PDS Lab - Sample Research Works

Identification of disturbance sources in power systems
When an electricity customer such as an industrial facility experiences power disturbances, it is very difficult to know where the disturbances originate. An example is the telephone interference caused by various electrical noise-producing appliances (called harmonic sources). The objective of this research is to develop a platform and associated algorithms for on-line tracking of the harmonic sources in a power system, using a multi-point metering/sensor network. Although the project has a well-defined application, the research subject itself is quite general. It deals with a class of problems called “cause identification”. Here we use the term “cause identification” to describe methodology that identifies the causes underlying abnormal observations recorded by a sensor network. In this project, we intend to research two potential ideas for the cause-identification problem. One is statistical inference. The basic idea is to use disturbance signatures as indices for segmenting data and then apply statistical inference analysis on the segmented data. After more than one year of intensive effort, we have established a statistical inference technique for harmonic source identification. The technique has been verified by computer simulation, experimental results, and some field data. We are currently developing a user-friendly computer program so that the project collaborators can test the tool on actual problems. This program development is to be completed by September 2009. The research findings also open up two new directions: (1) identification of sources of other types of power quality disturbances and (2) implementation to the smart meters as a value-added feature. The second potential idea we intend to investigate is the principal component analysis.

Measurement of advanced power system parameters
The objective of this research project is to make individual sensors “measure” advanced parameters of power systems unattainable in the past. We propose to use disturbance responses recorded at substation meters to perform continuous estimation of the Thevenin impedance of the upstream system and the load-to-voltage dependency characteristics of the downstream system. These parameters are of significant value to Albertan and other power companies. In this research, a disturbance recorder is placed at the sending end of each feeder. A disturbance from downstream can be used to estimate parameters of the upstream system. Conversely, a disturbance from upstream can be used to estimate the downstream feeder characteristics. After more than one year of intensive effort, we have developed a method to estimate the upstream system parameters. The method has been tested successfully using measurements taken from three substations in Alberta. We also conceived a practical and simple method to measure the load parameters.

Power electronic signaling technologies
When considering the application of power electronics to power systems, we often think about the FACTS type devices for power flow control and stability improvement. We believe there is a parallel and equally important class of applications for power electronics. It is the creation of controllable disturbances for communication, protection and other signaling-oriented applications. We call this power electronic signaling technology. One of our research findings in this direction has resulted in a license agreement with an Alberta company. The company is currently commercializing this patented technology (EnerPulsar). We are currently collaborating with the company to prepare a field-demonstration project in Ontario in the Hydro One system. Additional technology in this direction, a device to detect a broken neutral conductor, has attracted the interest of Manitoba Hydro in addition to that of the consortium companies. Manitoba Hydro just approved a research grant to support the development of a prototype device for testing in the Manitoba system. We are also making progress on a third technology in this direction: active detection of human contact on downed conductors. Such conductors must not be energized as it may result in human fatality. Our idea is to use a power electronic device to generate controlled-voltage pulses to detect human contact with unenergized lines.

Techniques to model industry facilities for power system stability assessment
Large electric-power-consuming facilities such as oil refineries, steel mills and shopping complexes are major loads to a power transmission or distribution system. For computer-simulation-based power system planning and operation studies, it is critical to model such loads properly due to their large power demands and complex responses to power disturbances. Inadequate models may result in a poorly designed power-supply system or costly system-investment decisions. However, it is a great challenge to establish adequate models for large facilities. For power-system-planning studies, basic information such as circuit diagrams, load composition and motor sizes etc. of the future industrial or commercial facilities are often not available. So, it is extremely difficult to develop models of future facilities for long-range planning studies. In this project, we have developed a technique that can create models for various facilities with minimal user input. Due to the interest of a power-system software vendor, a provisional patent has been filed for this technique in April 2009.

Power quality impact of distributed harmonic sources
The characteristics of electric loads such as consumer electronics and energy efficient appliances have changed significantly in recent years. Such loads draw pulse-like currents from the supply systems and cause distortions to the sinusoidal voltage waveforms. These distortions are called voltage or current harmonics and are a significant power quality concern. Due to the increased awareness of energy conservation and environmental protection, mass distributed-harmonics-producing loads (called harmonic sources) have proliferated in current power distribution systems. One clear indication of this is the mandate to use energy efficient (but harmonics-producing) lighting devices. Plug-in hybrid vehicles could become the next major distributed-harmonic source. Unfortunately, research results obtained for large, centralized harmonic sources over the past two decades have been of limited value in analyzing such situations. In view of the significant interest by industry, this chair program is mounting a major effort in researching techniques for modeling, analyzing, measuring, and mitigating harmonics for systems with distributed harmonic sources. Progress made in this direction involves three projects so far, including (1) the modeling and analysis of the harmonic-producing characteristics of energy efficient appliances, (2) field measurement to understand the collective harmonic impact of multiple residential houses, and (3) proposition of a novel adaptive-harmonic-filtering device. The field measurement work is one of the project highlights in this reporting period. Supported by Epcor and Telus, our research team conducted a five-week extensive field measurement in a residential distribution line and collected close to 1TB field test. We are currently analyzing the data for the purpose of establishing a theory explaining how distributed harmonic sources affect the power quality of a distribution system. The theory will eventually be used to predict the harmonic pollution levels of a power system with distributed harmonic sources.

Management of commercial and residential loads
Commercial buildings, small industry plants and residential loads are considered as “dumb” loads by power system engineers. With the “smart grid” technologies (i.e. information and communication technologies applied to power systems), the loads are no longer that passive and reactive to power system situations. They can actually participate in power system activities and create a mutual support to each other. The best example is the virtual power plants created through smart demand response technologies. In this research, we try to develop innovative techniques for these loads and facilitate their participation in power system activities. The focus are on consumption monitoring, energy conservation, demand response and electric safety.

A Novel Network Decoupling Transform for Power System Analysis
Due to the rapid advancement of measurement and telecommunication technologies, a large amount of data are available nowadays for power system monitoring and control. One example is the synchronized phasor data. Various research works have been conducted to develop applications for the phasor data. The challenges of extracting new and unique information from the phasor data may be partially due to the lack of a support theory for phasor data processing and interpretation. This situation may be understood by examining the use of three-phase voltage phasor data, Va, Vb and Vc. One can process the data in various ways. However, operations such as (Va+Vb+Vc)/3 (=Vzero-sequence) or (Va+a2Vb+aVc)/3 (=Vnegative-sequence) (where a=1Đ120o) have been recognized as the best means to analyze the data. The symmetrical components transform is the support theory for these operations. Because of the theory, converting abc phasors to 012 sequences has become a standard approach to process three-phase phasor data and a number of monitoring and protection schemes have been developed. The wide-area monitoring systems nowadays have made multi-location (positive sequence) voltage phasor data, V1, V2, V3 … Vn available. Inspired by the success of the symmetrical components theory, one would wonder if operations such as T1V1+T2V2+…+TnVn can be rigorously derived for the multi-location (i.e. multi-bus) phasor data, and if such operations can reveal unique characteristics of a power system. This work shows that such a transform can be derived. The transform is able to extract important information about a power system from the phasor data. The transform is conceived from the following observation: a power network can be represented as a multi-node, multi-branch Thevenin circuit connecting the loads to the generators. If one applies eigen-decomposition on the Thevenin impedance matrix, the network can be decoupled into a set of single-node, single-branch equivalent circuits. These circuits are much easier to analyze and they carry valuable information of a power system. Similarly, if the variations of the transformed variables can be evaluated, one may be able to predict the complex behaviors of the actual network.