Re: Re[2]: Electronic Dosimetry

[Bob Flood <bflood@SLAC.Stanford.EDU>]

At 04:56 PM 3/7/97 -0600, you wrote:
>     planning on using the TLD as a back-up (I also spoke with ANI when 
>     they were here for an inspection).  Some benefit can be derived from 
>     this.  One could set up a program where the TLD is not processed 
>     unless it is necessary (e.g., ED failure, or ED dose >= to 100 mrem or 
>     some other dose setpoint), or the TLD processing frequency could be 
>     extended to say once per 6 months.

This could be a reasonable measurement system, although the idea of using
an accredited system as a backup to an unaccredited one may have some
political or image problems. I can state based on experience that the
Panasonic UD-802 can be used for a monitoring period as long as a year,
since fade dwindles to almost nothing after 5-6 months. Using a criterion
based solely on ED data (malfunction or dose total) would be very
managable. I STRONGLY RECOMMEND THAT YOU READ THE TLD FOR DOSE AT YEAR'S
END, EVEN IF YOU DON'T RUN ANY CALCULATIONS WITH THE DATA. IF YOUR ED
SYSTEM EVER GETS CHALLENGED, THE READINGS COULD BE RETRIEVED AND DOSES
CALCULATED AT THAT TIME TO BACKUP/VERIFY THE ED DATA.

>     As far as investigation of a failed ED, with the back-up TLD, you have 
>     an alternative, also with the Siemens EPD and associated software, we 
>     have the ability to identify (with a couple of computer keystrokes) 
>     every user of the particular EPD in question since it was issued from 
>     Dosimetry.  Not really a problem!

Could do this at TVA, too. That's the easy part. Suppose the user list is
1,000 people since its last good QC check. They all worked varying jobs in
varying locations, working with a variety of different coworkers or no
coworkers. Assembling the data on every case, looking over the coworker
data and choosing what applies best to the user of the failed ED is not a
trivial task. Multiplied by 1,000 users, it's huge!

>     Consider this:  TLDs are issued to 
>     different workers also.  If the worker's dose is typically low each 
>     processing period (20-30 or anything <100 mrem for that matter), an 
>     under-responding TLD may be hard to detect.  At the end of the 1 or 
>     two year period when it comes time to generate ECFs, and a particular 
>     TLD(s) ECF changes by, oh say 25%, what do you do then?

That's a routine situation in any TLD program. But in that situation, the
dosimeter has been by a maximum of 4 people. Not the same problem.

>     Data Loss.  With the Siemens EPD (I hope with the others also) the 
>     data is written to an EEPROM every 15 minutes.  If the EEPROM were 
>     totally obliterated, then the dose would be lost.  Anything short of 
>     obliteration, dose can be retrieved from the EEPROM.
>

Agreed, as long as access to the EPROM is in tact or can be restored.
Dropped EDs generally don't lose data, unless it's dropped in the fuel pool
or Rx vessel. When you do something 1,000's of time a month (use EDs),
those 1 in 100,000 rare events become inevitable! I don't expect anyone to
see this problem frequently, but I think every program must be prepared to
deal with it.
     
>     Calibration.  With two irradiators, EPDs can be calibration verified 
>     at a rate of 1 per minute with one person.

As you have no doubt learned from users of other brands, not all EDs lend
themselves to such productivity. Users should consider the calibration and
QC workload in their purchase decision.

>     If you 
>     want to pursue Electronic Dosimeter Dose of Record, you need to have 
>     at least the following (in my opinion):
>     
>     An ED that performs like a TLD.  
>     A back-up Dosimeter.
>     NVLAP Accreditation, including proficiency testing in all categories 
>     except 3A and Neutrons.
>     A sound Calibration or Calibration Verification program.
>     A well documented basis to support ED as dose of record.
>     Regulator (NRC etc.) concurrence.
>     Plant Management concurrence.
>     

That's a good list of fundamentals for any dosimetry program.

>     A major problem that we have in this industry is that there is no 
>     REALISTIC Standard for us to measure against as far as EPD 
>     performance.

And with a ponderously slow standard development process, that problem
won't go away soon.

There's nothing of a technical nature that prohibits one from doing good
dosimetry using EDs. All of the technical issues can be resolved, but it's
still my fear that the workload issue of a failed ED and the associated
dose investigations is a bone crusher.

Opinion. Nothing but opinion. 


Bob Flood
Stanford Linear Accelerator Center
(415) 926-3793     bflood@slac.stanford.edu
Unless otherwise noted, all opinions are mine alone.