John S. Donnal

Noncontact Power Meter

Energy metering is increasingly important in todays power grid. With real-time power meters, utilities can efficiently incorporate renewables and consumers can tailor their demand accordingly. Several high-profile attempts at delivering realtime energy analytics to users, including Google Power Meter and Microsoft Hohm, have essentially failed because of a lack of sufficient richness and access to data at adequate bandwidth for reasonable cost. High performance meters can provide adequate data, but require custom installation at prohibitive expense, e.g., requiring an electrician for installation. This paper presents hardware and signal processing algorithms that enable high bandwidth measurements of voltage, current, harmonics, and power on an aggregate electrical service (such as a residential powerline) for nonintrusive analysis with hardware that requires no special skill or safety considerations for installation.

NilmDB: The Non-Intrusive Load Monitor Database

This paper presents NilmDB, a comprehensive framework designed to solve the “big data” problem of non-intrusive load monitoring and diagnostics. It provides the central component of a flexible, distributed architecture for the storage, transfer, manipulation, and analysis of time-series data. NilmDB is network-transparent and facilitates remote viewing and management of large data sets by utilizing efficient data reduction and indexing techniques.

VAMPIRE: A magnetically self-powered sensor node capable of wireless transmission

This paper presents a power electronic topology for providing a vibration monitor with in-situ magnetic energy harvesting. The energy harvesting circuit uses low voltage MOSFETs and a nano-crystalline magnetic core to extract energy from an operating electrical load like a motor. This topology can be used to power sensors and wireless circuitry for diagnostics with no need for battery or special power wiring. Harvesting energy from a 90 W-60 Hz load, a Vibration Assessment Monitoring Point with Integrated Recovery of Energy (VAMPIRE) produces 7.5mW with the core volume of 2.9 cm3, and powers up a 6.5mW electrical monitor that provides sampling, signal processing, and periodic wireless data transmission through an RF channel. The system samples the vibration data at 125 Hz, and enables wireless burst transfer every 160 ms.

Energy Applications for an Energy Box

Changes in the electric utility will necessitate new needs and opportunities for monitoring and controlling electric power consumption and generation. Technical solutions exploiting these opportunities and answering these needs would ideally preserve best practices like reliability, privacy, efficiency, and flexibility. A nonintrusive load monitor (NILM) can serve as an ideal platform for constructing an “energy box” capable of sophisticated monitoring and control. This paper introduces a data processing and analysis framework, NILM manager. NILM manager creates a business model for handling power data by minimizing network bandwidth and placing intelligence and feature expansion in easily transmitted “energy apps.”

Hunting Cyclic Energy Wasters

Smart grid and smart meter initiatives seek to enable energy providers and consumers to intelligently manage their energy needs through real-time monitoring, analysis, and control. An essential element of intelligent management and control is access to information. The nonintrusive load monitor (NILM) identifies the operation of electrical loads from an aggregate service, making installation inexpensive, and speeding data collation and analytics. Smart meters are likely, in the near future, to be tasked with finding energy waste without requiring unreasonable demands for communication bandwidth. This paper presents NilmManager and NilmDB; tools for finding electromechanical energy wasters with a minimum of network bandwidth.

Energy Accountability Using Nonintrusive Load Monitoring

Conventional power meters measure total kilowatt-hours yet reveal little about how power was used. Modern solid-state metering solutions are not necessarily taking full advantage of the inexpensive but high performance computation capability that is available. This paper contains the details of a field test of a new software architecture for nonintrusive utility monitoring that endeavors to solve the big data problem of handling interesting but high bandwidth data streams from many monitored sites, e.g., residences and commercial buildings. Results from a field test are used to illustrate the utility of this system.

Current and Voltage Reconstruction From Non-Contact Field Measurements

Non-contact electromagnetic field sensors can monitor voltage and current in multiple-conductor cables from a distance. Knowledge of the cable and sensor geometry is generally required to determine the transformation that recovers voltages and currents from the sensed electromagnetic fields. This paper presents a new calibration technique that enables the use of non-contact sensors without the prior knowledge of conductor geometry. Calibration of the sensors is accomplished with a reference load or through observation of in situ loads.

Utilizing Spin-Down Transients for Vibration-Based Diagnostics of Resiliently Mounted Machines

This paper presents a vibration measurement and analysis technique for use during a machines spin-down procedure. During spin-down, the machines operation covers a continuous wide frequency band, from operating speed to standstill, which allows the estimation of the machines vibration transfer function (VTF). This transfer function is rich in information for detecting and differentiating not only machinery pathologies but also problems with vibrational mounts. Utilizing a back-electromotive force sensor to infer rotor speed and a single-axis accelerometer for vibration measurements, this technique allows minimally intrusive estimation of a machines VTF. Data collected in laboratory and field tests aboard U.S. Navy ships are presented to demonstrate the usefulness of this monitoring technique.

Water Nonintrusive Load Monitoring

Resource conservation decisions require detailed consumption information. This paper presents sensors and signal processing techniques that use pipe vibration signatures to non-intrusively identify water consumption at the appliance level. The method requires as little as one easily installed vibration sensor. This method provides a no-fuss retrofit solution for detecting the operation of a buildings water consuming appliances. In addition, flow rate is nonintrusively obtained from a conventional water meter via a new, high sensitivity strap-on magnetic sensor. Together, these two sensors track load operating schedule and water consumption in a building, demonstrated here at three different field test sites.

The Sinefit Spectral Envelope Preprocessor

This paper presents a new spectral envelope preprocessor based on sinusoid fitting and the discrete Fourier transform. This preprocessor is well-suited for nonintrusive condition monitoring and diagnostics due to its high noise resiliency and flexibility. It reduces data storage, transfer, and processing requirements by extracting only relevant harmonic signatures. This paper analyzes the resolution and accuracy benefits of spectral envelopes, including the effects of additive white Gaussian noise and presence of higher frequency spectral harmonics.