Barrow 2000 Icing sensor and heated pitot tube data

Document Sample
Barrow 2000 Icing sensor and heated pitot tube data Powered By Docstoc
					                                    .




Barrow August 2000 Operation:

   Icing Sensor Data report




       September 21, 2000




            Written By:

       Dan Thomas, BSEE
     thomasda@ece.orst.edu
        Engineering Intern

            Cloud Cap
              Technology
     PO Box 1500, No. 8 Fourth St
        Hood River, OR 97031
        Phone: 541.387.2120
         Fax: 541.387. 2030
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

1 Introduction
      The Barrow 2000 Aerosonde Operation occurred August 15-28, 2000 in Barrow, Alaska. The
      operation was conducted in support of NSF contract # 99-101, Aerosonde Operations in the
      Arctic. Detailed background information of the operation can be found in the Barrow
      Planning document.

      One of the research and development objectives of the operation was to conduct
      experiments to improve the robustness of the Aerosonde in Arctic conditions. In support of
      this objective, data was collected from two icing sensors to test their ability to detect ice
      accretion and to correlate their readings.

      The purpose of this report is to analyze the icing sensor tube data collected during the
      Barrow operation. Recommendations and conclusions will be made about the icing sensors
      based on test data.

2 Icing Sensors
      Two icing sensors were evaluated during the Barrow 2000 operation. One was constructed
      around a piezo element and the other was a commercially available sensor, modified
      especially for the operation.

2.1    Piezo Icing Sensor
       The icing sensor currently used by Aerosonde is constructed with a piezoelectric element
       in the feedback path of an oscillator circuit. The piezo element is placed in the air stream
       along the leading edge of the aircraft’s wing and as ice builds up on its surface, its
       resonant frequency increases. In turn, the oscillation frequency of the sensor circuit (i.e., its
       output signal) also increases.

2.2    ODS Icing Sensor
       Optical Detection Systems (ODS) of Littleton, Colorado manufactures the other sensor
       evaluated during the Barrow operation. It is commercially available and is designed for use
       in passenger and personal aircraft.

       The sensor consists of a control unit and a sensor probe. The sensor probe contains a
       control surface monitored with two parallel, redundant optical beams with corresponding
       detectors. The logic inside the control unit monitors intensity of the beams to detect ice
       buildup. The parallel beams are used to protect against false readings of ice accretion. The
       sensor probe also contains temperature sensors used by the control unit logic to verify that
       the air temperature is below a threshold temperature (i.e., 10C) where icing could occur.

       The output of the sensor’s control unit used during the Barrow August 2000 Aerosonde
       operation was specially modified to provide a continuous DC output voltage. In effect, this
       modification allows the ODS icing sensor to be used as an ice accretion sensor, in contrast
       to its original design as an ice detector.

       The ODS sensor used during the Barrow operation was also specially modified to disable
       the deicing mode, used to clear the control surface of the sensor probe of ice, because the
       mode required a significant amount of power (54 Watts, 4.5 Amps at 12V). Therefore, the
       sensor output displays only one icing “cycle” (i.e. the thickness of ice detectable without
       deicing the control surface).




                                             Page 2 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

3 Baseline Data
3.1   Piezo Icing Sensor
       A historical plot of the response of three different piezo sensors is shown in Figure 2 on the
       next page. This plot is the best illustration of the piezo sensor’ s response to ice accretion
       prior the Barrow operation.

3.2   ODS Icing Sensor
       Two freezer tests were also run with the ODS icing sensor prior to the Barrow operation to
       characterize its response. Figure 1 shows the results of the first test and Figure 3 shows
       the results of the second test. Each data point represents the application of two squirts of
       water from a water bottle. The bottle dispensed 0.7ml of water per squirt The DC output
       voltage of the ODS sensor was measured using a digital voltmeter. Each test concluded
       when the ice build-up of the sensor probe triggered the “ICE” indicator, an LED, on the
       ODS sensor control unit. Based on the two tests, the average value of the ODS output
       voltage when the unit triggered was 0.67 Vdc.

       Note also that because the continuous DC output voltage modification of the sensor was
       done exclusively for Aerosonde no baseline data was available from the manufacturer.
       Thus, these freezer tests served as baseline data for comparison to the Barrow test
       results.




                                            Page 3 of 31
     Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Figure 2: Historical Plot Of The Piezo Icing Sensor’s Response To Ice Accretion




                                         Page 4 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-




4 Barrow Test Descriptions
4.1   Test Location
       The icing sensor tests were conducted at the Naval Arctic Research Laboratory (NARL)
       airstrip in Barrow, Alaska.

4.2   Mounting location of sensors, data logger, and special battery pack
       The ODS icing sensor probe, which weighed 113.4 grams, was mounted at the center of
       the wing of the test Aerosonde aircraft (“Heimie”, aircraft 56). It was mounted using a
       double-sided adhesive and rubber wedges to compensate for local curvature of the wing.
       Figure 4 is a photograph that displays of the mounting location of the sensor probe on the
       test aircraft. The control unit for the ODS sensor weighed 184.3 grams and was mounted
       directly behind the flight avionics box in the fuselage of the test aircraft. The photo in Figure
       4 displays the mounting location of the ODS control unit.

       The piezo-element icing sensor (which weighs 4.3 grams) was mounted along the leading
       edge of the test aircraft’s wing, near the center of its span. The piezo-element icing sensor
       was self-contained, with its control circuitry located behind the sensor “probe” (the piezo
       element). The mounting location of the piezo sensor on the test aircraft is also shown in
       Figure 4.




                                             Page 5 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


                                            ODS Sensor Probe




                                          Piezo icing sensor




             Figure 4: Photo of the mounting location of the ODS icing
                       sensor probe and piezo icing sensor.




                                          ODS sensor
                                          control unit




        Figure 5: Photo of the mounting location of the ODS sensor control unit.

  The data logger used to record data from the icing sensors (as well as temperature data
  from the heated pitot tube prototypes, also tested in Barrow using this aircraft) was
  placed above the fuel tank of the test aircraft. A special battery pack, intended for
  supplying power to the data logger, heated pitot tube prototypes, and icing sensors, was




                                    Page 6 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

         also placed above the fuel tank near the rear bulkhead. See Figure 6 for a photo of the
         mounting locations of the data logger and the special battery pack.




                                  Data logger

                                                             Special battery pack




           Figure 6: Photo of the mounting location of the data logger and battery pack (area
                                   below the wing of the test aircraft “Heimie”)

4.3    Electrical Connections Of The Icing Sensors
       The ODS icing sensor was powered off a special battery pack mounted in the test aircraft
       for the heated pitot tube prototypes (also tested during the Barrow 2000 operation), the
       data logger, and for the icing sensors. The sensor probe was connected to the control unit
       via a cable routed through the wing and center bulkhead of the plane. The control unit
       generated the sensor’s continuous DC output voltage using input signals from the sensor
       probe. This continuous DC output voltage was then routed to the data logger via a cable
       through the center bulkhead, and recorded at a rate of one sample point per second.

       The output of the piezo-element sensor was connected to the data logger via a cable
       routed under the wing. The output (a frequency) of the sensor was recorded at a rate of
       one sample point per second.

5 Data Plots
      Data from the icing sensors was collected in Barrow during flights of the test aircraft “Heimie”
      on August 20,21,22,25, and 26, 2000. A total of ten flights were made with the test aircraft
      during these dates. Flight times and altitude were chosen to improve the likelihood of aircraft
      icing during the test flights.

      Data from six of the ten flights was selected for analysis because the test aircraft returned
      with airframe icing and the icing sensors displayed noticeable response to this icing. This
      data is presented on the next 12 pages, with two pages per test The plot on the top of the
      first page for each test flight displays the complete ODS and piezo icing sensor data sets
      collected during the flight, plotted on different y-axes. The next plot displays outside air
      temperature (OAT) and humidity data recorded by the RS90 sondes during the flight and
      generated for plotting by replaying the flight telemetry files using Ground Base.

      The plots on the second page are detailed views of the piezo and ODS icing sensors data.
      These plots show pertinent times where the sensors were reacting to ice accretion.


                                             Page 7 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

5.1    Data Plots: Flight #1, August 21, 2000
                                        Figure 7




                                        Figure 8




                                     Page 8 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #1, August 21, 2000 (continued)-
                                      Figure 9




                                      Figure 10




                                    Page 9 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

5.2    Data Plots: Flight #2, August 21, 2000

                                       Figure 11




                                       Figure 12




                                     Page 10 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #2, August 21, 2000 (continued)-
                                      Figure 13




                                      Figure 14




                                   Page 11 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

5.3    Data Plots: Flight #3, August 21, 2000
                                       Figure 15




                                       Figure 16




                                     Page 12 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #3, August 21, 2000 (continued)-
                                      Figure 17




                                      Figure 18




                                   Page 13 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


5.4    Data Plots: Flight #4, August 21, 2000
                                       Figure 19




                                       Figure 20




                                     Page 14 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #4, August 21, 2000 (continued)-
                                      Figure 21




                                      Figure 22




                                   Page 15 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


5.5    Data Plots: Flight #1, August 25, 2000
                                       Figure 23




                                       Figure 24




                                     Page 16 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #1, August 25, 2000 (continued)-
                                      Figure 25




                                      Figure 26




                                   Page 17 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


5.6    Data Plots: Flight #2, August 26, 2000
                                       Figure 27




                                       Figure 28




                                     Page 18 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


Data Plots: Flight #2, August 26, 2000 (continued)-
                                      Figure 29




                                      Figure 30




                                   Page 19 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-



6 Observations From The Data Plots
6.1    Flight #1, August 21, 2000 Test Plots (Figures 7-10)
      The piezo element sensor’s output, as observed in Figure 7 and in relation to OAT and
      humidity data shown in Figure 8, had a nearly monotonic response to ice accretion during
      the flight. A magnified view of the sensor’s response is shown in Figure 9. The sensor’s
      output frequency started at its non-iced value of 4985 Hz and ramped to a maximum
      frequency of 7750 Hz. Light icing was noted on the airframe at the end of the flight.

      ODS sensor’s data curve, also observed in Figure 7 and in relation to OAT and humidity
      data in Figure 8, has two possible interpretations. The first interpretation was that once
      OAT and humidity was suitable for icing, the sensor reacted to ice build-up before the piezo
      sensor. Ice accretion was tracked until it icing stopped between 17:52 and 18:04. Icing then
      resumed until the sensor reaches an output “saturation” level at 18:14. The output of
      sensor had a non-monotonic response to ice accretion.

       The other interpretation of the ODS sensor’s data curve is that it displayed temperature
      dependence. The sensor’s response appeared to track the OAT changes, with a time
      delay. For example, when the OAT transitions to –1.5 C at 17:42, the output of the sensor
      transitions from 0.95 to 0.85 Vdc. These interpretations will be explored further in the
      remaining graphs.

      Another observation from the ODS sensor’s curve is a noticeable, periodic measurement
      “noise”. Figure 31 shows a highly magnified view of the ODS sensor’s response. A distinct
      50-millivolt peak-to-peak, 0.012 Hz signal is observed riding on the sensor’s output.
      Because the noise was not observed on other data collected by the data logger, the ODS
      sensor circuitry is the suspected source of the noise.

                                                       Figure 31

                                Detailed view of ODS sensor measurement "noise" -
                                           Heimi flight #1, August 21, 2000
                        0.925


                          0.9          period of measurement oscillation = 1 min,26 sec.
                                                  (i.e. frequency of 0.012 Hz)
                        0.875
              Voltage




                         0.85


                        0.825


                          0.8


                        0.775
                           18:00:00   18:00:43   18:01:26   18:02:10   18:02:53   18:03:36   18:04:19
                                                       Tim e (UTC/GMT)




                                                    Page 20 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-


6.2    Flight #2, August 21, 2000 Test Plots (Figures 11-14)
      The piezo icing sensor’s response was nearly monotonic and ramped from 4985 Hz, its
      non-iced output frequency, to 11,400 Hz (best observed in Figure 13). Icing was again
      observed on the test aircraft at the end of the flight.

      As in previous test, the ODS sensor’s output responded faster than the piezo sensor (the
      ODS sensor’s output began to change 4 minutes before the piezo sensor’s output). It’s DC
      output voltage ramps from 0.95 Vdc with no ice to a final value of 0.74 Vdc. Note this final,
      saturated output value was different than the previous test, indicating more icing has
      occurred. Also note that though the piezo sensor continued to respond ice accretion
      (indicated by the changes in it output frequency) the ODS sensor’s output remains at this
      value. The ice thickness had reached the detectable thickness possible without clearing the
      control surface (recall the heater inside the ODS sensor probe was disabled for the Barrow
      operation).

6.3    Flight #3, August 21, 2000 Test Plots (Figures 15-18)
       The piezo icing sensor ramped from its non-iced frequency to a final output frequency of
       11,000 Hz before entering a period of “over-frequency”. The piezo sensor circuit is
       venerable to unstable modes, an inherent characteristic of oscillators. The observed over-
       frequency mode could have been caused by instability of the piezo sensor circuit or a
       build-up of ice to the point where the piezo element could not vibrate.

       The test aircraft returned with 7.2-mm of ice build-up on the leading edge of the wing and
       piezo sensor. There was no audible output from the piezo sensor, which indicated either
       the output frequency was above the human hearing range or that the sensor had stopped
       vibrating. The latter thought was suspected state of the sensor. See section 7 Visual Data
       for photos of the piezo icing sensor after this flight.

       The ODS sensor’s output responded three minutes before the output of the piezo sensor.
       The DC output voltage ramped from 0.95 Vdc to 0.75 Vdc in two minutes. Note that during
       between 01:09 and 01:16 the output ODS sensor had a large amount of aperiodic noise
       (unlike the low frequency noise still riding on the output signal and very evident in the
       output plot shown in Figure 18). This noise is attributed to intermittent heater connection for
       the heated pitot tube that was also tested during this flight. See the plot in Figure 32 of pitot
       tube temperature measurements recorded during this flight to observe a similar period of
       instability.

       The test aircraft returned with 9-mm of ice build-up on all ODS sensor surfaces facing the
       air stream. The larger amount of ice build-up on the sensor in comparison to the leading
       edge is evident in the photographs of the icing sensors shot following this flight (see
       section 7) and most likely due to its mounting location.




                                            Page 21 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

                                                                   Figure 32

                                                   Heated Pitot tube Prototype #1, test # 1:
                                              ("Prototype #1b", Heimi flight #3, August 21, 2000)
                                   40


                                   30


                                   20
               Temp erature (°C)




                                   10


                                     0
                                    0:43:12       0:57:36      1:12:00      1:26:24     1:40:48     1:55:12

                                   -10


                                   -20
                                                                Tim e (UTC/GMT)


6.4    Flight #4, August 21, 2000 Test Plots (Figures 19-22)
       The piezo icing sensor again had a period of almost monotonic output response from the
       beginning of the flight at 02:42 until 03:06, during which the output frequency varied from
       4985 to 11,750 Hz. At 03:06 the sensor went into an over-frequency mode until the end of
       the flight. The test aircraft returned with 2.4-mm of ice on the leading edge and piezo
       sensor, and 4-mm of ice on the ODS sensor.

       On this flight the ODS sensor’s response appeared nearly monotonic, ignoring the already
       noted low frequency, periodic noise. The output ramped from 0.95 Vdc to 0.75Vdc in 3
       minutes, as illustrated in the plot shown in Figure 22.

6.5    Flight #1, August 25, 2000 Test Plots (Figures 23-26)
       The piezo sensor behaved “normally” on this flight. That is, the output frequency ramped
       from the non-iced output frequency to 9500 Hz and didn’t enter an unstable mode. Note
       from Figure 24 that icing conditions were not optimal during this flight (we found an OAT at
       or below -4C, and with humidity above 90% provided the best icing conditions) and thus
       only relatively small amount of ice was measured on the airframe after the flight (i.e., only
       2.5-mm of ice was measured on the leading edge when the aircraft returned).

       The ODS icing sensor also behaved very similarly to previous tests. The output voltage
       ramped from 0.90Vdc to 0.75 Vdc during the flight. The transition region of the output
       voltage occurred 22:36 to 22:40 (i.e., a 4 minute transition time).

6.6    Flight #2, August 26, 2000 Test Plots (Figures 27-30)
       The icing sensor only ramped to 6000 Hz before going into an unstable mode on this flight.
       Note from Figure 28 that very high humidity conditions occurred during this flight and that
       plane returned with 10-mm of ice build-up on the leading edge.

       The ODS sensor ramped from 0.925 to 0.750 Vdc, during a 4 minute transition region.




                                                               Page 22 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-



7 Visual Data
7.1    Photos
      All of the photographs shown in this section were shot following the third flight of the test
      aircraft “Heimie” on August 21, 2000. The aircraft returned with 7.2-mm of ice build-up on
      the leading edge of the wing and piezo sensor, and 9-mm of ice build-up on the ODS
      sensor. The photograph in Figure 33 shows ice build-up observed on both icing sensors
      following the flight. Figure 34 is a close-up photograph of the iced piezo icing sensor and
      photograph in Figure 35 shows a close-up view of the iced ODS sensor.




            Figure 33: A photograph of both icing sensors following the third flight on August 21, 2000




                                                Page 23 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-




    Figure 34: A photograph of the piezo icing sensor following the third flight on August 21, 2000




    Figure 35: A photograph of the ODS icing sensor following the third flight on August 21, 2000




                                          Page 24 of 31
      Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-



8 Other Observations
8.1    Maximum Piezo Icing Sensor Frequency Displayed By The Flight Avionics
       One of the other test aircraft used during Barrow operation (“Lincoln”) had been modified
       with the capability to monitor the piezo icing sensor output frequency during flight. It was
       observed that the measured output frequency never exceeded 8000 Hz before reading
       over-frequency values.

       In order to test the piezo sensor output frequency measurement technique used by the
       flight avionics the piezo icing sensor mounted on the icing test aircraft “Heimie” was
       connected the flight avionics on “Lincoln”. This piezo icing sensor was labeled “63”. It was
       known from preliminary data recorded by the data logger that the output of this sensor
       could read output frequencies up to 12,000 Hz. Next, a water bottle and a can of refrigerant
       were used to build-up ice on the sensor. The piezo output frequency displayed on Ground
       Base never exceeded 8000 Hz, confirming the results observed during flights of “Lincoln”.


9 Follow-Up Tests Of The ODS Sensor
      More tests of the ODS sensor were run following the Barrow operation to investigate the
      measurement noise observed in the Barrow flight tests and to further characterize the
      sensor. The ODS sensor probe was placed in a common household freezer with the ODS
      control unit located outside the freezer. The probe was placed in the freezer in order to cool
      it to a temperature where the logic in the control unit would assume ice could form (i.e., the
      temperature threshold for the control logic is 10 C). Next, individual pieces of tape were
      wrapped around the control surface, with the idea that the tape would provide a known
      thickness that could be correlated with the change in the ODS sensor’s DC output voltage.
      The data was recorded using the data logger used for the Barrow operation.

      ODS sensor was tested with three different kinds of tape – black electrical tape, masking
      tape, and clear “Scotch” tape. Different types of tape were used to investigate if the ODS
      sensor’s DC output voltage were affected by the opacity of the material on the control
      surface. Two trials were run for each type of tape. The thickness of a single piece of tape
      was measured using brass calipers as 0.1905-mm, 0.08255-mm, and 0.05715-mm, for the
      black electrical, masking, and clear tape, respectively.

      The results from the black electrical tape test are shown in Figure 36. Note the pattern of the
      overall data for each trial (shown in the upper graph boxes, one of each trial) closely
      resembles the characteristic pattern of the ODS DC output voltage observed during icing
      sensor tests in Barrow. Note also that the data from the test trials contains the same
      measurement noise observed in the Barrow data.

      The collected data was “filtered” to further correlate the ODS icing sensor’s DC output
      voltage to the thickness of tape applied to the control surface. That is, only one data sample
      recorded 10 seconds after the application of each piece of tape was plotted. This plot is
      shown in Figure 37. The sensor triggered with 5 pieces tape on the control surface, resulting
      a total thickness of 0.762-mm. The calculated thickness of black electrical tape on the
      control surface of the sensor probe is also noted on the plot. Because this tape is the
      thickest of the three types used for testing, the accuracy of the measurement is limited by its
      thickness.




                                            Page 25 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-




                Figure 36: ODS icing sensor test trial results for black electrical tape




   Figure 37: Graph correlating thickness of black electrical tape to ODS sensor DC output voltage




                                          Page 26 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

The results from the masking tape test are shown in Figure 38. These plots again resemble
the response of the ODS sensor to ice accretion as observed in the Barrow tests.
Figure 39 shows the filtered data plot. The ODS sensor triggered with 6 pieces of tape
applied to the control surface, resulting an overall thickness of 0.4953-mm.




                    Figure 38: ODS icing sensor test trial results for masking tape




                                       Page 27 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-




       Figure 39: Graph correlating thickness of masking tape to ODS sensor DC output voltage



The results from the clear tape test are shown in Figure 40. Figure 41 shows the filtered data
plot. The ODS sensor triggered with 13 pieces of tape applied to the control surface,
resulting an overall thickness of 0.743-mm.




                                         Page 28 of 31
Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-




                       Figure 40: ODS icing sensor test trial results for clear tape




       Figure 41: Graph correlating thickness of clear tape to ODS sensor DC output voltage




                                         Page 29 of 31
    Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

10 Conclusions And Recommendations
   The flight conditions Barrow operation provided ample conditions to test the piezo and ODS
   icing sensors. However, each sensor has problems that made neither a clear choice as an
   optimal icing rate sensor. The advantages and disadvantages of each sensor are
   summarized below. Recommendations of how to correct some of disadvantages of each
   sensor are also discussed.

10.1 Piezo Icing Sensor
     For test flights in Barrow this sensor provided predictable, nearly monotonic output
     frequency variation during ice accretion. This sensor is compact, lightweight, and
     inexpensive to build. The sensor also consumes very little power.

     The sensor did suffer from instability problems with heavy ice build-up. The heavy build-up
     caused the piezo element to go into an over-frequency mode or stop oscillating. Adding a
     small heater element and control circuitry to deice the piezo element surface could solve
     this problem. However, this increases the power consumption of the sensor.

     The piezo element doesn’t conform to the leading edge of the wing and thus causes
     undesirable side effects such as turbulence and drag. An alternate piezo material that can
     be conformed to the leading edge is currently being studied.

     As mentioned in last section, the interface circuit in the flight avionics used to measure the
     output frequency of the piezo sensor doesn’t properly represent the complete output
     frequency range of the icing piezo sensor. Modifying the flight avionics interface circuit or
     improving the measurement routine in the avionics software could remedy this problem.

10.2 ODS Icing Sensor
     The results of the Barrow test flights shows that the continuous DC output voltage
     modification made by ODSystems exclusively for the Barrow operation did effectively
     convert the sensor from an ice detector into a ice accretion rate sensor. The output voltage
     varied from 0.95 to 0.75 Vdc on most tests, giving a range of voltages for ice rate
     detection. The sensor output also showed changes in its level 3 to 4 minutes before the
     piezo sensor, though this result was probably due to its mounting location.

     A low frequency, small amplitude periodic noise was observed on all test flight data
     collected during the Barrow operation. Because this noise is periodic, it could be filtered.
     If the source could be identified it might be possible to eliminate the noise.

     The follow-up tests performed after the Barrow operation did confirm the measurement
     noise source is within the ODS sensor circuitry. These tests also provided further
     confirmation of the pattern of the sensor’s response to ice accretion. Further, these tests
     also provided data that correlated thickness measured by the sensor and the DC output
     voltage of the sensor. The results of these tests also indicate the sensor has some
     sensitivity to the opacity of the material deposited on the control surface.

     The ODS sensor probe and control unit were large relative to the size of the Aerosonde.
     The heater unit is the sensor probe consumed too much for practical use on long-term
     flight missions. A smaller, more lightweight version of the probe and control unit with
     reduced power consumption could be designed for the Aerosonde.




                                         Page 30 of 31
   Barrow 2000 Aerosonde Operation: Icing Sensor Data Report (continued)-

11 Short Term Icing Sensor Recommendation
  The piezo sensor is the best short-term choice for an icing rate sensor between the two
  sensors. However, with modifications to physical size and power requirements the ODS
  sensor, it could be a competitive solution.


12 Future Work
  Improvements to the piezo icing sensor required in the long term are circuitry modifications
  that would boost the dynamic range of the current sensor. Because this circuit was designed
  at Aerosonde, knowledge of the circuit already exists. This fact would make cost of circuit
  redesign relatively low.

  Another prospect for future piezo sensors is the use of the material PiezoFlex manufactured
  by AIRMAR Technology Corporation of Milford, New Hampshire. This material would
  conform to the wing’s leading edge, thus overcoming this problem with the current piezo
  sensor. A sensor constructed from this material would also be very lightweight and
  inexpensive to build.

  Improvements required for the ODS icing sensor in the long term would be reduced size,
  power consumption, noise, and increased dynamic range. One possibility for the physical
  redesign of this sensor would be to mount it into a removable wing “section” that could be
  added to the airframe easily and quickly. The sensor should conform to the leading edge of
  the wing in the section.

  Both sensors need to be tested in higher accretion rates than that experienced in Barrow
  during August 2000. Higher accretions rates are likely in more severe Arctic weather
  conditions.




                                      Page 31 of 31