Abstract
Precise prediction of power consumption is critical to power generation planning, which is important to overall power demand and supply management. Power consumption for university campuses is different from that of general office buildings and shopping malls due to, among other factors, the particular academic calendar that includes summer and winter vacations. GAM (Generalized Additive Model) model [1] is a powerful tool to fit a set of response data to a collection of explanatory variables for its exceptional capability of modeling interacting explanatory variables though nonlinear smooth functions. GAM has been successfully employed in power prediction for modeling targets like research covering office buildings and nation-wide consumption. However, the most difficult problem in applying GAM modeling is how to find the right fitting GAM formula in the first place. In this regard, this paper proposes a general form of GAM demand model based on human daily activity. On top of this general form, we propose an AOO (Adding One by One) algorithm to find a fitting GAM model for a particular target based on historic data collected from that particular target. In particular, in the context of power consumption modeling, we train the general GAM model by a year-long historic power consumption data of year 2014 from a university campus to derive prediction models for both the expected value and its variance. The two resultant fitting GAM formulas thus obtained for both expected value and the associated variance have been applied to predict the power consumption for the following three years of 2015–2017. The prediction results have shown preferable performance, MAPE for expected value and CAE for variance, over that predicted by the GAM models reported in the literature [2].
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.