Kaustav Basu

Nonintrusive Load Monitoring: A Temporal Multilabel Classification Approach

The article tackles the issues related to the identification of electrical appliances inside residential buildings. Each appliance can be identified from the aggregate power readings at the meter panel. The possibility of applying a temporal multilabel classification approach in the domain of nonintrusive load monitoring is explored (nonevent-based method). A novel set of metafeatures is proposed. The method is tested on sampling rates based on the capabilities of current smart meters. The proposed approach is validated over a dataset of energy readings at residences for a period of a year for 100 houses containing different sets of appliances (water heater, washing machines, etc.). This method is applicable for the demand side management of households in the current limitation of smart meters; from the inhabitants or from the grid operators point of view.

Appliance usage prediction using a time series based classification approach

Energy management for residential homes and offices require the prediction of the usage(s) or service request(s) of different appliances present in the house. The hardware requirement is more simplified and practical if the task is only based on energy consumption data and no other sensors are used. The proposed model tries to formalize such an approach using a time-series based multi-label classifier which takes into account correlation between different appliances among other factors. In this work, prediction results are shown for 1-hour in the future but this approach can be extended to predict more hours in the future as per the requirement(with restrictions). The learned models and decision tree showing the important factors in the input information is also discussed.

Residential appliance identification and future usage prediction from smart meter

Energy management for residential homes and/or offices requires both identification and prediction of the future usages or service requests of different appliances present in the buildings. The aim of this work is to identify residential appliances from aggregate reading at the smart meter and to predict their states in order to minimize their energy consumption. For this purpose, our work is divided in two distinct modules: Appliance identification and future usage prediction. Both identification and prediction are based on multi-label learners which takes inter-appliance co-relation into account. The first part of the paper concerns the identification of electrical appliance usages from the smart meter monitoring. The main objective is to be able to identify individual loads from the aggregate power consumption in a non-intrusive manner. In this work, high energy consuming appliances are identified at 1-hour sampling rate using novel set of meta-features for this domain. The second part of the paper concerns future usage prediction. A comparison of algorithms for future appliance usage prediction using identification and direct consumption reading is presented. This work is based on a real residential dataset, called IRISE: 100 houses monitored every 10 minutes to one hour during one year (including weather informations).

A comparative study of low sampling non intrusive load dis-aggregation

Non-intrusive load monitoring (NILM) deals with the identification and subsequent energy estimation of the individual appliances from the smart meter data. The state of the art applications typically runs once per day and reports the detected appliances. In this work, data driven models are implemented for two different sampling rates (10 seconds and 15 minutes). The models are trained for 20 houses in the Netherlands and tested for a period of 4-weeks. The results indicate that the disaggregation methods is applicable for both sampling cases but with different use-case.

Online forecasting of electrical load for distributed management of plug-in electric vehicles

The paper aims at making online forecast of electrical load at the MV-LV transformer level. Optimal management of the Plug-in Electric Vehicles (PEV) charging requires the forecast of the electrical load for future hours. The forecasting module needs to be online (i.e update and make forecast for the future hours, every hour). The inputs to the predictor are historical electrical and weather data. Various data driven machine learning algorithms are compared to derive the most suitable model. The results indicate that an online forecasting method has an error between 2-5% for the future 24-hour. The decentralized management system works well with the forecasting data.

On the most convenient mixed strategies in a mixed strategist dynamics approach for load management of electric vehicle fleets

This manuscript explores the selection of appropriate mixed strategies (MSs) in a Mixed Strategist Dynamics (MSD) application for load management of Plug-in Electric Vehicle (PEV) fleets. This selection is based on the convenience of PEV owners, aiming to choose those MSs that privilege early high (or fast) charging rates when it is possible. The previously published MSD and Maximum Entropy principle (MSD-MEP) approach is revised and illustrated with several examples, specially in the context of selection of MSs. This revision allows a wider understanding of the method, and aims to inspire new contributions on domains where distributed optimization methods are pertinent. Results obtained without any management structure are compared to those obtained with the MSD-MEP approach under different scenarios, where full sets of MSs and reduced sets of convenient MSs are applied. The performance of the method, using conveniently reduced sets of MSs, is tested with real historical active power measurements, provided by the SOREA utility grid company in the region of Savoie, France.