Alternative IceCOLD Testing Strategies - The Case for Enhanced Data

     As an investigator looking into the usefulness of IceCOLD for increasing performance and reducing costs of refrigeration and air conditioning, we have learned the hard way to wary of unplanned variables.  Twice, we have even "tested the testing" trying out different logging strategies to find consistent pre-test data that is likely to remain consistent throughout the entire testing period.

    Sometimes, variables creep in that were entirely unplanned.  At the Golden Corral test, we understood that human activity (going in and out of the coolers) would be a difficult variable to track.  Therefore, total power use could vary independently of tracked variables like outdoor temperature.  We hoped that tracking coil temperatures would allow us to demonstrate the usefulness of the product regardless of how the cooler was used.  

    Thankfully, this did work.  We saw the type of benefits we expected even though the Thanksgiving week fell in the middle of the test.  Coil capacity increased remarkably as we hoped and cycle lengths were reduced as well as compressor starts.  

    Had we relied on the traditional cooling degree day methods, we would not have been able to account for the vastly different usage patterns we found in this establishment.  The customer deserves to see what results he is getting and know just how well that the product is working for him.  Enhanced data was the answer!

     Regardless of what applies in the rest of life, in data-logging there is no such thing as too much information.  Unless someone has done the exact same type of testing many times before, you can never have too much data.  As of right now, every investigation we have been involved in had surprises.  To provide the customer with the best value, we need the best data.

    The Nebraska City Pizza Hut test illustrates this well.  During the study, the data loggers were not "synchronized."  The internal timing of the loggers was incorrect and the data appeared to be useless.  This researcher noted that power and temperature data was "off" by about 36 hours.  Simply, the higher power consumption times occurred in the middle of the night and peak temperature times had lower power consumption.  We did not give up when this became obvious.

    We took some liberty with the data and shifted the times so higher temperatures corresponded with higher power consumption.  Indeed, the curves match rather well.  To get all the details of the Pizza Hut test we invite you to read that post.