A new home energy dataset can be used to create a list of energy consumption activities is Smart* dataset. The information can be found at this link: http://traces.cs.umass.edu/index.php/Smart/Smart.
They have two house: B, and C that measured the aggregate electrical data, and environmental data; and House A that measured the aggregated energy consumption as well as electrical data from individual energy meters. As we are only interested in the individual appliances, so House A could be a dataset.
In house A, basically they recorded all sub-meters in 3 months, started from 01 May 2012 to 31 July 2012.
There are 91 days. Each day contains a list of labels, timestamp, and real power consumption at that timestamp. The resolution is about 3 seconds.
I attempted to get the data of every single event/label during these 3 months. The aim is to convert these dataset so that it would look like REDD dataset. Then, I can apply similar methods to get the list of ON-OFF EVENTS.
The first task is to separate the timestamp, and power usage of each label to single file. Having done that, it showed 53 labels altogether. To this far, the data should be looked at REDD dataset. Then, I converted these files to list of ON-OFF EVENTS.
I manually check the output. Surprisingly, majority of labels doesn't have data. For some reason, there are only few data on some irrelevant labels such as basement-dehumidifier, basement-freezer, bedroom-ac, kitchen-coffeepot, kitchen-toaster, livingroom-tv, master-fan. It is roughly 10/53 labels that working fine.
They don't have the data for major appliances like washing machine, microwave,...
This is a shame, because I have had a hope for this dataset.
Now I know it is unused for me. Such a waste of time :(.
No comments:
Post a Comment