Figures 1 & 2. Rear and side views of the Mk. II Cart showing the DAS equipment compartment (right, 32K jpg) and the sensor arrays at 0.1m, 0.6m, and 1.1m (left, 49K jpg). page links:
In 1990, we were approached by the Pacific Gas & Electric Company to evaluate thermal comfort in a set of buildings developed for the Advance Customer Technology Transfer (ACT2) Project. This endeavor followed an earlier thermal comfort assessment project sponsored by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) and offered an opportunity for us to improve the apparatus used in that project. The new Mark II cart would be battery powered for a full day's operation, automate the collection procedure, and conform to the measurement protocols of ASHRAE's 55-81 comfort standard. To meet these specifications, we assembled an array of sensors and transducers, constructed an integrated signal processing and data acquisition system, and programmed two laptop computers for data reduction and display in real-time. All equipment is mounted on a two-wheeled chassis of 3 inch by 1 inch rectangular aluminum tubing with a wooden "chair" attached to the front. In addition to battery storage, the "seat" of the chair shields the sensors and carries the laptop computer that issues the subjective survey questionnaire at each workstation visit. We had several signal processing devices custom-built to our specifications for this application and they are mounted under the seat and in the seat back. Behind the seat back is a second laptop computer that provides the cart operator with a real-time view of the transducer values and presents a stripchart-format time history of the previous ten minute's data.
The cart's sensors were chosen to meet the response time and accuracy requirements of ASHRAE Standard 55-81 and ISO Standard 7730 for thermal assessment. The sensors we used are YSI series 700 probes having a vinyl-coated tip on a flexible signal wire. The temperature sensors are accurate to within 0.2 degrees Celsius and have a time-constant of several seconds. Where globe temperature was measured, we mounted a table tennis ball on the cart with one of the YSI temperature sensors in the center of the "globe". The globe is painted gray for the proper emissivity and responds to the balance between radiation and convection in the physical environment. In an office environment where the differences between workstations are relatively small, the globe should reach equilibrium well within the 5 minute measurement period. A short discussion of globe temperature and its measurement is provided in Benton, Bauman and Fountain (1990).
Figures 3 & 4. View of a typical sensor array at 0.6 meters above the floor. From left to right you can see a shielded drybulb thermometer, the omnidirectional anemometer, and the grey globe thermometer. Above the drybulb thermometer is the air inlet for the chiller-mirror dewpoint sensor (right, 38K jpg). On the left is a closer view of the 1.1 meter sensor bar with similar devices (left, 36K jpg).
Air velocity was measured at three heights by Dantec 54R10 anemometers. The 54R10 is an omni-directional fully temperature-compensated sensor with a time constant of 0.1 seconds. A fast response time is essential for accurate estimation of turbulence in the airflow. Each sensor has two nickel-plated quartz spheres supplied with a small electrical current. The current heats the spheres which in turn are cooled by passing airflow. Velocity is measured by regulating the electrical current to maintain the spheres at a constant temperature. Dewpoint temperature is measured by a General Eastern DEW-10 chilled mirror dewpoint transducer. In this transducer, a heated chimney draws a small sample of room air into a measuring chamber where a small mirror is continuously cooled. A nearby LED shines a beam of light at the mirror where it is reflected to a photosensor. When the mirror reaches the dewpoint temperature of the air sample, water condenses on the mirror scattering the light beam so the signal to the photosensor is interrupted. The temperature of the mirror is then measured and sent to the central datalogger.
Radiant asymmetry is measured by a Bruel and Kjaer Plane Radiant Asymmetry sensor. Plane radiant temperature is defined as the uniform surface temperature of a hemisphere that produces the same incident radiation on a surface as the actual environment. Radiant asymmetry is the difference between the plane radiant temperatures of small planes facing opposite directions. The radiant asymmetry probe consists of two pairs of gold-plated and black-painted elements connected to thermopiles, mounted on opposed planes. The measurement is based on the fact that the gold element exchanges heat primarily by convection while the black element exchanges heat by both convection and radiation. Thus any voltage generated across the thermopiles can be attributed to heat transfer by radiation between the black element and the environment. Illuminance is an ancillary parameter that may be useful in later analysis. A cosine-corrected silicon photometer manufactured by LiCor measured illuminance in the horizontal plane.
The cart's data acquisition system consists of several signal processors feeding a central datalogger programmed to poll the sensors and relay the data to a laptop computer for display and storage. Since both the air velocity and air temperature sensors are inherently non-linear, signal processing is required to convert these measurements to engineering units. In the case of the temperature sensors, a linearization bridge on the signal side is required while more extensive circuitry is necessary for controlling the current supplied to the anemometers and providing temperature compensation. The signals from all transducers and signal conditioning are sent continuously to the heart of the system, a Campbell Scientific 21x datalogger. The datalogger measures the sensor signals and converts each to engineering units using polynomial curve fits or linear conversions as appropriate. The 21x is connected to a lightweight laptop computer that serves as data display, operator interface, and data storage device.
Figures 6 & 7. The cart's "chair back" contains the systems signal conditioning and data acquisition systems. From upper right clockwise you can see the dewpoint transducer, the power distribution board (wooden), the Campbell Scientific datalogger, and the thermistor signal conditioning (black) (left, 80K jpg). The back of the cart supports the laptop computer that provides realtime display of measured variables and archival data storage (right, 52K jpg). The Campbell Scientific datalogger also controls the timing and sequence of measurements. A data-collection sequence is initiated by the operator flipping a switch mounted on the top of the cart. This instructs the datalogger to begin a data collection sequence, a state indicated by an LED glowing solid green near the cart switch. The laptop computer continuously displays data in a stripchart fashion with an indicator showing whether the data are being stored or not. After one minute of monitoring the transducers as they come into equilibrium with the physical environment at the workstation, the datalogger shifts into "burst" mode. Burst mode is the only state in which the datalogger can sample the anemometers quickly enough to measure turbulence intensity. During the next three minutes, the datalogger is completely occupied with the air velocity measurement, collecting 60 readings per second while the LED blinks green. After the burst measurement is complete, the 21x collects data from the other sensors at the rate of one sample per second for the remaining one minute. During the last minute of data collection, the LED glows solid red while the system calculates averages from the one-second data. Finally, the system collects measurements from the globe thermometers after they have had a 5-minute period to equilibrate. The cart then turns off the red LED to indicate that the measurement sequence is complete and the cart can be safely moved. The total number of air velocity readings taken during the three minute measurement burst is 7,200, too much data to process in real-time. So, as the cart is being moved to the next workstation, a post-measurement processing sequence reduces the 7,200 readings to engineering units, calculates turbulence intensity, and stores the final values on the hard disk with data from the other transducers.
The Mark II cart was designed in late 1990 by Professor Charles C. Benton and graduate student David Lehrer. Construction began in 1990 and was completed by master craftsman Robert Marcial, then an undergraduate student at Berkeley. Research Associate Fred Bauman managed the specification and procurement of measurement components. Ph. D. student Marc Fountain worked with Benton on modification of the data acquisition components and programming of the system. The cart became operational in April 1991 and has performed well ever since. In addition to serving in Berkeley's Controlled Environment Chamber, the cart has been used extensively in the Bay Area, in Michigan (twice), and in Australia (twice). The cart was designed so that all sensitive components and wiring can remain within the central plane of the "chair back". This allows relatively easy disassembly of the cart for shipment by air.
Figures 9 & 10. Cart contributors Robert Marcial (left, 56K jpg) and Marc Fountain (left, 69K jpg).
|
Figure
5. The B&K Radiant Asymmetry sensor pivots on its vertical axis to face local
sources of radiation. The sensor reports the difference in plane radiant temperature of
opposing hemispherical fields. (24K jpg).
Figure 8. The cart's components laid out before final
assembly in March 1991 (74K jpg).
