeds (20 to 40 mph) in extra tropical storms.
The field measurement program initiated in FY 03 as part of the Hurricane Loss Mitigation Program seeks to develop the technology that will allow monitoring of wind loads on buildings with flat roofs. This has involved the re-design of the sensor packaging so that the sensor can be stably installed on a flat roof without creating penetrations in the roof surface.
2 Instrumentation and Data Acquisition
2.1 Sensors
The instrumentation system, building off experiences of the Florida Coastal Monitoring Program, uses absolute pressure transducers to avoid having to run vinyl tubing over the roof in order to supply a reference pressure to each sensor. The selected transducer was the Micro switch 142 PC 15 A. This transducer operates with one excitation between 7.0 and 16.0 Volts DC, the necessary supply current is less than 20 mA, and the response time is 1 msec. The device can operate in the range of -40 to +1850 F. and the device is connected to a circuit for the purpose of matching to the communication line. Each sensor is connected to a central station through a single 4-conductor shielded cable. Wind speeds are monitored using an RM Young wind monitor with a hardened propeller and the speed range is set at 0 to 200 mph. The technical development of this system is described in [1].
The original device was prepared for its installation only on sloped roofs. When dealing with flat roofs, the sensor cannot be fixed to the roof without creating penetrations on its surface. Due to this a new packaging has been designed, as shown in Figure 1. This design is strong enough to resist possible damage by debris and the electrical circuitry is kept dry inside the housing. A sensing port is connected to the housing unit. This external sensing port allows us to measure the outside pressures captured by six ¼’’ copper tubes that permitted the wind to enter the sealed enclosure and a rubber hose connected through a soldered connection, where the six tubes meet to the electronic circuit sensor. The device is inserted in a ¼’’ x 24” x 24” steel base, where weights of 10 lbs are installed on the borders, to avoid any movement of the whole device
The sensors are connected to the data logging system in a star configuration. Data is collected using a National Instruments data acquisition board and a program developed using National Instrument’s LabView object oriented software.
Each of the sensors has been calibrated against a Setra Model 370 Digital Pressure Gage with an operating range of 800 to 1100 milli-bars. The results from calibration for the 16 sensors are listed in Table 1 along with the R2 value from the regression analysis.
The enclosure has been constructed using an aluminum pan and sealed with an aluminum back plate [6].
Figure 1. Pressure Sensor Packaging
The system has been installed on the Engineering and Applied Sciences (EAS) building, at Florida International University which is a three-story building 40 ft height, to continue monitoring of wind conditions and wind loads.
2.2 Data Logging System
The data acquisition system is based on the National Instruments LabView, version 7.0. It was selected due to the versatility offered by this software-hardware system. Having installed the system opens new possibilities of utilization not only in this project, but also in general for collecting information and/or control in any other project. The data collection system structure is shown in Figure 2 [7].
As can be seen in Figure 2, there are connected 16 pressure sensors and one anemometer, for a total of 18 measures. The sensors are located on the roof and connected through shielded cables to the FieldPoint. National Instruments FieldPoint is a modular real-time, distributed I/O system for measurement, control and data logging applications that demand industrial-grade hardware with easy installation and configuration [8]. This I/O system gives us the freedom to quickly and easily place measurement nodes near sensors. FieldPoint includes a variety of bank isolated analog and digital I/O modules, and network interfaces for easy connection to open, standard networking technologies. The FieldPoint system includes two general types of I/O modules- standard 8 and 16-channel modules and dual-channel modules for maximum mix-and match flexibility. In the present situation, three 8-channel modules have been employed. The connection between the FieldPoint and the virtual instrument can be realized using Ethernet, RS-232, RS-485, CAN, FOUNDATION Fieldbus, or wireless networks. The presented system uses Ethernet for connecting to the virtual instrument.
The FieldPoint is located also on the roof of the same building, in a covered space. The FieldPoint and the virtual instrument (VI) are connected to the source of energy through a UPS each in order to assure the supply of the necessary voltage to the system under any conditions.
The transmitted information from the FieldPoint through the Ethernet Link is received by a National Instruments PXI, where LabView has been installed. The PCI eXtensions for instrumentation (PXI) specifications define a rugged PC-based platform for measurement and automation systems. PXI uses the high-speed Peripheral Component Interconnect (PCI) bus, which is de facto standard driving today’s desktop computer software and hardware designs. PXI combines the PCI electrical bus with the rugged modular Eurocard mechanical packaging of compact PCI [8].
3 Results
Figure 3 shows the Virtual Instrument screen. The information from each sensor may be viewed on the two analog indicators, as well as is represented in two digital indicators. The wind speed and direction is also displayed in two digital indicators. The operator can select the sample rate, the date and time for starting and finishing recording information, or the wind speed for starting and finishing recording.
A run of measurements obtained with the VI is shown on Table 2. In the table only two values for each parameter are shown. The sample rate for this run was 1 msec, which explains why the two values are practically the same for all the sensors. At the time of a hurricane, the collected information permits among other results, to create pressure curves that indicate the points to be reinforced on the roof.
4 Conclusions and Recommendations
The data collection system is capable of sampling data from up to 24 sources at a sampling rate that can be fixed from a few milliseconds to several seconds. The data collection system has a precision higher than 1%. The sensors and anemometer give the system precision, which is not better than 2%. This means that the data logging system will not reduce the overall precision of the measurement.
Several runs have been made to show the system working properly. One example of a run at a sampling rate of 1 sample/second can be seen on Table 2. These runs have also given the idea of the necessary memory for a real experiment. For a sampling rate of 1 sample/second, the memory needed in one hour is less than 0.5 MB. This result confirms the suggestion of installing lower capacity computing units in future projects.
The electronic cards in the sensors have presented some reliability problems due to the extremely hard conditions of temperature and humidity they are forced to work with. It is recommended to realize a future work studying the following:
- Improving the sensor packaging with respect to the heat absorption, the installation on the roof and simplifying the way the card is installed inside the device.
- Simplifying the electronics used on the cards to improve reliability through the use of few active components.
The system permits at the time of a hurricane, to collect the necessary information for creating pressure curves that indicate the points to be reinforced on the roof.
References:
[1] Michot, B., Full-Scale Wind Pressure Measurement Utilizing Unobtrusive Absolute Pressure Transducer Technology, MS Thesis, Department of Civil Engineering, Clemson University, December, 1999.
[2] Vickery, P. J., Wind Loads on the Aylesbury Experiment House: A Comparison Between Full Scale and Two different Scale Models, Master of Science Thesis, University of Western Ontario, London, Canada, September, 1984.
[3] Marshal, R. D., Aerodynamics of Structures and wind Tunnel Modeling, National Bureau of Standards, Washington DC, November 1973.
[4] Marshal R. D., Reinhold T. A., and Tieleman, H. W., Wind Pressures on single-Family Dwellings, ASCD-EMD Specialty conference, dynamic Response of Structures: Instrumentation, Testing Methods and System Identification, University of California, Los Angeles, CA, 1976, pp. 228-242.
[5] Marshal R. D., The Measurement of wind Loads on a Full-Scale Mobil Home, Center for Building Technology, Institute of Applied Technology, National Bureau of standards, September 1977.
[6] Caballero A., Mitrani J., Arencibia L. Modification for the Sensors Field Measurement Applications. Report presented to the IHC at Florida International University, June 2002.
[7] Reinhold T., Caballero A., Field Measurements of Wind Loads on Flat Roofs. Report presented to the International Hurricane Center at Florida International University, Florida, USA. July 2003.
[8] National Instruments. LabView Version 7.0
Sensor # 1
------Ethernet Link
------
Sensor # 16
Anemometer
Figure 2. System Structure
Figure 3. The Virtual Instrument Screen
Table 1. Calibration Equations for FIU Field Measurement Sensors
Sensor Number / Slope of Calibration / Calibration Offset / Goodness of Fit – R2FIU 01 / 22.693 / -223.5 / 0.9992
FIU 02 / 22.856 / -200.1 / 0.9992
FIU 03 / 26.762 / -277.3 / 0.9992
FIU 04 / 21.697 / -182.3 / 0.9988
FIU 05 / 22.813 / -208.6 / 0.999
FIU 06 / 26.774 / -279.4 / 0.9993
FIU 07 / 21.732 / -182.5 / 0.9988
FIU 08 / 21.588 / -187.6 / 0.9988
FIU 09 / 26.816 / -279.9 / 0.9991
FIU 10 / 22.8 / -187.4 / 0.9992
FIU 11 / 26.952 / -272.9 / 0.9985
FIU 12 / 26.809 / -281.7 / 0.999
FIU 13 / 20.536 / -206.7 / 0.9964
FIU 14 / 20.511 / -212.1 / 0.9964
FIU 15 / 26.81 / -282.3 / 0.9992
FIU 16 / 22.908 / -196.6 / 0.9991
Table 2. Flat Roof Run
Sensor # 1 / Sensor # 2 / Sensor # 3 / Sensor # 4 / Sensor # 5 / Sensor # 114.423 / 13.141 / 4.142 / 11.012 / 58.019 / 14.423
14.423 / 13.141 / 4.142 / 11.012 / 58.019 / 14.423
Sensor # 6 / Sensor #7 / Sensor # 8 / Sensor # 9 / Sensor # 10 / Sensor # 6
-24.411 / 5.933 / 11.216 / -25.317 / 13.338 / -24.411
-24.646 / 5.933 / 11.216 / -25.317 / 13.338 / -24.646
Sensor # 11 / Sensor # 12 / Sensor # 13 / Sensor # 14 / Sensor # 15 / Sensor # 11
-41.989 / -281.359 / -6.482 / -7.537 / -18.201 / -41.989
-41.989 / -281.359 / -6.482 / -7.537 / -18.201 / -41.989
Sensor # 16 / Wind Direction / Wind Speed
15.791 / 0.12 / 0.001
15.59 / 0.12 / 0.001