Translate

Monday, May 26, 2014

Testing for Evidence of Fukushima : Part V

This is my third attempt at testing my air filters with my Aware Electronics RM-80 this year. My first two attempts (1, 2) produced ambiguous results. However, I have refined my test methods. Let me give a fuller explanation of my testing and test environment. I have a now ten year old Carrier 'Infinity 16'  zoned heat pump system that provides continual fresh air. This is recommended for allergic asthmatics by some and I find that by and large I and our family is healthier with it. However, my filters trap lots of what appears to be 'carbon black' or essential carbon. I assume most of this is from wood burning in Bellingham which I discuss in detail  here.  Code for this post can be found  here and here.

My hypothesis is that the fallout from naval submarine bases like Bangor-Kitsap and most recently Fukushima has resulted in increased levels of radioisotopes in the air, water, hydrological cycle, and in the smoke derived from the wood burned in many homes in the Pacific Northwest for heat. My testing device (Aware Electronics RM-80) is a sensitive and capable Geiger-Mueller counter. However, I can not directly identify radioisotopes without the purchase of an expensive gamma spectrometer. Another hypothesis I have is that I should be able to see an increase in radioactivity as the months progress and the plume of radiation from Fukushima becomes more and more part of our water cycle.

Because I am testing inside air filtered by mulitple layers of air filters, my base line probably suffers some bias. I could use a more portable Geiger-Mueller counter. Come to think of it, a rad lab, a small team of scientists, and multiple outdoor stationary test facilities might be nice as well. Maybe I will pitch something at KickStarter.comIn any event, my most recent filter testing produced good evidence that the air outside my house may have higher levels of radioactivity than the air inside would have otherwise.  Let me examine these results below:

These tests represent 24 minutes of data averaged every two seconds. The RM-80 is stationary on the test substrate in all these tests. All of the tests measure the rate MicroRoentgens/HR  (μ/hr) which I also identify as MicroRads_HR; although they are not quite the same thing. The first two graphs are effectively control samples. The first is the open air in my office. The second is the RM-80 filter head placed down on  the shrink wrap plastic covering of a brand new filter. The last two tests are the dirty or used external filter which sits outside the fresh air intake under an external sheet metal hood.  In the first of these tests (Sample 2),  the used filter is covered in freezer bag; in the second test (Sample 3) the  RM-80 is placed upon the used filter uncovered. 

Even to the untrained eye, one can see that the used air filter is much thicker in the mid range measurements than the controls. One can get a feel for this by examining the statistics I derive from the non-zero data:

      Samples.V1 SampleEqZero.V1 MinNot0.V1 max.V1   mean.V1    sd.V1
Control        718             341       8.47  50.85  6.702563 7.972580
Sample1        718             297       8.47  50.85  7.481462 7.957133
Sample2        718             194       8.47  42.37 10.467354 9.091690
Sample3        718             220       8.47  42.37  9.924708 9.029365

Because I have used code to remove zero readings in the samples above, the statistics are not evenly weighted. Here is another look from the R summary function:


 s1.MicroRads_HR  s2.MicroRads_HR  s3.MicroRads_HR s4.MicroRads_HR 
 Min.   : 0.000   Min.   : 0.000   Min.   : 0.00   Min.   : 0.000  
 1st Qu.: 0.000   1st Qu.: 0.000   1st Qu.: 0.00   1st Qu.: 0.000  
 Median : 8.470   Median : 8.470   Median : 8.47   Median : 8.470  
 Mean   : 6.703   Mean   : 7.481   Mean   :10.47   Mean   : 9.925  
 3rd Qu.: 8.470   3rd Qu.: 8.470   3rd Qu.:16.95   3rd Qu.:16.950  
 Max.   :50.850   Max.   :50.850   Max.   :42.37   Max.   :42.370 


The single 50.85 reading in both of the control group (Control and Sample1) disturbs the scale of visual comparison. However, Samples 2 and 3 (the air filter data) have less 0 measurements to begin with and their means are much higher:



The colored graph below is more complicated at first glance but designed to give the viewer some idea of how the levels compare between tests in number.  The RM-80 measures radiation in levels so that the non-zero readings look like this when counted:


[1] "Sample3"
  MicroRads_HR freq
1         8.47  258
2        16.95  167
3        25.42   46
4        33.90   24
5        42.37    3


By overlapping these groups as sorted frequency bins displayed by partially transparent color, we get a visual impression of the difference in profile between any two samples. In the comparison between the Control and Samples 2 and 3 (below far right), we see that the numbers for the control sample are only stronger for the zero level.



As a final test, I do row based matrix multiplication of all non zero samples  and their frequencies and then add those totals.  The matrix sums from the used filter samples (2 and 3) result in  much greater weights:

[1] "(Matrix) Sums of row based multiplication for non zero samples."
        MR.HR.Mtrx.Mult
Control         4812.44
Sample1         5371.69
Sample2         7515.56
Sample3         7125.94

This round of testing is enough to prove to myself that my filters are trapping some part of the  radioactive isotopes that would be otherwise not be found in my home. The composition of  those isotopes remains undefined for now.

[1] "Four samples:"
[1] "Control is Room Air."
[1] "Sample1 is a high (MERV rating) rated blank filter with metal grate still in sealed plastic."
[1] "Sample2 Tests high (MERV rating) air filter located outside 8 inch external air intake for  approximately one month."
[1] "Sample2 RM80 stationary against center of filter while covered in plastic sack."
[1] "Sample3 Tests high (MERV rating) air filter located outside 8 inch external air intake for  approximately one month."
[1] "Sample3 : RM80 stationary against center of filter without barrier."
[1] "Time Base Unit is 2 seconds. MicroRads_HR = microroentgens/hr"
[1] "Sample Statistics:"
        Samples.V1 SampleEqZero.V1 MinNot0.V1 max.V1   mean.V1    sd.V1
Control        718             341       8.47  50.85  6.702563 7.972580
Sample1        718             297       8.47  50.85  7.481462 7.957133
Sample2        718             194       8.47  42.37 10.467354 9.091690
Sample3        718             220       8.47  42.37  9.924708 9.029365

[1] "Matrix Information for Non Zero Data"
[1] "Control"
  MicroRads_HR freq
1         8.47  237
2        16.95  100
3        25.42   33
4        33.90    4
5        42.37    2
6        50.85    1
[1] "Sample1"
  MicroRads_HR freq
1         8.47  259
2        16.95  122
3        25.42   32
4        33.90    6
5        42.37    1
6        50.85    1
[1] "Sample2"
  MicroRads_HR freq
1         8.47  282
2        16.95  150
3        25.42   67
4        33.90   21
5        42.37    4
[1] "Sample3"
  MicroRads_HR freq
1         8.47  258
2        16.95  167
3        25.42   46
4        33.90   24
5        42.37    3

[1] "Sending Four time-based Graphs of MicroRads_HR Averages Non Zero Data"
[1] "Sending Four frequency-based Graphs of MicroRads_HR Averages for Non Zero Data"

[1] "(Matrix) Sums of row based multiplication for non zero samples."
        MR.HR.Mtrx.Mult
Control         4812.44
Sample1         5371.69
Sample2         7515.56
Sample3         7125.94

No comments: