RE: Instrument Response / Training Issues

[Gv1@AOL.COM]

The Radiation Protection Manager here at VC Summer Nuclear Station, wants

>>a training class to be given on the following HP survey instruments: 



A couple of quick thoughts:



REM Ball



I believe you'll find from NUREG docs that your average unscattered neutron 

energy is approximately 100 keV and the multi-scattered flux is approximately 

50 keV.  I don't have any references with me, but I seem to recall that the 

sources most often used to calibrate rem balls have neutron energies much 

higher.  In this case for each neutron detected, the pulse counter is 

providing a signficant dose impact, whereas the true dose/unit flux in the 

plant is not that much.  Without any good comparison studies, I believe this 

might be a primary contributor to the fact that a rem ball seems to read 

signficantly higher than some of the new neutron electronic dosimeters with 

their detector response factors set to the average neutron energy in the 

field.  There is also a significant variance in TLD response in neutron 

fields.  I believe one utility multiplies their rem ball reading by 0.3 to 

get close to the neutron electronic dosimeter reading.  This is an area where 

a comparison to the rem ball really hasn't been available and further 

industry data should prove to be interesting.  I'm no longer at the plant 

with the electronic neutron dosimetry and wish I had more to get to the 

bottom of the issue...



Teletector



The teletector is a pulse counter that assigns 662 keV of dose to each 

counting event whereas the ion chamber measures the dose response of all of 

the energies.  I don't have my teletector in N-16 info available, but I have 

seen data from U of Mass at Lowell with I believe was the MG Ram Ion ion 

chamber.  I believe the density thickness for this ion chamber was 

approximately 300-400 mg/cm^2 and the response was within approximately 10% 

of the expected value.  If this low level of density thickness yields a 

comparison this close, then I'd expect the RO-2 chamber thickness to be this 

much or greater and yield similarly good comparisons.  This essentially says, 

there's enough density thickness in the chamber walls to achieve a decent 

electonic equilibrium.



PCM 1B/2



I belive you'll find that the tables for false alarm are for a single 

detector, so you'd effectively multiply that value by the number of detectors 

to get the overall false alarm rate for the monitor.  You can also probably 

get calculators from the major vendors to play what-if scenarios to help you 

better understand the performance of the monitor.  Sometimes you need better 

tools to obtain the higher level of knowledge.  One thing to cover is 

detector response vs. beta energy.  Higher energy betas are easier to detect. 

 So, if you have failed fuel and your average energy goes from ~96 keV for 

Co-60 to 200-300 keV, your monitor set at 5000 dpm will alarm at lower levels 

because these energies are more easily detected.  Radon descriminating 

mechanisms that work perfect almost sound to good to be true...  Think 

carefully about the body geometry and distances from the detectors vs. alpha 

range in air and the potential for licensed alpha to be present.  The EPRI 

doc concerning alpha at commercial nuclear power facilities suggests that you 

only need alpha monitoring if your beta/alpha ratio is <50:1.  This is 

equivalent to 5000 dpm/100 cm^2 fixed + removable beta and 100 dpm/100 cm^2 

alpha.  Additionally this would be 1000 dpm/100 cm^2 removable and 20 dpm/100 

cm^2 alpha.  The last value I recall seeing from INPO was a sensitivity of 

300 dpm for a personnel monitor.  Performance data will show you that 20 dpm 

is essentially unobtainable.  Remember it takes a SAC-4 (30% eff ZnS) 

approximately a minute to achieve an MDA of <20 dpm.  With respect to 

performing plateau's for all of the detectors, once I got the monitor set up, 

I'd only perform a plateau if the efficiency for a detector during the cal 

was outside your admin variance against the mean.  You'd also perform one 

when you replaced a detector.  If you read the plateau you can put in a new 

detector and tune it to the performance of the others just from the plateau 

data.  The key is you'll have a "worst" detector that has the worst 

signal-to-noise or efficiency to background and this detector will 

essentially drive the count time for the whole monitor.  I have a methodology 

for finding this detector and tuning the others around this detector to get 

the lowest background with the highest average efficiency with a tight fit.  

A tight fit for the average is also important or you'll get false alarms on 

the one with an efficiency that is too high.



PM-7



The PM-7 is a very predictable instrument.  It can be made to work in high 

background levels (>13,000 cps) and it will appropriately take itself out of 

service if the background fluctuates too much.  The shield factors in the 

instrument might run 6% or so which means at the end of the count cycle, the 

counts are inflated by 6%.  If you aren't the same size as the person that 

generated those values, you're susceptible to excessive false alarm rates in 

higher background areas.  In general adjust the false alarm probability as 

high as you can and still obtain a count time <2 sec.  The longer the count 

time in high and changing background levels, the more the count will be 

affected.  Minimize the count time and accept higher RDA values to get good 

performance in high background areas.  Essentially everyone is dragging these 

monitors deeper into the plant into increasingly higher background levels and 

it forces us to learn more about operating in adverse environments.  Most 

plants pick their calibration source Co-60 or Cs-137 to be the one most close 

to the average photon energy.  Think about the photon emission rate of your 

plant mix vs. the source your using.  Additionally, more work needs to be 

done to understand the impact of performing a plateau to set a detector 

voltage on a fixed descriminator and variable detector gain system with 

different photon energies.  Where do the photons go with energies that are 

below the "knee"?  I believe work in this area will generate some interesting 

insights...  Remember David Cardine in the Kung Fu series when the master 

would offer an intriguing riddle which held much reward?...  These 

philosophies would similarly apply to tool monitors as well.



AMS-4



Very reliable instrument.  Field operations will match the performance 

formulas in the manual to a "T".  For example the time needed for the monitor 

to go into service relative to the background level and alarm setpoint will 

generate curve fits of plant data that match the manual.  The primary issue 

to remember is that the bkg/source detector responses vary in the z plane.  

This condition can cause an under-response or over response.  For example, 

the draining of a reactor cavity will create a difference in the flux 

gradient and cause the background subtraction factor to change.  Reset the 

value periodically for changes in plant conditions like this to help you 

"see" the effect.  The key it to understand the personality of the monitor.  

It is not likely that a large group of rad techs will ever interface with the 

monitor enough to be good at diagnosing issues in the field.  This is where 

simplified written guidance is key to monitoring success.  The AMS-4 could 

take up much more space, but it's time to go... 





As you can tell, I don't have much of an interest in instruments...



Glen



************************************************************************

You are currently subscribed to the Radsafe mailing list. To unsubscribe,

send an e-mail to Majordomo@list.vanderbilt.edu  Put the text "unsubscribe

radsafe" (no quote marks) in the body of the e-mail, with no subject line.

You can view the Radsafe archives at http://www.vanderbilt.edu/radsafe/