|Publication type:||Conference other|
|Type of review:||Editorial review|
|Title:||How to collect current usage data outside the laboratory|
|Conference details:||7th RME Research Conference, Virtual Conference, 19-20 October 2020|
|Subjects:||Energiemessung; NILM; Stromdatenanalyse|
|Abstract:||Everybody is talking about saving energy. Since it’s always a good idea to reduce electrical power consumption as much as possible, it is not so easy to find the best way to do so. Without knowing in detail, how current is used, it is only a rough guess which electrical device could be optimized. But not only the device is responsible for power usage also the user operating the device has a big impact. Although it is easy to collect (electrical) data in a laboratory, just use a power-meter, getting information out from and insight in a “real” environment is more complicated. The goal is to get a meter between the power grid and an existing household – right after the official electrical meter. It is always easy to develop a customized solution, but usually, it is the slowest and most expensive solution. The usage of standardized and already available equipment solves many problems (e.g. certification, robustness) and is much cheaper. On the downside, you must work with what you get – almost no tweaks are possible. Through my work, I was able to collect many experiences in the collection of huge amounts of raw electrical data mostly with a self-developed smart-meter. In short; a raspberry-Pi computer combined with a standard energy-consumption-meter. Collecting precise electrical data over periods from weeks to over a year with a resolution of one second from different swiss households and industrial plants resulted in a wide collection of raw data. Just to collect constantly the raw data is only the first step, a big part of the work is the cleaning and preparation of the database. For example, further development of the firmware during measurement changes the data output; It is mandatory to harmonize all the data for further analysis. After this step, it is possible to use the data for many purposes. Getting a detailed insight of the daily electrical usage is just the easiest one. How about detecting some patterns? Now it is possible to do research far beyond “just saving energy”. The collected data is predestined for data mining, either statistical or by advanced artificial intelligence algorithm. Different methods lead to different results - this is the actual state of work I’d like to present.|
|Fulltext version:||Published version|
|License (according to publishing contract):||Licence according to publishing contract|
|Departement:||School of Engineering|
|Organisational Unit:||Institute of Energy Systems and Fluid Engineering (IEFE)|
|Appears in Collections:||Publikationen School of Engineering|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.