Electric Vehicle Charging Sessions Profiling and Modelling
Perform mclust::Mclust clustering for multivariate GMM
Extract models parameters from mclust object
Visualize BIC indicator to choose the number of clusters
Cluster sessions with mclust package
Modify datetime values according evprof.start.hour
Convert datetime values to sorted numeric values considering a start t...
Convert numeric time value (hour-based) to character hour in %H:%M for...
Convert numeric time value to a datetime period (hour-based)
Cut outliers based on minimum and maximum limits of ConnectionHours an...
Define each cluster with a user profile interpretation
Detect outliers
Divide sessions by disconnection day
Divide sessions by time-cycle
Drop outliers
Get charging rates distribution in percentages
Get a tibble of connection GMM for every user profile
Get the daily average number of sessions given a range of years, month...
Get daily number of sessions given a range of years, months and weekda...
Get the minPts and eps values for DBSCAN to label only a specific perc...
ggplot2 type function to plot a division line
Get Mclust object of univariate Gaussian Mixture Models
Get energy univariate Gaussian Mixture Model
Get a tibble of energy GMM for every user profile
Get the EV model object of class evmodel
Logarithmic transformation to ConnectionStartDateTime and ConnectionHo...
Plot Bivariate Gaussian Mixture Models
Density plot in 2D, considering Start time and Connection duration as ...
Density plot in 3D, considering Start time and Connection duration as ...
Iteration over evprof::plot_division_line function to plot multiple li...
Compare density of estimated energy with density of real energy vector
Histogram of a variable from sessions data set
Grid of multiple variable histograms
Plot kNNdist
Plot all bi-variable GMM (clusters) with the colors corresponding to t...
Plot outlying sessions
Scatter plot of sessions
print method for evmodel object class
Read an EV model JSON file and convert it to object of class evmodel
Round numeric time value to half hour basis.
Round to nearest interval
Save iteration plots in PDF file
Save the EV model object of class evmodel to a JSON file
Classify sessions into user profiles
Statistic summary of sessions features
Tools for modelling electric vehicle charging sessions into generic groups with similar connection patterns called "user profiles", using Gaussian Mixture Models clustering. The clustering and profiling methodology is described in Cañigueral and Meléndez (2021, ISBN:0142-0615) <doi:10.1016/j.ijepes.2021.107195>.
Useful links