Zainul Charbiwala

Multi-User Energy Consumption Monitoring and Anomaly Detection with Partial Context Information

Anomaly detection is an important problem in building energy management in order to identify energy theft and inefficiencies. However, it is hard to differentiate actual anomalies from the genuine changes in energy consumption due to seasonal variations and changes in personal settings such as holidays. One of the important drawbacks of existing anomaly detection algorithms is that various unknown context variables, such as seasonal variations, can affect the energy consumption of users in ways that appear anomalous to existing time series based anomaly detection algorithms. In this paper, we present a system for monitoring the energy consumption of multiple users within a neighborhood and a novel algorithm for detecting anomalies by combining data from multiple users. For each user, the neighborhood is defined as the set of all other users that have similar characteristics (function, location or demography), and are therefore likely to react and consume energy in the similar way in response to the external conditions. The neighborhood can be predefined based on prior customer information, or can be identified through an analysis of historical energy consumption. The proposed algorithm works as a two-step process. In the first step, the algorithm periodically computes an anomaly score for each user by just considering their own energy consumption and variations in the consumption of the past. In the second step, the anomaly score for a user is adjusted by analyzing the energy consumption data in the neighborhood. The collation of data within the neighborhood allows the proposed algorithm to differentiate between these genuine effects and real anomalous behavior of users. Unlike multivariate time series anomaly detection algorithms, the proposed algorithm can identify specific users that are exhibiting anomalous behavior. The capabilities of the algorithm are demonstrated using several year-long real-world data sets, for commercial as well as residential consumers.

UrJar: A Lighting Solution using Discarded Laptop Batteries

Forty percent of the worlds population, including a significant portion of the rural and urban poor sections of the population in India, does not have access to reliable electricity supply. Concurrently, there is rapid penetration of battery-operated portable computing devices such as laptops, both in the developing and developed world. This generates a significant amount of electronic waste (e-waste), especially in the form of discarded Lithium Ion batteries which power such devices. In this paper, we describe UrJar, a device which uses re-usable Lithium Ion cells from discarded laptop battery packs to power low energy DC devices. To understand the usability of UrJar in a real world scenario, we deployed it at five street-side shops in India, which did not have access to grid electricity. The participants appreciated the long duration of backup power provided by the device to meet their lighting requirements. To conclude, we present an ecosystem which consists of a community-level energy shed and UrJar devices individually owned by households, as a mechanism for DC electrification of rural areas in developing countries. We show that UrJar has the potential to channel e-waste towards the alleviation of energy poverty, thus simultaneously providing a sustainable solution for both problems.

DC picogrids: a case for local energy storage for uninterrupted power to DC appliances

An increasing number of appliances now operate on DC and providing uninterrupted power supply (UPS) to them through outages requires two conversions: first from an energy store, typically a DC battery, to AC mains and then from AC mains to the DC input required by the appliance. The energy storage and DC-to-AC inversion are usually centrally located and tied to existing AC distribution lines to amortize costs and battery capacity. In this paper, we argue that adding energy storage locally to each DC appliance and managing it intelligently can lead to higher efficiency and lower average cost. We term this topology a DC picogrid as it mimics a scaled down independent microgrid. Our contribution is the design and evaluation of a smart picogrid controller that a) identifies the power source and b) decides on battery charging or discharging based on the power source. As we expect DC picogrids to co-exist with AC UPSes, we must ensure that the DC picogrid does not draw power from the UPSs battery but charges from the macrogrid when available. To accomplish this, we exploit the fact that AC distribution from the macrogrid exhibits sufficiently distinct characteristics compared to an AC UPS or a diesel generator. Our picogrid controller uses a Hidden Markov Model for state estimation that uses temporally correlated fluctuations in line voltage and frequency for discrimination. We show through data from four settings that the controller can identify its supply source with over 90% accuracy, and that efficiency recovered from conversion losses could result in 30% reduction in energy consumption.

DC Picogrids as power backups for office buildings

Office buildings in developing countries employ battery backups with inverters and/or diesel generators to power essential loads such as lighting, air conditioning and computing loads during power cuts. Since these backup solutions are expensive and inefficient, they form a significant proportion of the operating expenses. To address this problem, we propose using a personal comfort system (an illustrative configuration can comprise a LED light and a DC desk fan) that is powered by batteries in computing devices. With this approach, cost savings are realized through two mechanisms, (i) by reducing the dependence on high-power lighting and air conditioning during times of power outage, and (ii) by charging the batteries at optimal times, taking advantage of the variable cost of power supply. Simulations show that the expected energy savings from this methodology are in the region of 26%, compared with the current system. In this paper, we present various architectures for the load-battery combination, a dynamic programming based framework that generates optimal charging/discharging schedules, and an experimental evaluation of the proposed approach.

Collaborative energy conservation in a microgrid

KBFSC (Kuala Belalong Field Studies Centre) is a research centre located in a remote tropical evergreen rainforest in Brunei Darussalam in South East Asia. It is visited by biologists and ecologists from all over the world. Power is available at the centre for 8-10 hours per day from a diesel generator (DG). The diesel travels2-3 hours by road, by boat and on foot over harsh terrain to reach the centre from the closest gas station. This paper describes the software and hardware of a microgrid system that was designed and deployed at KBFSC to reduce the fuel consumption while improving duration of power availability. A key feature of the energy management software is a collaborative scheduler interface that provides visitors at the centre the choice of scheduling appliance usage. The system optimises generator active hours using a customised DG Optimiser technique, to ensure minimum diesel consumption. Simulations extrapolating from empirical data suggest that our system could reduce diesel consumption by a third, and total cost by 20%, while making power available 24 hours a day. In addition, a user study with 8 visitors and 4 administrators showed that the collaborative scheduler interface is effective and usable.

A Framework for Evaluating the Costs and Benefits of Instrumentation in Smart Home Systems

The goal of this paper is to establish a framework for evaluating the marginal utility of adding smart sensing and metering hardware to residential premises, in terms of efficiency improvement (reduction in energy consumption) and economic benefits (reduction in energy cost). A simulation procedure is developed for experimenting with different types of hardware architectures. In-house analysis algorithms suitable to the installed suite of sensing and metering hardware are applied, which take into account realistic practical constraints. The proposed methodology allows us to perform a cost-benefit analysis of several potential smart home solutions. This analysis is meant to enable home owners to evaluate a priori the real cost saving potential of these solutions, when applied to their home.

Algorithms for upgrading the resolution of aggregate energy meter data

Metering of the energy supplied to consumers is an important component of operations for utility providers. Several schemes have been employed for this purpose, including traditional postpaid and prepaid metering, and more advanced smart metering technology. Analysis of the data generated by these meters has the potential to provide insights into consumer characteristics and power consumption patterns, including consumer segmentation and anomaly detection. We describe the different types of power purchase and consumption data, as well as the analytics algorithms that can be applied to them. Most applications developed for energy meter data require high resolution information of the type provided by smart meters, thus leaving aggregate prepaid or postpaid meter schemes at a disadvantage. In this paper, we present analytics-based methodologies to upgrade aggregate prepaid and postpaid meter data resolution, which will allow smart meter analytics to be applied without expensive infrastructure upgrades.

SMOME: A framework for evaluating the costs and benefits of instrumentation in smart home systems

The goal of this paper is to establish a framework called SMOME for evaluating the marginal utility of adding smart sensing and metering hardware to residential premises, in terms of efficiency improvement (reduction in energy consumption) and economic benefits (reduction in energy cost). In order to isolate the effect of smart home technologies on energy efficiency, it is assumed that no changes are made to the appliances already installed. SMOME is developed for experimenting with different types of hardware architectures. Analysis algorithms suitable to the installed suite of sensing and metering hardware are described, which take into account realistic practical constraints. Validated with empirical appliance-level data, the proposed methodology allows us to perform a cost-benefit analysis of several potential smart home solutions. This analysis is meant to enable home owners to evaluate a priori the real cost saving potential of these solutions, when applied to their home.

Unlocking the hidden potential of data towards efficient buildings: Findings from a pilot study in India

Energy cost is one of the significant contributors to the operational expenses of commercial buildings. In developing countries facing problems of frequent power outages and deficient grid connectivity, diesel generators are used as backup power source which significantly increase the costs incurred in management of commercial establishments. Integration of information and communication technologies to building management systems provides a reliable platform to analyze various aspects of the building such as energy consumption trends and occupancy inferences thereby proposing reactive or pro-active strategies directed towards efficient and cost-effective building management. Usually, this potential of data available to building management agencies stays untapped in developing countries. In this paper, we take a data-driven approach to understand various operational aspects of a commercial establishment. To demonstrate the scope for optimization of building operations by exploiting the energy consumption data, a pilot study was conducted in an IT office building in India.

UrJar: A Device to Address Energy Poverty Using E-Waste

A significant portion of the population in India does not have access to reliable electricity. At the same time, is a rapid penetration of Lithium Ion battery-operated devices such as laptops, both in the developing and developed world. This generates a significant amount of electronic waste (e-waste), especially in the form of discarded Lithium Ion batteries. In this work, we present UrJar, a device which uses re-usable Lithium Ion cells from discarded laptop battery packs to power low energy DC devices. We describe the construction of the device followed by findings from field deployment studies in India. The participants appreciated the long duration of backup power provided by the device to meet their lighting requirements. Through our work, we show that UrJar has the potential to channel e-waste towards the alleviation of energy poverty, thus simultaneously providing a sustainable solution for both problems. Mode details of this work are provide in [3].