Show as: List Detail Sort by: Year Title Reset search

2017

BibTeX

Towards Robust Energy Demand Response At Airports

Abstract

Airports have a great motivation for taking advantage of demand response (DR) opportunities considering their large energy footprint and continuous operation. A robust energy baseline model, which calculatates what the power demand would have been without any curltailment, is crucial to realize this motivation as it assesses the DR potential and the effectiveness of DR strategies. Since such baseline models are specific to building types and operational characteristics, this study targets developing an airport-specific energy baseline model to help airport operators utilize DR opportunities. For the purpose, first we perform visual inspection to analyze the relationships between the power demand and explanatory variables, such as time-of-day, time-of-week, outside temperature, and the number of passengers of departure flight and arrival flights. Then, we develop airport-specific energy baseline models through linear regression analysis with ten different combinations of explanatory variables. Finally, we analyze the regression coefficients of each model to understand the impact of variables on the airport power demand. The results show the model with time-of-week and outside temperature has the lowest mean absolute percentage error (MAPE) of 2.72% (305.87 kW) and using time-of-week rather than time-of-day reduce the error by about 4.1 ~ 4.8 kW. It is also found that both departure and arrival flight schedules do not significantly increase the prediction accuracy. Through coefficients analysis, we also find the isolated impact of each variable on the airport power demand, which inform airport operators about the contribution of each variable, such as flight schedule, to the whole airport power demand.

2016

BibTeX
BibTeX

The Neural Energy Decoder Energy Disaggregation by Combining Binary Subcomponents

Abstract

In this paper a novel approach for energy disag- gregation is introduced that identifies additive subcomponents of the signal in an unsupervised way. In a subsequent step, combinations of these subcomponents are sought that constitute appliances. Once the subcomponents that constitute an appliance are identified, energy disaggregation can be viewed as non- linear filtering of high frequency current readings. The approach introduced here tries to avoid numerous pitfalls of existing energy disaggregation techniques such as computational complexity issues, data transmission limitation or unrealistic assumptions about prior knowledge of appliances. The proposed method is used to infer the states of appliances in the BLUED dataset.

BibTeX

Efficient Inference in Dual-Emission FHMM for Energy Disaggregation

Abstract

In this paper an extension to factorial hidden Semi Markov Models is introduced that allows modeling more than one sequence of emissions of the individual HMM chains, as well as a joint emission of all chains. Since exact inference in factorial hidden Markov Models is computationally intractable, an approximate inference technique is introduced that reduces the computational costs by first constraining the successor state space of the model, allowing state changes at statistically significant points in time (events) and by discarding low probability paths (truncating). Furthermore, by being agnostic about state durations the computational costs are further decreased. These assumptions allow for efficient inference that is less susceptible to local minima and allows one to specify the computational burden a priori. The performance of the inference technique is evaluated empirically on a synthetic data set whereas incorporating the feature emissions is evaluated on real world data in the context of energy disaggregation. Energy disaggregation tackles the problem of decomposing whole home energy measurements into the power traces of constituent appliances, and is a natural application for this type of models.

2015

Download PDF BibTeX

A feasibility study of automated plug-load identification from high-frequency measurements

Abstract

Plug-meters benefit many grid and building-level energy management applications like automated load control and load scheduling. However, installing and maintaining large and/or long term deployments of such meters requires assignment and updating of the identity (labels) of electrical loads connected to them. Although the literature on electricity disaggregation and appliance identification is extensive, there is no consensus on the generalizability of the proposed solutions, especially with respect to the features that are extracted from voltage and current measurements. In this paper, we begin to address this problem by comparing the discriminative power of commonly used features. Specifically, we carry out tests on PLAID, a publicly available high-frequency dataset of hundreds of residential appliances. By examining how the classification accuracy changes with sampling frequency, we also explore the computational complexity of these techniques to understand the feasibility and design of a hardware setup that can perform these calculations in near real-time.

Download PDF BibTeX

A Data-driven Meta-data Inference Framework for Building Automation Systems

Abstract

Building automation systems (BAS) are believed to hold the key to significantly reducing the average energy consumption of our residential and commercial building stock, which in the U.S. is responsible for 41% of the total annual energy use in 2014. As these systems become more widespread and inexpensive, the complexity and challenges associated with their installation, maintenance and upkeep will increase. One of the primary challenges is the generation and update of the meta-data associated with the sensors and control points distributed throughout the facility. Previous research has attempted to reduce the human input required to perform these activities, by leveraging different signal processing and statistical analysis approaches to infer the sensor types and locations from measurements and/or tags obtained through a BAS. However, because of the relatively small sample size, the feasibility of applying these type approaches on large buildings, as well as their generalizability, remain as unsolved questions. In this paper, we propose a meta-data inference framework to learn from BAS measurement data in a semi-automated way. Furthermore, we evaluate the framework on two large buildings instrumented with thousands sensors and show the feasibility of applying data driven approaches in the real world. We present the results of our study and provide recommendations for future work in this area.

BibTeX

Authors

BibTeX

Semi-Automatic Labeling for Non-Intrusive Load Monitoring Datasets

Abstract

In this paper we present and evaluate a semi- automatic labeling prototype to enable the creation of fully labeled energy disaggregation datasets from sub-metered data. Our results advocate in favor of our approach and show that it is possible to extract individual appliance transitions with considerable precision, as long as the individual appliance information is present in the sub-metered data, and its resolution is high enough.

2014

BibTeX

Quantifying Flexibility of Thermostatically Controlled Loads for Demand Response:A Data-driven Approach (Accepted)

Abstract

Power systems are undergoing a paradigm shift due to the influx of variable renewable generation to the supply side. The resulting increased uncertainty has system operators looking to new resources, enabled by smart grid technologies, on the demand side to maintain the balance between supply and demand. This study uses a unique data set to estimate and validate models of demand response from residential thermostatically controlled loads (TCLs)—specifically, HVAC units—and quantifies the extent to which a population of TCLs can provide demand response (DR). We use measured temperature setpoints, internal temperatures, compressor cycling ratio and metered energy data collected from over 4200 homes in Texas during the summer of 2012. Using transfer function models for individual households, we investigate the instantaneous power shed, the duration of the power shed, steady state energy savings and total energy savings. Specifically, we provide insight into the dependency of household DR availability to the temperature setpoint schedule, outdoor air temperature and time of the day.

Authors

BibTeX

Demo Abstract: Mortar.io: A Concrete Building Automation System

Abstract

The commoditization of wireless sensing systems makes it feasible to include BAS functionality in small and medium-sized buildings. The configuration complexity and cost of installation is now the dominant barrier to adoption. In this demo we introduce a platform called Mortar.io, which focuses on ease-of-installation, secure configuration, and management of BAS sub-systems in a manner that can scale from small to large installations. Unlike cloud-reliant systems, Mortar.io distributes storage and control functionality across end devices making it robust to network and internet outages. The system, once initialized, can run autonomously on a low-cost controller within a building or connect to the cloud for remote monitoring and configuration. We will also show our efficient multi-resolution data store that buffers data locally and replicates aggregate data across devices for reliability. A publish-subscribe model built on top of XMPP is used for messaging with per-device access control and a transducer schema. Finally, a web portal provides an interface to monitor and schedule lighting, plug-loads, environmental sensors and HVAC from a single uniform interface

BibTeX

Demo Abstract: PLAID: a Public Dataset of High-Resolution Electrical Appliance Measurements for Load Identification Research

Abstract

We introduce the Plug-Level Appliance Identification Dataset (PLAID), a public and crowd-sourced dataset for load identification research consisting of short voltage and current measurements (in the order of a few seconds) for different residential appliances. The goal of PLAID is to provide a public library for high-resolution appliance measurements that can be integrated into existing or novel appliance identification algorithms. PLAID currently contains measurements for more than 200 different appliance instances, representing 11 appliance classes, and totaling more than a thousand records. In this demo we summarize the existing dataset, demonstrate how new records can be added to the library using a web interface and, finally, walk through a live example of how the library can be integrated into an existing non-intrusive load monitoring (NILM) algorithm framework.

BibTeX
BibTeX

Authors

BibTeX
Download PDF BibTeX

Modeling Thermostatically Controlled Loads to Engage Households in the Smart Grid: Lessons Learned from Residential Refrigeration Units

Abstract

As renewable generation capacity in the power grid increases, keeping the balance between the supply and demand becomes difficult. This threatens the grid’s stability and security. Existing power reserve assets and regulation methodologies fail to provide the short-term responses required to keep the load and generation balanced as the amount of renewable generation increases. Hence, researchers proposed to increase the information exchange within the power network and to introduce real-time demand control to ensure robustness while accommodating the intermittent nature of these generation resources. Constituting a significant portion of the electrical demand of buildings, thermostatically controlled loads (TCLs) are well-suited to provide real-time demand control. In this paper, we shed light on challenges associated with engaging TCLs to the power grid using a centralized control strategy. We focus on the challenges associated with simulating a realistic TCL population using the models that are proposed in the literature. Specifically, we use data collected from residential refrigeration units operating in 214 different households to propose a strategy to select parameters when simulating a TCL population.

Download PDF BibTeX

Investigation on the effects of environmental and operational conditions (EOC) on diffuse-field ultrasonic guided-waves in pipes

Abstract

In spite of many favorable characteristics of guided-waves for Nondestructive Evaluation (NDE) of pipes, real-world application of these systems is still quite limited. Beside the complexities derived from multi-modal, dispersive and multi-path characteristics of guided-waves, one of the main challenges in guided-wave based NDE of pipelines is sensitivity of these systems to variations of environmental and operational conditions (EOC).

This paper investigates the effects of varying EOCs on guilded-wave based NDE of pipelines. We first provide a review of the studies to date in the field of guided-wave based testing to identify research gaps for enhancing the application of these systems in pipeline NDE. To study the identified gaps, guided-wave data from a fully operational piping system, with continuously varying flow rate and temperature, is used. Time-shift and amplitude drift effects due to flow rate variations are evaluated along with those of temperature. It is observed that masking effects of flow rate for damage detection can be at least as significant as temperature effects, and that such effects become more dominant when flow rate and temperature variations co-occur.

 

Authors

Download PDF BibTeX

Anomaly Detection on Piezometer Data Collected from Embankment Dams Using Physical Model-Based Simulation

Abstract

Embankment dams, like most other civil infrastructure systems, are exposed to harsh and largely unpredictable environments. However, unlike bridges, buildings and other structures, their design specifications and as-is properties are not generally known in the same level of detail due to, among other things, their age and the difficulties associated with assessing their internal structure. Hence, making sense of measurements collected from instruments used to monitor their behavior requires sound engineering judgment and analysis, as well as robust statistical analysis techniques to prevent misinterpretation. In the United States (US), the current practice of analyzing the structural integrity of embankment dams relies primarily on manual a posteriori analysis of instrument data by engineers, leaving much room for improvement through the application of automated data analysis techniques. In our previous work, we presented the effectiveness of applying statistical anomaly detection techniques – such as Principal Component Analysis and Robust Regression Analysis – when analyzing piezometer data collected from embankment dams. In this paper, we present how we could improve our work by testing with simulated anomalies that are indicative of internal erosion problems. In order to closely replicate more realistic anomalous scenarios, a physics-based model of an embankment dam was developed. By varying a hydraulic conductivity of a soil material in the model, corresponding detection accuracies and sensitivities of the statistical anomaly detection algorithm were evaluated. When we applied our proposed anomaly detection on more realistically simulated anomalous data using the numerical model, the detection accuracy came out to be 98.5%.

Read More: http://ascelibrary.org/doi/abs/10.1061/9780784413616.220

BibTeX

SURF and SURF-PI: A File Format and API for Non-Intrusive Load Monitoring Public Datasets

Abstract

In this paper we propose a common file format and API for public Non-Intrusive Load Monitoring (NILM) datasets such that researchers can easily evaluate their approaches across the different datasets and benchmark their results against prior work. The proposed file format enables storing the power demand of the whole house along with individual appliance consumption, and other relevant metadata in a single compact file, whereas the API supports the creation and manipulation of individual files and datasets in the proposed format.

Download PDF BibTeX

Toward characterizing the effects of environmental and operational conditions on diffuse-field ultrasonic guided-waves in pipes

Abstract

One of the main challenges in real-world application of guided-waves based nondestructive evaluation (NDE) of pipelines is their sensitivity to changes in environmental and operational conditions (EOC) that these structures are subject to. In spite of many favorable characteristics of guided-waves for NDE of pipes, their multi-modal, dispersive, and multi-path characteristics result in complex signals whose interpretation is a difficult task.

Studies that have considered the effects of EOC variations either fail to reflect realistic EOC scenarios (e.g., limited to particular effects of specific EOCs, like time shifting effects of temperature in plates) or lack the necessary understanding of the effects of EOC variations on different aspects of the developed damage detection approaches. Such gaps limit the extensibility of these approaches to pipeline applications outside of controlled environments.

This paper motivates the idea of analytically incorporating the effects of temperature and flow rate variations into damage diagnosis of pipes, through a number of case studies. A review of the existing literature on guided-wave based testing is also provided. For damage detection, a linear supervised classification method, namely linear discriminant analysis (LDA), is applied to experimental guided-wave data recorded from a hot water piping system under regular operation. Principal components, obtained through principal component analysis (PCA), and Fourier transforms of the signals are two sets of damage-sensitive features (DSF) that are examined for LDA-based classification. The effects of temperature and flow rate difference among testing and training datasets on (A) detection performance and (B) goodness of fit of the method to the data are investigated.

2013

Download PDF BibTeX

A Moving Horizon State Estimator in the Control of Thermostatically Controlled Loads for Demand Response

Abstract

The quality and effectiveness of the load following services provided by centralized control of thermostatically controlled loads depend highly on the communication requirements and the underlying cyber-infrastructure characteristics. Specifically, ensuring end-user comfort while providing realtime demand response services depends on the availability of the upstream information provided from the thermostatically controlled loads to the main controller regarding their operating statuses and internal temperatures. State estimation techniques can be used to infer the necessary information from the aggregate power consumption of these loads, replacing the need for an upstream communication platform carrying information from appliances to the main controller in real-time. In this paper, we introduce a moving-window mean squared error state estimator with constraints as an alternative to a Kalman filter approach, which assumes a linear model without constraints. The results show that some improvement is possible for scenarios when loads are expected to be toggled frequently.

Authors

Download PDF BibTeX

One Size Does Not Fit All: Averaged Data on Household Electricity is Inadequate for Residential Energy Policy and Decisions

Abstract

Residential electricity users need more detail than monthly bills to reduce consumption. With the emergence of technologies that provide detailed usage estimates for energy consumption, two questions arise. First, how many different energy-consuming appliances contribute to household electricity load, and secondly which appliances? Using national average penetration rates, the Residential Energy Consumption Survey (RECS), estimates that 42 unique appliances account for 93% of electricity consumption, while 12 appliances account for 80% of average household electric load. A typical scenario is developed from national and regional penetration rates and find that eight appliances are responsible for 80% of a household’s electric load in the United States. Four household scenarios are developed: a house that uses electric appliances, gas appliances, the average household, and typical household. It is concluded that RECS cannot be used as a representative household as it overestimates the number of appliances that contribute to a household electric load. The number of significant appliances is affected by appliance ownership and use, which is more variable between homes than between census divisions. These results can be used to design and maximize the value of residential energy information and management systems.

Download PDF BibTeX

Novel Techniques For ON and OFF states detection of appliances for Power Estimation in Non-Intrusive Load Monitoring

Abstract

Non-Intrusive Load Monitoring (NILM) is a method of extracting appliance-level power consumption information from aggregate circuit-level data with the goal of giving users feedback regarding their energy consumption so they can take control of their consumption habits. In this paper, we present a novel algorithm for classification of on and off states of appliances. We compare the performance of our algorithm in on state detection with a pervious paper that evaluated the same dataset and show that it performs up to 13% better. We also present the results of a case study where we collected data for different modes of a cooktop, microwave and dishwasher and used our algorithms to perform power estimation. The error on ten different setups in the test bed ranges from 1% to 32%. We discuss our results and lay out ideas for future work.

BibTeX

Towards automated appliance recognition using an EMF sensor in NILM platforms

Abstract

Non-Intrusive Load Monitoring (NILM) has been studied for a few decades now as a method of disaggregating information about appliance level power consumption in a building from aggregate measurements of voltage and/or current obtained at a centralized location in the electrical system. When such information is provided to the electricity consumer as feedback, they can then take the necessary steps to modify their behavior and conserve electricity.  Research has shown potential for savings of up to 20% through this kind of feedback.  The training phase required to allow the algorithms to recognize appliances in the home at the beginning of a NILM setup is a big hindrance to wide adoption of the technique. One of the recent advances in this research area includes the addition of an Electro-Magnetic Field (EMF) sensor that measures the electric and magnetic field nearby an appliance to detect its operational state. This information, when coupled with the aggregate power consumption data for the home, can help to train a NILM system, which is a significant step forward in automating the training phase. This paper explores the theory behind the operation of the EMF sensor and discusses the feasibility of automating the training and classification process using these devices. A case study is presented, where magnetic field measurements of 8 appliances are analyzed to determine the viability of using these signals alone to determine the type of appliance that the EMF sensor has been placed next to. Various dimensionality reduction techniques are applied to the collected data, and the resulting feature vectors are used to train a variety of common machine learning classifiers. A vector subspace obtained using Independent Component Analysis (ICA), along with a k-NN classifier, was found to perform best among the different alternatives explored. Possible reasons behind the findings are discussed and areas for further exploration are proposed.

Download PDF BibTeX

Towards automatic classification of appliances: Tackling cross talk in EMF sensors using blind source separation techniques

Abstract

Non Intrusive Load Monitoring (NILM) is the method of obtaining information about appliance-level power consumption inside a building from voltage and/or current measurements made at a central location in the electrical system. The system that does this kind of power disaggregation typically relies on some sort of training step to help it recognize what types of changes in the observed signal correspond to what kinds of appliances. To help train NILM systems recognize appliances within the house, auxiliary sensors like Electromagnetic Field (EMF) sensors have been proposed (Rowe et al., 2010). One of the problems with using EMF sensors to automatically recognize an appliance is that of cross talk. Since a typical home constitutes of settings where there are multiple appliances at close vicinity, EMF sensors are prone to picking up unwanted signals from appliances that are not of interest. In this paper, we use blind source separation techniques like Independent Component Analysis to remedy cross talk in EMF sensors.

Authors

Download PDF BibTeX

Interpreting the Dynamics of Embankment Dams through a Time-Series Analysis of Multiple Piezometer Data Using a Non-Parametric Spectral Estimation Method

Abstract

A common approach used by engineers to monitor and analyze data collected from piezometers installed in embankment dams is to generate time history plots and visually identify any spikes or anomalies in them. However, such practice has several limitations when capturing complicated relationships among a number of factors that affect piezometric readings. This is especially true when periodic or dominant variations that exist in time-series data are of concern, given that environmental and process noise can sometimes mask these variations. In this paper, we propose applying Moving Principal Component Analysis (MPCA) and Robust Regression Analysis (RRA), which have been shown to be successful in other applications, to extract relevant components and detect anomalies in piezometer measurements, which are one of the most important data to be monitored when evaluating the performance of embankment dams. The proposed anomaly detection method provides a more efficient way of understanding and detecting changes in piezometer data.

Download PDF BibTeX

Demand Response in Buildings: Engaging Thermostatically Controlled Loads in the Power Grid

Abstract

Buildings accounted for nearly 75% of the electricity use in 2010, the largest portion among all sectors in the United States. Despite their contribution to the overall electricity demand, they have traditionally been considered as passive end-users of energy, and research in building engineering has largely focused on solutions to curtail their energy usage and improve energy efficiency. However, recently, the potential of buildings to become active participants in the electricity grid by providing ancillary services via direct load control has been garnering interest in the research community. In this paper, we introduce different demand response programs that use thermostatically controlled loads (TCLs) available in buildings. Specifically, we shed light on the existing work on direct load control for TCLs and identify the upcoming challenges associated with this approach. Finally, we introduce BUFFER: the building frequency forecast and electricity regulation framework, a novel decentralized and autonomous framework that uses TCLs in buildings to do frequency regulation for the power grid.

BibTeX

Spatiotemporal Dimensions of Network Density-based Clustering for Water Pipe Maintenance

Abstract

In the U.S., many drinking water assets operate beyond their designed lifetimes, and constrained resources necessitate optimizing repair and replacement. For pipe infrastructure, clustering tools can characterize their susceptibility to failure by identifying relationships among descriptive or measured features. In these complex systems, algorithmic learning approaches can provide a first insight before expert knowledge is applied, reducing time and labor. The state-of-the-art techniques often rely upon static characteristics. In this work, pipe maintenance records are analyzed through the network OPTICS (“Ordering Points To Identify the Clustering Structure”), which forms a hierarchal density-based clustering structure. The study compares and extends OPTICS to a temporal context; exploring the evolution of clustering structure provides additional insight. The findings suggest this spatiotemporal approach is applicable for improved asset management.

Authors

Download PDF BibTeX

A Magnetic Field-based Appliance Metering System

Abstract

Understanding where energy is being used in buildings is an important CPS component that can help improve energy conservation and efficiency. Current approaches for appliance-level energy metering typically require the installation of plug-through power meters, which is often difficult and costly for devices with inaccessible wires or outlets, or appliances that draw large amounts of current. In this paper, we present an energy measurement system that estimates the energy consumption of individual appliances using a wireless sensor network consisting of contactless electromagnetic fi eld (EMF) sensors deployed near each appliance, and a whole-house power meter. We present the design of a battery-operated EMF sensor, which can detect appliance state transitions within close proximity based on magnetic and electric fi eld fluctuations. Each detector wirelessly transmits state change events to a circuit-panel energy meter, in a time-synchronized fashion, so that the overall power measurements can be used to estimate appliance-level energy usage. The time synchronization and data throughput requirements of this problem motivated the development of a new low-power TDMA sensor networking protocol. Our EMF sensors are able to detect significant power state changes from a few inches away, thus making it possible to externally monitor in-wall wiring to devices. We experimentally evaluate our proposed EMF sensor, three-phase power meter and communication protocol in a residential building collecting data for over a week. The system is able to estimate appliance energy consumption with an average accuracy of 95.8%.

Authors

BibTeX

Domain-specific querying formalisms for retrieving in- formation of HVAC systems.

Abstract

In order to save energy and improve the control of indoor environments, researchers have developed hundreds of computer algorithms that can automatically and continuously analyze the conditions of Heating, Ventilation and Air-Conditioning (HVAC) systems. However, the complex information requirements of these algorithms inhibit deploying them in real-world facilities. We propose an integrated performance analysis framework that automatically collects, merges and provides the information required by them. In previous studies, we have identified a general set of information requirements for the computerized approaches and formalized a semi-automated approach that integrates multiple data models to support the required information. In order to automatically retrieve the information required by different approaches, the research discussed in this paper explored a query mechanism that can represent the required information in a formal way that can be reasoned about. We categorize the information items that are used to represent the information needs, formalize a domain-specific query language that can formally represent the query statements, and develop a library of mechanisms that can automatically reason about and retrieve the needed information. In order to validate the performance of the query language and mechanisms, we also developed a prototype, which includes a graphic user interface that helps users to define the queries, and the implementation of the reasoning mechanisms that process the queries. The precision and recall of the query language and mechanisms were tested using the queries identified from previous research.

2012

Authors

BibTeX

An Integrated Performance Analysis Framework for HVAC Systems Using Heterogeneous Data Models and Building Automation Systems

Abstract

More than 20% of the energy consumed by heating, ventilation and air-conditioning (HVAC) systems is wasted due to undetected faults in these systems. In the past three decades, researchers have developed hundreds of computer algorithms to automatically and continuously analyze their energy performance. However, due to the complex information required by these algorithms, it is very difficult for facilities operators to deploy them in real-world buildings.

This paper presents an integrated performance analysis framework (IPAF) that can be used to integrate heterogeneous data models about the building and HVAC systems and the dynamic data from embedded sensors and controllers. This framework facilitates the deployment of multiple performance analysis algorithms in different buildings and HVAC systems by automatically providing the information required by these algorithms. We developed and tested our proposed framework using four different types of algorithms in a real-world facility. The IPAF is able to integrate three heterogeneous data models with 85% of precision and 91% recall. The precision and recall for retrieving data required by the four different types of algorithms are both 100%.

Authors

Download PDF BibTeX

Using Smart Devices for System-level Management and Control in the Smart grid: A Reinforcement Learning Framework

Abstract

This paper presents a stochastic modeling framework to employ adaptive control strategies in order to provide short term ancillary services to the power grid by using a population of heterogenous thermostatically controlled loads. A classical Markov Decision Process (MDP) representation is developed to leverage existing tools in the field of reinforcement learning. Initial considerations and possible reductions in the action and state spaces are described. A Q-learning approach is implemented in simulation to demonstrate the performance of the presented adaptive control framework on a reference tracking scenario.

Authors

  • Kyle Anderson
  • Mario Bergés
  • Adrian Ocneanu
  • Diego Benítez
  • José M. F. Moura
Download PDF BibTeX

Event Detection for Non Intrusive Load Monitoring

Abstract

Monitoring electricity consumption in the home is an important way to help reduce energy usage and Non-Intrusive Load Monitoring (NILM) techniques are a promising approach to obtain estimates of the electrical power consumption of individual appliances from aggregate measurements of voltage and/or current in the distribution system. In this paper, we discuss event detection algorithms used in the NILM literature and propose new metrics for evaluating them. In particular, we introduce metrics that incorporate information contained in the power signal instead of strict detection rates. We show that this information is important for NILM applications with the goal of improving appliance energy disaggregation. Our work was carried out on a publicly-available week-long dataset of real residential power usage.

Authors

  • Kyle Anderson
  • Adrian Ocneanu
  • Diego Benítez
  • Derrick Carlson
  • Anthony Rowe
  • Mario Bergés
Download PDF BibTeX

BLUED: A Fully Labeled Public Dataset for Event-Based Non-Intrusive Load Monitoring Research

Abstract

The problem of estimating the electricity consumption of individual appliances in a building from a limited number of voltage and/or current measurements in the distribution system has received renewed interest from the research community in recent years. In this paper, we present a Building-Level fUlly-labeled dataset for Electricity Disaggregation (BLUED). The dataset consists of voltage and current measurements for a single family residence in the United States, sampled at 12 kHz for a whole week. Every state transition of each appliance in the home during this time was labeled and time-stamped, providing the necessary ground truth for the evaluation of event-based algorithms. With this dataset, we aim to motivate algorithm development and testing. The paper describes the hardware and software configuration, as well as the dataset’s benefits and limitations. We also present some of our detection results as a preliminary benchmark.

Download PDF BibTeX

A study on the feasibility of automated data labeling and training using an EMF sensor in NILM platforms

Abstract

Non-Intrusive Load Monitoring (NILM) has been studied for a few decades now as a method of disaggregating information about appliance level power consumption in a building from measurements obtained at a centralized location in the electrical system. The training phase required at the beginning of a NILM setup is a big hindrance to wide adoption of the technique. One of the recent advances in this research area includes the addition of an Electro-Magnetic Field (EMF) sensor that measures the electric and magnetic field around an appliance to detect its state. This information, when coupled with the aggregate power data, can effectively train a NILM system almost automatically, which is a significant step towards automating the training phase. This paper explores the theory behind the operation of the EMF sensor and analyzes the feasibility in terms of automating the training and classification process. It then outlines our plan for further analysis.

Authors

BibTeX

Requirements for A Formal Approach to Represent Information Exchange Requirements of a Self-managing Framework for HVAC Systems

Abstract

In order to improve the energy efficiency of heating, ventilation and air conditioning (HVAC) systems, researchers have developed computer algorithms to automatically detect, diagnose and mitigate faults that cause energy waste. However, because the different types of information required by these algorithms are commonly generated by different stakeholders using various formats, it is difficult for the system operators to retrieve the information required to implement these algorithms. In order to overcome this issue, the problem of automatically collecting, integrating and providing the algorithms with the required information that is stored in heterogeneous formats needs to be solved. Only then it might be possible to streamline the applications and utilization of these algorithms and consequently achieve their energy saving potential.

Such an automated information retrieval approach requires a formal way to represent the information exchange requirements that map the needed building-related information and heterogeneous information sources. Because the needed information is stored in different data models, there is no existing approach that is able to represent the information in all data models. The objective of this paper is to analyse the requirements for a formal approach to represent the mappings of needed information from different sources, and compare the characteristics of the existing representation approaches. The discussions include an identification of the data models that are used to store the needed information, a synthesis of the formats and schemas of these data models, an analysis of the existing formal approaches for representing the information exchange requirements, and a comparison of these approaches. By analysing the requirements and comparing the existing approaches, this paper concludes with suggestions to guide the development of a formal approach to represent the information exchange requirements.

Authors

Download PDF BibTeX

The Design of a Hardware-software Platform for Long-term Energy Eco-feedback Research

Abstract

Researchers often face engineering problems, such as optimizing prototype costs and ensuring easy access to the collected data, which are not directly related to the research problems being studied. This is especially true when dealing with long-term studies in real world scenarios. This paper describes the engineering perspective of the design, development and deployment of a long-term real word study on energy eco-feedback, where a non-intrusive home energy monitor was deployed in 30 houses for 18 months. Here we report on the efforts required to implement a cost-effective non-intrusive energy monitor and, in particular, the construction of a local network to allow remote access to multiple monitors and the creation of a RESTful web-service to enable the integration of these monitors with social media and mobile software applications. We conclude with initial results from a few eco-feedback studies that were performed using this platform.

Authors

BibTeX

A Taxonomy for Depicting Geometric Deviations of Facilities Extracted from Comparisons between Laser-Scanned Point Clouds and 3D Models

Abstract

Building components are subject to diverse changes throughout their lifecycles. Several tasks, such as construction quality control and structural health monitoring, require accurate information of the existing condition, and they involve comparing the current status of the building with the models to identify possible discrepancies. However, there is a lack of formalisms for representing the identified deviations so that they can be easily understood, evaluated and addressed by engineers and managers. Accurate and complete understanding of how, and the extent to which, buildings deviate from models is necessary for a number of decisions made throughout the building’s lifecycle. This paper addresses this need by proposing a taxonomy for depicting geospatial deviations identified through comparison of as-is data (i.e. laser scanned point clouds in this paper) and building models. This taxonomy is an initial step toward formalizing the representation of geospatial deviations and communicating them in a machine-interpretable manner to support data-model comparison, and model evaluation tasks.

Authors

BibTeX

Authors

  • Ijung
  • James H. Garrett
  • Lucio Soibelman
Download PDF BibTeX

Application of Classification Models and Spatial Clustering Analysis to a Sewage Collection System of a Mid-Sized City

Abstract

Improving asset management of infrastructure systems has been an ongoing issue in the United States. Oliveira et al. (2010, 2011) developed several approaches to better understand the nature and location of pipe breaks in a drinking water distribution system. In this paper, we applied these two approaches to another infrastructure system..the pipe network of a sewage collection system. We first applied several classification approaches to analyze factors associated with higher density regions of deteriorating pipes in the sewage collection system. Relevant attributes that cause poorly conditioned sewer pipes could be found using this approach. We then applied the network version of a density-based clustering algorithm created by Oliveira et al. (2010), used to detect clustered regions of pipe breaks in water distribution systems, to detect hierarchically clustered regions in one of the high density regions of pipe deterioration in the same pipe network. This latter approach was found to provide useful information and additional insight about the local attributes that might be related to the high-density of pipe deterioration.

Authors

BibTeX

Effects of Pre-scanning and Scanning Approaches on the Quality of Processed Laser Scanned Data: Lessons Learned

Abstract

Several tasks in the Architectural, Engineering, Construction and Facility Management (AEC&FM) domain require real-time geospatial information. Traditional measurement methods are widely being replaced by more advanced technologies. 3D laser scanners are increasingly being used for diverse tasks during the lifecycle of a facility. However, a variety of process, environmental, scanner and analysis related factors can affect the quality of laser scanned data. Error sources, particularly those inherent in early stages of the laser scanning process, have not been well-addressed. This paper focuses on the effects of such error sources on the quality of the raw/processed data. Lessons learned through the planning, laser scanning and data processing of a facility are shared. The dataset consists of a total of 68 scans from both indoor and outdoor environments collected through using two different types of scanners. The effects of pre-scan site visits and proactive planning on reducing the consequences of different error sources are illustrated through examples. Lessons learned emphasize the need for formalization of the planning stage of laser scanning process, and data collection approaches for the AEC&FM domain applications, to enhance the quality of the raw/processed data.

2011

Authors

Download PDF BibTeX

Poster Abstract: Appliance Classification and Energy Management Using Multi-Modal Sensing

Abstract

In this demonstration, we introduce a low-cost energy management system that tracks appliance energy usage and identifies particular sources of waste that can be optimized. In order to better understand appliance usage patterns, we correlate electrical load information with environmental sensors to identify clusters. These patterns can be used to identify when devices are accidentally left active in unoccupied rooms and provide a means to identify excessive consumption. The correlation is based on learned information over time and hence requires minimal manual labeling. Our system combines measurements from a circuit-panel energy
meter with multiple low-cost wireless sensors. We utilize an EMF-based appliance state detector that when combined with circuit-panel and plug-load energy meters allows the system to track the energy consumption of loads at a lower cost and in a less invasive manner than previous metering systems. We deployed our system in a house, collecting data from over 60 sensing points for more than six months. During this period, the system was able to identify wasteful energy usage as high as 17% of the total daily consumption.

Download PDF BibTeX

Towards Automated Detection and State Tracking of Artificial Light Sources From Sequential Pictures Inside Buildings

Abstract

In this paper the problem of automatically determining the number, location and operational schedule of artificial light sources present in a sequence of photographs taken from a static location inside a building at regular intervals is discussed. In particular, the focus is on the necessary computational support for users reviewing such datasets. The most reliable way to measure an individual light fixture is to place a light meter much closer to the lamp than to any other light source (other lights, windows, etc.) This process is complicated for buildings with hardto- access light fixtures such as high-ceilinged warehouses, manufacturing facilities, hangars, outdoor parking lots, etc. Existing solutions to similar problems in other domains are reviewed and a prototype system based on simple image processing techniques is evaluated to illustrate the challenges that are specific to this problem and illustrate the needed future work.

Authors

Download PDF BibTeX

Towards Automated Detection and State Tracking of Artificial Light Sources From Sequential Pictures Inside Buildings

Abstract

In this paper the problem of automatically determining the number, location and operational schedule of artificial light sources present in a sequence of photographs taken from a static location inside a building at regular intervals is discussed. In particular, the focus is on the necessary computational support for users reviewing such datasets. The most reliable way to measure an individual light fixture is to place a light meter much closer to the lamp than to any other light source (other lights, windows, etc.) This process is complicated for buildings with hard-to-access light fixtures such as high-ceilinged warehouses, manufacturing facilities, hangars, outdoor parking lots, etc. Existing solutions to similar problems in other domains are reviewed and a prototype system based on simple image processing techniques is evaluated to illustrate the challenges that are specific to this problem and illustrate the needed future work.

Authors

Download PDF BibTeX

A time-frequency approach for event detection in non-intrusive load monitoring

Abstract

Non-intrusive load monitoring is an emerging signal processing and analysis technology that aims to identify individual appliance in residential or commercial buildings or to diagnose shipboard electro-mechanical systems through continuous monitoring of the change of On and Off status of various loads. In this paper, we develop a joint time-frequency approach for appliance event detection based on the time varying power signals obtained from the measured aggregated current and voltage waveforms. The short-time Fourier transform is performed to obtain the spectral components of the non-stationary aggregated power signals of appliances. The proposed event detector utilizes a goodness-of-fit Chi-squared test for detecting load activities using the calculated average power followed by a change point detector for estimating the change point of the transient signals using the first harmonic component of the power signals. Unlike the conventional detectors such as the generalized likelihood ratio test, the proposed event detector allows a closed form calculation of the decision threshold and provides a guideline for choosing the size of the detection data window, thus eliminating the need for extensive training for determining the detection threshold while providing robust detection performance against dynamic load activities. Using the real-world power data collected in two residential building testbeds, we demonstrate the superior performance of the proposed algorithm compared to the conventional generalized likelihood ratio detector.

Authors

  • S. Mohsen Shahandashi
  • S. N. Razavi
  • Lucio Soibelman
  • Mario Bergés
  • Carlos H. Caldas
  • Ioannis Brilakis
  • Jochen Teizer
  • Patricio Vela
  • Carl Haas
  • James H. Garrett
  • Burcu Akinci
  • Z. Zhu
Download PDF BibTeX

CEC: Data Fusion Approaches and Applications for Construction Engineering

Abstract

Data fusion can be defined as the process of combining data or information for estimating the state of an entity. Data fusion is a multi-disciplinary field that has several benefits, such as enhancing the confidence, improving reliability and reducing ambiguity of measurements for estimating the state of entities in engineering systems. It can also enhance completeness of fused data that can be required for estimating the state of engineering systems. Data fusion has been applied to different fields, such as robotics, automation, and intelligent systems. This paper reviews some examples of recent applications of data fusion in civil engineering and presents some of the potential benefits of using data fusion in civil engineering.

Authors

  • Mario Bergés
  • Ethan Goldman
  • Lucio Soibelman
  • H. Scott Matthews
  • Kyle Anderson
Download PDF BibTeX

User-centered Non-Intrusive Electricity Load Monitoring for Residential Buildings

Abstract

This paper presents a nonintrusive electricity load-monitoring approach that provides feedback on the energy consumption and operational schedule of electrical appliances in a residential building. This approach utilizes simple algorithms for detecting and classifying electrical events on the basis of voltage and current measurements obtained at the main circuit panel of the home. To address the necessary training and calibration, this approach is designed around the end-user and relies on user input to continuously improve its performance. The algorithms and the user interaction processes are described in detail. Three data sets were collected with a prototype system (from a power strip in a laboratory, a house, and an apartment unit) to test the performance of the algorithms. The event detector achieved true positive and false positive rates of 94 and 0.26%, respectively. When combined with the classification task, the overall accuracy (correctly detected and classified events) was 82%. The advantages and limitations of this work are discussed, and possible future research is presented.

Authors

  • Anthony Rowe
  • Mario Bergés
  • Gaurav Bhatia
  • Ethan Goldman
  • Raj Rajkumar
  • James H. Garrett
  • José M. F. Moura
  • Lucio Soibelman
Download PDF BibTeX

Sensor Andrew: Large-Scale Campus-Wide Sensing and Actuation

Abstract

In this paper, we present Sensor Andrew, an infrastructure for Internet-scale sensing and actuation across a wide range of heterogeneous devices designed to facilitate application development. The goal of Sensor Andrew is to enable a variety of ubiquitous large-scale monitoring and control applications in a way that is extensible, easy to use, and secure while maintaining privacy. To illustrate the requirements of Sensor Andrew, as well as the capabilities and limitations of the system, we outline one such application in which multiple classes of energy sensors are combined with environmental sensors to not only monitor energy usage but also identify energy waste within buildings.

Authors

  • Saurabh Taneja
  • Burcu Akinci
  • James H. Garrett
  • Lucio Soibelman
  • Mario Bergés
  • Guzide Atasoy
  • Pine
  • S. Mohsen Shahandashi
  • Engin Burak Anil
  • Essin Ergen
  • Anu Pradhan
  • Pingbo Tang
Download PDF BibTeX

CEC: Sensing and Field Data Capture for Construction and Facility Operations

Abstract

Collection of accurate, complete, and reliable field data is not only essential for active management of construction projects involving various tasks, such as material tracking, progress monitoring, and quality assurance, but also for facility and infrastructure management during the service lives of facilities and infrastructure systems. Limitations of current manual data collection approaches in terms of speed, completeness, and accuracy render these approaches ineffective for decision support in highly dynamic environments, such as construction and facility operations. Hence, a need exists to leverage the advancements in automated field data capture technologies to support decisions during construction and facility operations. These technologies can be used not only for acquiring data about the various operations being carried out at construction and facility sites but also for gathering information about the context surrounding these operations and monitoring the workflow of activities during these operations. With this, it is possible for project and facility managers to better understand the effect of environmental conditions on construction and facility operations and also to identify inefficient processes in these operations. This paper presents an overview of the various applications of automated field data capture technologies in construction and facility fieldwork. These technologies include image capture technologies, such as laser scanners and video cameras; automated identification technologies, such as barcodes and Radio Frequency Identification (RFID) tags; tracking technologies, such as Global Positioning System (GPS) and wireless local area network (LAN); and process monitoring technologies, such as on-board instruments (OBI). The authors observe that although applications exist for capturing construction and facility fieldwork data, these technologies have been underutilized for capturing the context at the fieldwork sites as well as for monitoring the workflow of construction and facility operations.

2010

Authors

Download PDF BibTeX

Contactless sensing of appliance state transitions through variations in electromagnetic fields

Abstract

Non-Intrusive Load Monitoring (NILM) is a promising technique for disaggregating per-appliance energy consumption in buildings from aggregate voltage/current measurements. One major limitation of the approach is that it typically requires a training phase during which users must manually label device transitions. In this paper, we present an inexpensive contactless electromagnetic field (EMF) event-detector that can detect appliance state changes within close proximity based on magnetic and electric field fluctuations. Each detector wirelessly transmits state changes to a circuit-panel energy meter, which can then be used to label and disambiguate appliance transitions detected from the aggregate signals as well as to track the associated energy consumption. Our EMF sensors are able to detect significant power state changes from a few inches away making it possible to externally monitor in-wall wiring to devices (e.g., overhead lights). We experimentally evaluate our proposed EMF sensor in terms of power consumption, accuracy and detection range on a variety of appliances to demonstrate its effectiveness towards augmenting NILM systems. We show that accurately detecting 100W loads from 10cm away is possible while maintaining multiple-year battery life from a coin-cell battery.

Authors

Download PDF BibTeX

Enhancing Electricity Audits in Residential Buildings with Nonintrusive Load Monitoring

Abstract

Non-Intrusive Load Monitoring (NILM) is a technique for deducing the power consumption and operational schedule of individual loads in a building from measurements of the overall voltage and current feeding it, using information and communication technologies (ICT). In this paper we review the potential of this technology to enhance residential electricity audits. First, we review the current commercially available whole-house and plug-level technology for residential electricity monitoring in the context of supporting audits. We then contrast this with NILM and show the advantages and disadvantages of the approach by discussing results from a prototype system installed in an apartment unit. Recommendations for improving the technology to meet the demands of residential audits are provided, along with ideas for possible future work in the field.

Authors

Download PDF BibTeX

Leveraging Data From Environmental Sensors to Enhance Electrical Load Disaggregation Algorithms

Abstract

The idea of sustainable or green buildings generally stops after the design and construction phases.
Little effort is made to continuously monitor and control the energy profile throughout the life-cycle
of these facilities. To effectively identify opportunities for consumption reduction, measurement and
feedback of current energy use is necessary. Monthly utility bills are inadequate for planning
conservation programs, or even for assessing their effectiveness once implemented. Extensive
hardware sub-metering, although very expensive, is sometimes used to obtain more granular
feedback. Non-Intrusive Load Monitoring (NILM), another method that has been studied for the past
two decades, follows an inexpensive approach for obtaining appliance-specific consumption
information. The idea behind this technique is that operation of individual appliances generates a
distinct signature in the power distribution system of the building, which can be detected by carefully
analyzing the overall voltage and current of the building. However, two of the main challenges
keeping the technology from reaching wide adoption are: (a) finding simple ways to train the
algorithms; and (b) obtaining robust appliance signatures that form spread-out clusters in the feature
space, especially for small loads.

In this paper we explore the feasibility of utilizing data from separate environmental sensors (e.g.,
light intensity, sound level, etc.) present in the building, for improving the training process by
enhancing the appliance signatures and providing an independent and trusted source of information
about the operation of appliances. We exploit the fact that the operation of appliances will likely be
reflected in both the power and environmental data streams. We present initial results from a case
study where a prototype NILM system was deployed in an occupied apartment building, along with a
number of environmental sensors. We also suggest two approaches for leveraging the environmental
data and provide descriptions for possible future research in the area.

Authors

Download PDF BibTeX

Evaluating the Electric Consumption of Residential Buildings: Current Practices and Future Prospects

Abstract

As the construction industry transitions towards green buildings, and the number of LEED certified facilities continues to increase, the question of how to evaluate as-built energy performance becomes more important. Most homeowners rely on a monthly bill to determine their electric consumption, which is not an effective way to understand the results of most energy-saving strategies. Real-time feedback and appliance-level information is necessary, but most solutions require extensive hardware sub-metering, with a high price due to the hardware and installation costs. We argue that, in order to achieve wide adoption, the solutions need to be simple, easy to install, inexpensive and be able to return the investment in a reasonable time. In this paper we first analyze and compare the different types of technologies that are currently available for allowing homeowners to monitor their energy expenditure. Then we discuss new approaches that balance the trade-off between information and cost, and present preliminary results from a prototype Non-Intrusive Load Monitoring (NILM) system we have installed in a building.