How to use temperature and humidity environmental control data?

D

Dudes

Hello Covers!

We monitor temperature and humidity of our rooms and fridges using electronic data recorders. They are read out once a year, results are printed out and stored in a folder. No analysis of the results is done and we don't even have an allowed data range (very useful, I know...:notme:)

I'd like to get some idea of how I could better use those data. Of course, I'll give T and Humidity tolerances and deviations will have to be justified (though it might be difficult with a yearly reading). Would you suggest any kind of statistical analysis of the data that could give us valuable infomation?

:thanx:
 
D

Darius

IMHO, there are some basic analysis methods:

Histogram of T and H (all data set)
IMR Chart (all data set)
Box Plot per month

Histogram, could give you the range of variation for each Variable

IMR Chart, could give you the limits where your process behabe and permit to see trend or patterns that can lead you to the Knoledge of your process

Box Plot per month, is usefull to compare months, looking for big changes, even some places recommend them insteed of Capability indexes.

And of course, try to correlate the two variables (correlation analysis), if there is a strong relation between both, maybe could not be necesary to track both.
 

Ajit Basrur

Leader
Admin
Hello Covers!

We monitor temperature and humidity of our rooms and fridges using electronic data recorders. They are read out once a year, results are printed out and stored in a folder. No analysis of the results is done and we don't even have an allowed data range (very useful, I know...:notme:)

I'd like to get some idea of how I could better use those data. Of course, I'll give T and Humidity tolerances and deviations will have to be justified (though it might be difficult with a yearly reading). Would you suggest any kind of statistical analysis of the data that could give us valuable infomation?

:thanx:

I would like to ask couple of questions to get more meaningful answers for you -

  1. What products are stored ?
  2. Are those products temperature and humidity sensitive ?

Secondly, if you are generating electronic data, I would suggest frequent monitoring of the data and look for temperature / humidity spikes (we do daily in Pharmaceuticals). If at all there are any spikes, you could take remedial actions at the earliest.

If you start doing this routine monitoring, statistical analysis could follow.

Hope this answers your question :)
 
D

Dudes

I would like to ask couple of questions to get more meaningful answers for you -

  1. What products are stored ?
  2. Are those products temperature and humidity sensitive ?

Secondly, if you are generating electronic data, I would suggest frequent monitoring of the data and look for temperature / humidity spikes (we do daily in Pharmaceuticals). If at all there are any spikes, you could take remedial actions at the earliest.

If you start doing this routine monitoring, statistical analysis could follow.

Hope this answers your question :)

The products (polymers) stored in fridges are temperature sensitive (they wouldn't be stored in fridges otherwise ;-)). Other products are not sensitive. Daily reading is unfortunately not an option as the loggers are dispersed all over the place (as well as off-site... I know it would be the best as it would allow to discover deviations almost as they happen (and not months later :bonk:)
 
D

Dudes

IMHO, there are some basic analysis methods:

Histogram of T and H (all data set)
IMR Chart (all data set)
Box Plot per month

Histogram, could give you the range of variation for each Variable

IMR Chart, could give you the limits where your process behabe and permit to see trend or patterns that can lead you to the Knoledge of your process

Box Plot per month, is usefull to compare months, looking for big changes, even some places recommend them insteed of Capability indexes.

And of course, try to correlate the two variables (correlation analysis), if there is a strong relation between both, maybe could not be necesary to track both.

Thanks for those suggestions! There is a link between both variables. Relative humidity will vary with temperature. However, a peak in humidity doesn't necessarily mean a peak in T. It can have other causes. We therefore need to monitor both... (I don't know if I was very clear :eek:)

Comparing month to month wouldn't bring much (at least for the rooms) as there are seasonal variations.

BTW, we don't monitor a process per se, but the temprature of the rooms (clean room, storage, ...).
 

somashekar

Leader
Admin
Hello Covers!

We monitor temperature and humidity of our rooms and fridges using electronic data recorders. They are read out once a year, results are printed out and stored in a folder. No analysis of the results is done and we don't even have an allowed data range (very useful, I know...:notme:)

I'd like to get some idea of how I could better use those data. Of course, I'll give T and Humidity tolerances and deviations will have to be justified (though it might be difficult with a yearly reading). Would you suggest any kind of statistical analysis of the data that could give us valuable infomation?

:thanx:
I gather that you must be having some limits for the temperature and humidity required in your process and what you have generated is a record of the periodic monitoring. While the record is the evidence of your monitoring, any out of limit record is a trigger point for your immediate action as Ajit said here in post #3 based on the daily observation of the data.
Unless otherwise you intend to make some changes in the temperature and humidity control system so as to narrow the limits further, I do not see a reason why you want some statistical analysis of this data (or record)
 
D

Darius

Thanks for those suggestions! There is a link between both variables. Relative humidity will vary with temperature. However, a peak in humidity doesn't necessarily mean a peak in T. It can have other causes. We therefore need to monitor both...

So you can use this relationship to detect some other patterns and of course, if this could be on line, desto improve your process with a lower variation.
 
S

samsung

We monitor temperature and humidity of our rooms and fridges using electronic data recorders. They are read out once a year, results are printed out and stored in a folder. No analysis of the results is done and we don't even have an allowed data range (very useful, I know...:notme:):thanx:

If you don't have an 'allowed' (permissible) data range, why don't you yourself define an 'acceptable' data range against which you can monitor and compare the actual measurements. Ofcourse, your initial assumptions would be based on the past experience that you gained from day to day observations which should enable you to determine at what point you normally get the 'acceptable' results.
 

ScottK

Not out of the crisis
Leader
Super Moderator
I have used temp and humidity measurement to analyze racking locations for food and cosmetic product storage in an air conditioned warehouse.
I was able to identify "warm spots" in the summer and mark them for storage of less succeptible product.
I suppose this could have been extended to have the air flow in the building modified to be more uniform, but we didn't feel the need to take it that far.

I did a pretty simple analysis with run charts and histograms.

but you have to look at it more than once a year. I did a week to week analysis.
*edit to add* by that I mean I collected the data on a weekly bases. The data loggers I used logged the measurements every 10 minutes, I think.

I also compared the measurements to outside temperature but didn't see much of a correlation and took it no further.
 
Last edited:

Ajit Basrur

Leader
Admin
The products (polymers) stored in fridges are temperature sensitive (they wouldn't be stored in fridges otherwise ;-)). Other products are not sensitive. Daily reading is unfortunately not an option as the loggers are dispersed all over the place (as well as off-site... I know it would be the best as it would allow to discover deviations almost as they happen (and not months later :bonk:)

Did you consider installing alarms (both audible and visual) to alert you of excursions ? This would alert you to take corrective actions immediately !

Technology, these days is so advanced that it could also send you short messages to your cellphone :)
 
Top Bottom