Effects of obstructions, sample size and sample rate on ultrasonic anemometer measurements underground

In fluctuating airflow, continuous air velocity recording is the most reliable method of air velocity measurement. It allows for fast recognition of changes and the calculation of long-term averages. Also, it enables the mine operator to identify when the airflow has decreased to a point requiring action. Using ultrasonic anemometers provides an accurate option for continuous air velocity monitoring. This paper provides information about the effect of common obstructions in underground mining on air velocity readings. Stationary and moving obstructions are used to represent workers and equipment that would cause discrepancies in measured airflow. Also, it is important to know how large of a sample size is required to ensure reasonable accuracy of results. Statistical analysis is used to evaluate the required sample size. The sampling procedure is further studied by comparing two different sample rates. The results show that obstructions provide noticeable differences in air velocity measurements. Also, movement of obstructions can be recognized from changes in results. Surprisingly small sample sizes provide reliable air velocity information. Standard sample rates are found to be suitable for the underground environment. Introduction Ultrasonic velocity transducers have been used extensively in fluid flow applications, but are a relatively new addition to the underground mine environment. The operation of an ultrasonic instrument is based on the principle that the speed of a sound pressure wave varies with the local air speed. The air velocity is calculated from measurements of air-pulse transit times between sound transmitter and receiver (Hall et al., 2007). Ultrasonic anemometers have a linear response to airflow and an absolute calibration that depends only on sensor spacing and transit time measurement accuracy (Taylor et al., 2004). As opposed to conventional forms of velocity measurement, this technique requires no correction for air density, there are no moving parts to wear and there are no start-up friction or inertial problems when the air velocity changes rapidly (Casten et al., 1995). An important advantage of this method is the ability to provide a directional sign to the air velocity. Ultrasonic instruments fall into two distinct categories: variable-distance instruments and fixed single-point instruments (Casten et al., 1995). Variable-distance instruments consist of two ultrasonic transceivers mounted on each side of an airway, pointed axially towards each other and measuring the difference in time of flight. Variable-distance instruments are limited to one-axis measurements, but are often capable of calculating airflow for a known area. Fixed-distance units operate on the same principle as variable-distance units, but the measurement is performed inside one unit, with a typical sensor array of about 0.2 m (0.7 ft). Fixed-distance ultrasonic anemometers are categorized as one-, twoand three-axis. From the fixed-distance ultrasonic anemometer options available, the one-axis instrument measures flow in one direction, the direction of instrument orientation. Thus, it has the same limitations as the current standard vane anemometer, which is dependent upon orientation with respect to airflow. The two-axis instrument measures flow velocity in a plane defined by the U and V flow components in a direction relative to a reference direction. The three-axis instrument measures flow in a three-dimensional space defined by the U, V and W components of flow. Some new additions to the variable distance instrument category have been developed recently in Canada for fan airflow monitoring in underground mines (Synergy Controls Corporation, 2010; Accutron Instruments, 2010). Fixed, single-point instruments currently available are primarily for meteorological and research purposes (R. M. Young Company, 2009; Gill Instruments, 2009; Vaisala, 2009). Coal mine operators can use an ultrasonic instrument to help them comply with ventilation requirements, such as air velocity and direction in the belt entry and total air quantity in the belt entry and primary escapeway (Martikainen et al., 2010). However, ultimate compliance is determined by the U.S. Mine Safety and Health Administration (MSHA) by taking traverse velocity measurements in the entries using a vane anemometer. Currently, some ultrasonic anemometer manufacturers have either applied for or are looking into applying for MSHA certification of permissibility, as defined under 30 CFR § 75.506, to enable the use of their instruments in U.S. underground coal mines. Several instruments have been classified as intrinsically safe, based on other certifications in countries including Poland, Canada and the U.K. Permissible, MSHA-certified instruments could be used in return airways of coal mines in the U.S. instead of only in fresh air. This study evaluates the feasibility of single-point twoaxis and three-axis ultrasonic anemometers for air velocity measurements in underground coal mines. The specific issues addressed by the study are the effects of common obstructions underground, required sample size to achieve accurate results and an adequate sample rate. Test setting Tests were performed underground in three locations of the U.S. National Institute for Occupational Safety and Health (NIOSH) Bruceton Experimental Mine, which is driven into the Pittsburgh coal seam. The first location (Location 1) is in a long, straight section of a tunnel with a cross-sectional area of 5.3 m2 (57 sq ft). The second test location (Location 2) is in a curve of about 45°. The cross-sectional area of the tunnel is about 7.7 m2 (83 sq ft). Locations 1 and 2 were chosen to represent airflow in a straight and curved tunnel and because of instrument cable length restrictions. Location 3 is in an entry to an opening used to run cables through a bulkhead. The cables block part of the opening, causing a very uneven airflow. Also, the change in cross-sectional area was expected to result in high turbulence at measurement Location 3. The cross-sectional area of Location 3 is 3.0 m2 (32 sq ft), and the cross-sectional area of the opening is 0.7 m2 (7.5 sq ft). All instruments were located in the larger area. Air velocity was varied in the mine by opening and closing doors and changing fan settings. All testing locations are shown in Fig. 1. Figure 1 — Underground test locations. At all three test locations, the ultrasonic anemometers were set up in the tunnels with the three-axis instrument positioned between the two-axis instruments. The three-axis anemometer was kept stationary throughout the measurements in all locations, while the two-axis instruments were attached to adjustable poles with swinging arms. This allowed for point measurements to be taken at several spots across the entry. A total of six measurements were taken across the entry simultaneously, two at a time with two two-axis instruments at Locations 1 and 2. Figure 2 shows the ultrasonic anemometer setup for testing in Locations 1 and 2. The same setup was used during a previous study (Martikainen et al., 2010). A similar setup was used at Location 3, but due to a smaller cross-sectional area, the swinging arms were not turned and only two heights, high and low, were used for the two-axis instruments. As a result, only five measurements were taken over the cross-section at Location 3 (two two-axis measurements on each side, plus a three-axis measurement in the middle), while in Locations 1 and 2 the number of measurements was 13. A 180-s data collection time was used with the ultrasonic anemometers. The sample rate used was one sample per second (one sample/s) except for the sample rate study, during which a sample rate of four samples/s was used. These sample rates were readily available in all measurement instruments. Davis rotating vane anemometers were used to measure air velocities for comparison. Three vane anemometer traverses of 60 s were taken in all locations to compare with the averages of the 180-s data measured by the ultrasonic anemometers. The 180-s data collected by the three ultrasonic anemometers at all 13 points in Locations 1 and 2, as well as at five points in Location 3, was averaged for reliable comparison with the vane anemometer traverse results. The vane anemometer was used with an extension rod to minimize errors caused by measurement-taker proximity. The averages of the vane anemometer readings were compared with the averages of the results obtained by the ultrasonic anemometers. Obstruction analysis Obstruction test setup. A potential difficulty with continuous airflow monitoring may arise with the presence of obstructions (i.e., equipment and personnel) upstream of an anemometer station. Depending upon the distance from the anemometer, such obstructions can interfere with the flow around the sensor head and may generate vortices and eddies that seriously impact the accuracy and stability of the output of this instrument. A series of airflow evaluations with two different types of obstructions, a person and a personnel and equipment carrier cart, were conducted to examine the impacts of both stationary and mobile obstruction at two different distances from the instrument location. Two different stationary obstructions were tested with two airflows at Locations 1 and 2. The first obstruction (Stationary Obstruction 1) was a test subject with a height of 1.75 m (5.9 ft) and an approximate average width of about 0.4 m (1.3 ft) in full mine gear. Stationary Obstruction 1 was placed between the instrument poles. To distribute the effect evenly between instruments, the obstruction was moved to four locations from left to right in the entry, as follows: (1) between rib and left pole, (2) between left pole and the three-axis instrument, (3) between the three-axis instrument and right pole (Fig. 3) and (4) between right pole and rib. Stationary Obstruction 2, an electrician’s personnel and equipment carrier cart, was placed in front of the instrument setup, 3.0 m (10 ft) upstream from the instruments. The height of this cart is about 1.17 m (46 in.) and the width is 1.14 m (45 in.). The cart is shown in Fig. 4. The test subject was also used as a moving obstruction. The test subject moved with a steady pace of about 0.25 m/s (50 fpm) back and forth across the entry at two different distances, first at 1.2 m (4 ft) (Moving Obstruction 1) and then at 3.0 m (10 ft) (Moving Obstruction 2) upstream from the instruments. Tests were performed at Locations 1 and 2 with two different airflows. Figure 2 — Anemometer placement for Locations 1 and 2. Figure 3 — Stationary Obstruction 1 between two-axis and three-axis ultrasonic anemometers in Location 2. Figure 4 — Stationary Obstruction 2 upwind of Location 1. Results of obstruction testing With both airflows in Location 1, Stationary Obstruction 1 increased the average air velocity measured across the entry by about 15%. An air velocity increase was also observed at Location 2. The change in air velocity corresponds to an area decrease of about 0.7 m2 (7.2 sq ft) at both locations, which correlated well with the size of the obstruction. Similar air velocity differences due to the effect of Stationary Obstruction 1 were observed in previous tests when comparing the results of a rotating vane anemometer to the results obtained by the ultrasonic anemometer setup (Martikainen et al., 2010). Even when an extension rod was used according to the suggested practice of keeping a minimum distance of 0.9 to 1.2 m (3 to 4 ft) between operator and instrument (Boshkov and Wane, 1955), with care taken to keep the anemometer upstream of the measurement taker, all the vane anemometer readings were significantly higher than those recorded by the ultrasonic anemometers, ranging from 13 to 22%. These differences were recognized to correlate well with the size of the measurement taker and were comparable to the effect of Stationary Obstruction 1 on the ultrasonic anemometer measurement results. Stationary Obstruction 2 affected the airflow in Location 1 in the same way as Stationary Obstruction 1. In this case, the measured change in air velocity was about 22%. The expected air velocity increase from the free cross-sectional area decrease caused by the obstruction based on a calculation was 1.4 m2 (14.9 sq ft) and the measured amount was 1.3 m2 (14.4 sq ft). This discrepancy can be explained by the circumstances in Location 2 being more complicated. The average change in air velocity was different for every instrument. In some cases, a slight decrease in air velocity was observed instead of an increase. It was determined that the stationary obstruction caused an air velocity distribution change over the entry cross section, moving higher airflow from the left side of the entry to the right. Also, the three-axis instrument recorded noticeable W-axis values (air moving upor downwards). Moving Obstruction 1 decreased the average air velocity values in comparison to the air velocity with no obstruction. Comparison of the ultrasonic anemometer outputs with the positions of the obstruction showed a significant decrease of airflow immediately after Moving Obstruction 1 had passed the instrument. After this disturbance, the air velocity quickly returned to slightly above that measured with no obstruction. As the obstruction moved only 1.2 m (4 ft) away from the instruments, the drop in air velocity due to the disturbance was large enough to cause a decrease in the average air velocity. The close proximity of the obstruction to the instruments, the turbulence caused by its movement and the relatively small size of the obstruction were considered the main reasons for the air velocity decrease. The disturbances are easily recognizable in comparison to the undisturbed airflow and are shown in Fig. 5. In this figure, the impact of the obstruction on the airflow can be clearly seen as pronounced dips in the plot. It is important to realize that the readings quickly returned to near normal, unobstructed levels once the obstruction passed the anemometer. Similar results were observed for Moving Obstruction 2. In this case, however, the average air velocities were approximately the same as with no obstruction. The disturbances caused by the obstruction moving further away from the instruments were not long enough or large enough to affect the averages. An example of the air velocity averages with no obstruction and with obstructions from all three instruments is shown in Fig. 6. Air velocity values recorded by the two-axis and three-axis instruments are very similar in all cases, except for Stationary Obstruction 2. This result shows that in a straight airway, even with an obstruction, the performance of a two-axis ultrasonic anemometer is comparable to the performance of a three-axis instrument. The large difference with Stationary Obstruction 2 can be explained by instrument positioning. Directly behind the obstruction, air velocity was decreased, but on both sides, air velocity increased. This was detected by the instruments when the three-axis instrument was placed behind the obstruction with the two-axis instruments located on both sides. Figure 5 — Results showing airflow with no obstruction compared to airflow disturbed by Moving Obstruction 1