Naveed Arshad

An Empirical Investigation of V-I Trajectory Based Load Signatures for Non-Intrusive Load Monitoring

Choice of load signature or feature space is one of the most fundamental design choices for non-intrusive load monitoring or energy disaggregation problem. Electrical power quantities, harmonic load characteristics, canonical transient and steady-state waveforms are some of the typical choices of load signature or load signature basis for current research addressing appliance classification and prediction. This paper expands and evaluates appliance load signatures based on V-I trajectory-the mutual locus of instantaneous voltage and current waveforms-for precision and robustness of prediction in classification algorithms used to disaggregate residential overall energy use and predict constituent appliance profiles. We also demonstrate the use of variants of differential evolution as a novel strategy for selection of optimal load models in context of energy disaggregation. A publicly available benchmark dataset REDD is employed for evaluation purposes. Our experimental evaluations indicate that these load signatures, in conjunction with a number of popular classification algorithms, offer better or generally comparable overall precision of prediction, robustness and reliability against dynamic, noisy and highly similar load signatures with reference to electrical power quantities and harmonic content. Herein, wave-shape features are found to be an effective new basis of classification and prediction for semi-automated energy disaggregation and monitoring.

AdOpt: An Adaptive Optimization Framework for Large-scale Power Distribution Systems

Optimizing self-evolving and dynamically changing systems is a grand challenge. In order to apply optimizations almost all conventional optimization techniques require a runtime system model. However, system models and their solution techniques vary in their strengths and limitations. For a rigid system, a single system model is acceptable. But if the system is constantly changing its structure then a rigid model is not able to represent the system properly, resulting in an inefficient use of technique in some cases. Therefore, in this paper we propose a framework for an optimization engine that adapts the optimization technique based on the system state. The adaptation involves selection of techniques based on historical statistics and current data, and dynamic generation of a model at runtime. This runtime model is then used to apply a relevant optimization technique to find a desired optimization plan for the system. We have evaluated the proposed framework on an electricity distribution system. Our results show that the proposed framework is adaptable, fast and able to manage numerous situations.

A Penny Saved is a Penny Earned: Applying Optimization Techniques to Power Management

Shortage of electricity is a major problem in many developing countries. Unfortunately, for some of these countries the only solution to this problem is to shut down complete electricity supply to a few neighborhoods to make up for the gap between demand and supply. To this end, we have developed a self-optimization approach to reduce the gap between demand and supply through remotely controlling high powered electric devices such as air conditioners. In this approach we have used mathematical optimization techniques such as linear programming to intelligently manage the electricity distribution. Not only through this approach we have been able to provide service-level guarantees to the consumers but we have also shown that our approach is fast, scalable and has the ability to handle unscheduled spikes in the system.

On the Use of Linear Programming in Optimizing Energy Costs

Efficient energy consumption in large sets of electric devices is a complex problem since it requires a balance between many competing factors. Presently, self-optimization techniques work expeditiously on small and relatively less complex problems. However, these techniques are not shown to be scalable on large and complex problems. In this paper we have used linear programming to optimize the use of energy in a typical environment that consists of large number of devices. Our initial results show that LP is fast, predictable and scalable. Moreover, we have also observed that modeling in LP is quite simple as compared to other self-optimization techniques.

An adaptive optimization model for power conservation in the smart grid

Dynamically adaptive systems (DAS) such as smart grids, cloud computing applications, sensor networks and P2P networks tend to change their structure at runtime. Therefore, design-time modeling for such systems are sometimes not enough to incorporate self-* properties. To this end, we have developed a dynamic mathematical modeling framework for runtime optimizations for DAS. In this paper, we describe how our system engineers a linear programming model by using a smart-grid application for power distribution as a case-study. At runtime whenever an optimization is desired this modeling framework captures the state of the system, converts it into an appropriate linear programming model, plan the changes using mathematical manipulations and apply the changes to the actual system. Our results show that this framework is able to capture accurate runtime models of large power systems and is able to adapt itself with the change in the size or structure of the system.

Engineering Optimization Models at Runtime for Dynamically Adaptive Systems

Dynamically adaptive systems (DAS), such as smart grids, cloud computing applications, sensor networks and P2P networks tend to change their structure at runtime. Therefore, design-time modeling for such systems are sometimes not enough for self-management. To this end, we have developed a dynamic mathematical modeling framework for runtime modeling for DAS. In this paper, we describe how our system engineers a linear programming model for self-optimization by using a smart-grid application for power distribution as a case-study. At runtime whenever, an optimization is desired this modeling framework captures the state of the system, converts it into an appropriate linear programming model, plan the changes using mathematical manipulations and apply the changes to the actual system. Our initial simulation results show that this framework is able to capture accurate runtime models of large power systems and is able to adapt itself with the change in the size or structure of the system by constructing a succinct model which is faster and more efficient than a design time model.

Towards a Self-managing Tool for Optimizing Energy Usage in Buildings

Smart grid is the next generation of electricity generation, transmission and distribution technology. A major component of smart grid is an overlay communication network for two-way communication between the power providers and the customers. With this feature smart grid provides exciting new ways of energy management and conservation. One of ways to conserve energy using a smart grid is to control and optimize energy usage in buildings. Buildings consume more than one third of the energy produced in the world. Therefore, conserving energy in buildings is cited as the “most important fuel” in energy generation. To this end, we have developed a self-managing approach to optimize energy usage in buildings. We have evaluated our approach using a software tool called Power Conservation Analysis Tool (PCAT). Our initial results using PCAT show upto 38% savings in the energy bills of customers that could directly translates into reduction in energy production costs for power producers.

Poster: Maximizing Renewable Energy Usage in Buildings using Smart Energy Switching Platform

Solar energy is slated to be an important energy source for reducing dependence in fossil fuels. Past few years have seen unprecedented deployments of solar energy in many countries of the developed world. However, solar energy uptake in developing countries is rather slow. This is particularly true for solar energy installations on buildings. Since buildings consume more than 40% of energy, it is important that greener buildings are encouraged through on-site production of renewable energy [2]. However, limited possibility of energy buyback programs in developing countries is one of the reason for less solar deployments. Also, in some countries the electricity infrastructure is so fragile that energy buyback programs at smaller scale are not feasible. Therefore, if the building owners like to go green then huge battery banks are needed to make the best use of solar energy. But battery banks add quite a bit of cost to the overall solar energy infrastructure in buildings. This extra upfont and maintenance cost is one of the reasons that hampers the growth of solar on buildings. Hybrid solar energy systems are used when energy buyback programs are available [1]. In this work we make a case that traditional hybrid solar energy deployments are not feasible for places without energy buyback programs. Instead we propose an idea of a solar energy system for buildings

SmartDSM: a layered model for development of demand side management in smart grids

Growing power demand and carbon emissions is motivating utility providers to introduce smart power systems. One of the most promising technology to deliver cheaper and smarter electricity is demand side management. A DSM solution controls the devices at user premises in order to achieve overall goals of lower cost for consumer and utility. To achieve this various technologies from different domains come in to play from power electronics to sensor networks to machine learning and distributed systems design. The eventual system is a large, distributed software system over a heterogeneous environment and systems. Whereas various algorithms to plan the DSM schedule have been proposed, no concerted effort has been made to propose models and architectures to develop such a complex software system. This lack of models provides for a haphazard landscape for researchers and practitioners leading to confused requirements and overlapping concerns of domains. This was observed by the authors in developing a DSM system for their lab and faculty housing. To this end in this paper we present a model to develop software systems to deliver DSM. In addition to the model, we present a road map of software engineering research to aid development of future DSM systems. This is based on our observations and insights of the developed DSM systems.

Software engineering for simulation systems in medical training: some initial experiences

In this paper, we describe our experiences of an ongoing project to develop a simulation system for ultrasound training. Ultrasound is a non-invasive technique to scan bodily organs using harmless sound waves. Because of the effectiveness of the ultrasound technology it is widely used by medical practitioners in health care establishments. However, gaining expertise in ultrasound technology requires a long time because of the complex hand-eye coordination required to perform ultrasound scans. To this end, the goal of this simulation system is to help medical practitioners gain expertise in ultrasound technology in relatively short period of time. In this paper we describe the requirements engineering, design and developmental challenges of this simulation system along with some initial evaluations.