Detecting Faults and Estimating Missing Values in Smart Meter Data

Title	Detecting Faults and Estimating Missing Values in Smart Meter Data
Summary	Finding outliers and missing energy consumptions, and replace them with estimated values
Keywords	data mining, smart meter data
TimeFrame
References	http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5524054 http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1425550
Prerequisites	Cooperating Intelligent Systems and Learning Systems courses
Author
Supervisor	Sławomir Nowaczyk, Anita Sant'Anna, Hassan Mashad Nemati
Level	Master
Status	Open

This project is about investigating available techniques and methods for: outlier analysis, missing data detection, and electricity consumption estimation on smart meters data. The idea is to use data mining and machine learning techniques to check for inconsistencies in the data and replace them based on information of customer’s electricity usage.

Data are collected from real Smart Grid distribution network and contain active power consumption (daily or hourly consumption).