Phasor Measurement Units - From Exotic to Everyday

by Gary Roskos, Electric Utility Industry Open Systems International, Inc. and Bill Robertson, Senior Principal Engineer, Salt River Project

Over the past ten years, direct Phasor Measurement technology has progressed from research and development to commercial availability. The development of open standards for measurement and communication has provided a solid foundation for device manufacturers. In addition, the availability of inexpensive and reliable measurement devices, secure communications and synchronous time sources have removed most of the barriers to widespread implementation of this important technology.

However, everyday use of measured synchronous phasor data has remained largely limited to research or pilot installations or for historical analysis. This is largely due to the perception that a large roll-out would require large, concentrated investments in phasor measurement unit (PMU) hardware, communications systems and application software.

Salt River Project, Phoenix AZ (SRP) is working with Open Systems International (OSI) and Schweitzer Engineering Laboratories (SEL) to show that PMU technology can be implemented in a logical, simple and effective manner without significant costs. In addition to gathering and storing data for historical analysis, SRP is using PMU measurement data to improve the performance and results of existing real-time transmission applications, particularly State Estimation and Contingency Analysis.

The goal is to effectively and efficiently incorporate PMU technology to improve everyday operations.

The promise of PMU data is that the actual state of the transmission grid can be measured in detail as fast as the data can be communicated. Currently, the determination of an estimated state uses scalar measurements taken over several seconds and the state-model processing times that are often in the order of tens of seconds. Reducing the time to determine an accurate system state could allow for accurate assessments of dynamic system conditions resulting in automatic stability control feedback and more useful presentation of system state and condition information to human operators.

Other benefits include the realization of a measured system connection and impedance model rather than using a calculated system model based on nominal construction and environmental assumptions. This detailed state model can be utilized in background analysis to determine system security and stability margins with a high degree of confidence.

At SRP, the short term goal is to utilize PMU data in an everyday manner to improve the quality and speed of the State Estimation process.

Field Installation Approach
SRP has developed a very simple approach to implementation of PMU technology. In 2006, as a result of the availability of PMU devices (a common feature in newer standard protective relays) a group was formed to plan and prioritize implementation of phasor measurement devices. Priorities were established by considering locations that were important for the utility and its neighbors (control area boundaries, generation sites, etc.), and coordinating installations with existing plans for maintenance and construction. Using this strategy, they are able to methodically add to the installed base of phasor measurement devices each year and now have PMUs installed on approximately 10% of their transmission system buses.

SRP also took advantage of concurrent projects to improve substation communications and automation, allowing them to immediately use secure and reliable, high-speed communications channels for the collection of PMU data.  This benefit was a direct result of a successful philosophy in the application of automation investments. By choosing solutions for automation that utilize open architectures and “off-the-shelf” components, SRP is able to cost-effectively meet specific current project needs while establishing basic infrastructure suitable for a wide variety of future needs. Making choices that take into account both current requirements and future capability allows the utility to roll out automation incrementally, building on the foundations laid by previous projects.

A small group meets regularly to maintain the implementation plan, coordinate installation between departments and to help promote the PMU program within the company.

Central Operations Implementation
Initially, PMU data was stored in a local historian and forwarded to the Bonneville Power Administration (BPA), another federal bulk power producer, like SRP, serving the Pacific Northwest.

SRP updated their system control center in 2007 and in 2008 a project was begun to utilize PMU data in their real-time systems alongside traditional data sources.

This project includes a new interface to the real-time system database that can collect streaming PMU data. This interface includes a clear and simple way to add devices and map the data to the real-time database, taking advantage of the self-describing features of the IEEE C37.118 protocol. This new application creates a time-aligned vector of input points from the incoming data stream which is written to the real-time database in a single write operation.

SRP is then able to utilize features available within their new OSI State Estimator that allows use measured phase angle data along with traditional sources. It was important that the advantages of this new data were realized even without a full system deployment and that changes in performance could be measured over time as more PMUs were installed.

A diagram of the SRP PMU data collection system is shown in Figure 1.

As shown, the data is concentrated and presented to various systems in a common protocol format. One path leads to the SRP data historian, one leads to a BPA data collection system, and the third path leads to the real-time database interface application.

To evaluate the effects of the new data and to verify deployment strategies, SRP is running two State Estimation processes in parallel.

The first is their legacy process without any synchronized phasor data input. The second process substitutes PMU data for traditional measurements where it is available. See Figure 2.

With this configuration, a side-by-side comparison is available where solution results, estimation errors, and solution times can be compared.

In addition to providing a simple interface to the PMUs utilizing the standard IEEE 37.118 protocol, the OSI PMU application time-aligns incoming data and provides a deterministic write time to the real-time database so that other applications can avoid reading the data while it is being written. This application is very scalable, handling existing requirements at a reasonable cost with provisions for long-term future needs.

Within the State Estimation application, magnitude and angle data is used as input along with traditional non-synchronized scalar measurements and a level of confidence is assigned to each measurement. Where magnitude and actual angle measurements are available contiguously, the observability of the system is increased and the State Estimator solution is improved.

If fully deployed, the State Estimation process reduces to a state measurement process removing the need for a traditional Weighted Least Squares or Similar Givens Rotation algorithm. Power flows can be calculated in one step and branch impedances could be further refined based on accurate real-time phase measurements.

Central Operations Project Schedule
The project began in the second quarter of 2008 with the development of the PMU communications interface. It also included development of deterministic methods for handling specific-time data within the Energy Management System. In October, OSI installed the software on a parallel production system, the SRP “Development and Quality Assurance (QA) system.” The system will be fully functional before the end of 2008.

These development and deployment initiatives at OSI and SRP were generously supported by Schweitzer Engineering Laboratories, which provided development hardware advice.

PMU Deployment Strategies
As SRP evaluated deployment priorities, they worked with their partners to determine simple strategies for prioritizing future deployments. These ideas, along with their current planning considerations, help SRP realize the benefits of gradual installations more quickly.

In making these analyses, some questions to consider include:

• What data is needed by regional organizations and my neighbors?
In many cases, Regional Operating, Planning and Reliability organizations require that PMU data be collected and stored at critical system locations such as control area boundaries, generation sites, etc.

• Where do I have poor system observability or the high estimation/measurement errors?
By reviewing error results for State Estimation solutions at different times, locations can be identified where additional high quality data will have the most impact on the quality of the solution results.

• Where do I have elements with varying impedance models?

  • LTC Transformers or Regulators?
  • Phase Shifters?
  • Varying or Switchable VAR Sources?

Analysis of these elements in the State Estimation application involves an inherently inaccurate impedance model or relies on more complexity in modeling and in the estimation algorithm. The benefit is found in the simplification of these processes, which contributes to solution accuracy and speed.

Once these early considerations are satisfied, it is best to deploy in a growing contiguous pattern, segregating the system into “calculated” and “estimated“ areas, rather than using a “shot gun” installation approach. This results in an ever-smaller system that still requires an estimation process.

A complementary strategy is to begin at the highest voltage levels. For example, a measured system model can be established for the 500kV system while the rest of the system is estimated. Then, as one deploys to lower system voltages, the size of the calculated system is increased and the portion of the system that requires estimation is decreased.

Other Data Collection Strategies
PMUs can be effectively deployed even without a dedicated communications channel. Many PMU devices allow the user to select sample rates slow enough to allow for polling from a central site as part of a standard SCADA data stream. It is also important that the protocol used supports the measurement time stamp. Although the sample and collection rate of this data is no better than traditional data, the samples can be coordinated to coincide with samples taken anywhere in the system, and given the sample time, the data can be properly aligned with other data available from other sources. This method will not increase the cycle time of the state solution, but will contribute to the quality of the solution.

The Future
As PMU technology is fully deployed and as communications latencies are reduced, accurate system state models can be determined very rapidly. This will lead to advances in distributed real-time dynamic stability control applications. Using the PMU derived state of the power system in faster time domain than a few seconds, would allow detection of unfeasible operating conditions, and would lead to development of control algorithms for isolation and islanding of the network components to preserve the integrity of the overall Grid.

In addition, problems with development of accurate large system models will be greatly reduced through the ability to directly measure and refine the models in real-time. These models can be three-phase or phase-independent models as both sequence and phase models are measured.

At the same time, advances in contingency and other dynamic analysis applications are allowing systems to analyze many more “what if” scenarios in a considerably reduced time-frame and visual presentation applications are being developed to enhance operator situational awareness. Using the PMU generated data as RADAR sweep for the operator to be able to operate the power system and delineate any impending dynamic problems in the Grid.

The coincidence of improved information and application tools is leading quickly toward the promise of enhanced reliability, security and survivability of the grid. Smart Grid technologies are on the horizon which will use such advances to create truly self-healing networks, true fast identification of impending problems and isolation of sub-networks and re-routing of power through alternate paths.

With clear goals, thoughtful and creative planning and with focused execution, these benefits can be realized successfully with little fanfare and without exorbitant investments today.

About the Authors
Gary Roskos has over twenty five years of direct engineering experience in the Electric Utility industry. Roskos is an industry expert with a wealth of practical knowledge regarding grid design, operations, maintenance, protection and automation. He specializes in advancing fields such as next-generation Remote Telemetry, Substation and Distribution Automation, Outage Management, Metering and Smart Grid Design. Mr. Roskos has a Masters Degree in Electrical Engineering from the University of Minnesota and is a member of IEEE and NSPE.

Bill Robertson is a Senior Principal Engineer with Salt River Project. He supports Network Applications functions including State Estimator within the Energy Management System. Robertson received a MS degree in Electrical Engineering from New Mexico State University in 1987. He joined SRP in 1987 and has worked in support of Network Applications and Energy/ Transaction Scheduling.

Most consulted news