Abstract
Buildings play a central role in energy transition, as they were responsible for 67.8% of the total consumption of electricity in France in 2017. Because of that, detecting anomalies (outliers) is crucial in order to identify both potential opportunities to reduce energy consumption and malfunctioning of the metering system. This work aims to compare the performance of several outlier detection methods, such as classical statistical methods (as boxplots) applied to the actual measurements and to the difference between the measurements and their predictions, in the task of detecting outliers in the power consumption data of a tertiary building located in France. The results show that the combination of a regression method, such as random forest, and the adjusted boxplot outlier detection method have promising potential in detecting this type of data quality problem in electricity consumption.
Highlights
Of the total consumption of electricity in France in 2017
Several methods for the detection of outliers have been used in recent times. Classic statistical methods, such as the three-sigma rule [6] and the boxplot method [7], have been highly used. These techniques assume a symmetrical data distribution and the performance of these techniques is highly dependent on this feature, which is commonly unknown for power consumption data
In order to find a model for the GreEn-ER energy consumption, the random forest gressor only in test time interval because of their effect in the statistical v method was applied using the data exposed in the previous section as the regression median, used to data detect these abnormal samples
Summary
Of the total consumption of electricity in France in 2017. Because of that, detecting anomalies (outliers) is crucial in order to identify both potential opportunities to reduce energy consumption and malfunctioning of the metering system. This work aims to compare the performance of several outlier detection methods, such as classical statistical methods (as boxplots) applied to the actual measurements and to the difference between the measurements and their predictions, in the task of detecting outliers in the power consumption data of a tertiary building located in France. The results show that the combination of a regression method, such as random forest, and the adjusted boxplot outlier detection method have promising potential in detecting this type of data quality problem in electricity consumption. In France, according to Bilan RTE 2018 [1], approximately 67.8% of electricity is consumed in buildings, both residential and tertiary. These figures indicate that buildings play a central role in energy transition. In [10] the author used linear regression to detect outliers and in [11] an auto regressive moving average (ARMA) was used as the regression technique
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: 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.
