Addressing Emerging Network Management Needs with Enhanced ...

2 downloads 0 Views 1MB Size Report
with Enhanced WAMS in the GB VISOR Project. S. Clark, D. Wilson,. N. Al- ..... higher reporting rate of 200fps is fully compliant with the. C37.118 standard, and is ...
Addressing Emerging Network Management Needs with Enhanced WAMS in the GB VISOR Project S. Clark, D. Wilson, N. Al-Ashwal

F. Macleod, P. Mohapatra, J. Yu

P. Wall, P. Dattaray, V. Terzija

GE’s Grid Solutions Edinburgh, UK [email protected]

SP Energy Networks Blantyre, UK

University of Manchester Manchester, UK

Abstract—This paper describes a number of novel Wide-Area Monitoring System (WAMS) capabilities and analysis tools, including a new monitoring solution that covers the 4-46Hz subsynchronous oscillation range, and the deployment of a new oscillation source location algorithm. These are being demonstrated under VISOR – an innovation project that has created the first integrated GB WAMS in order to demonstrate the capabilities and tangible business benefits of WAMS to the GB system. Other work from the project is discussed including hybrid state estimation, dynamic model validation, line parameter estimation and a new boundary constraint approach that uses angle difference in addition to power. Finally, selected learning from a review of existing GB WAMS performance is presented, along with details of next steps in the project. Index Terms--governor mode, hybrid state estimation, line parameter estimation, sub-synchronous oscillation, wide-area monitoring system (WAMS).

I. INTRODUCTION The GB (Great Britain) power system is undergoing a period of major change – with the addition of series compensation and intra-network HVDC transmission, the increasing penetration of renewable generation, and the closure of large synchronous generation plant. These changes alter existing dynamic behavior and introduce new issues – raising the need for improved monitoring and understanding of system behavior to support reliable and efficient operation of the GB grid. The VISOR project [1] was instigated to address this need – involving all three GB Transmission Owners, the GB System Operator, academia and vendors. The objective of VISOR is to demonstrate, using the first integrated GB Wide-Area Monitoring System (WAMS), a suite of new wide-area monitoring capabilities and analysis tools including novel algorithms – with a focus on management and early warning of system risks, and on relieving transmission constraints through increased situational awareness and operator confidence. There is strong emphasis on clear and concise presentation of real time decision-making information to operators, and on valuable long-term analysis. VISOR will showcase the tangible benefits from these tools, taking account of the operational processes that will also be The VISOR Project is a GB innovation demonstration project funded by the GB electricity consumer through the Network Innovation Competition mechanism, awarded by Ofgem – the GB regulator.

P. Ashton, M. Osborne National Grid UK Wokingham, UK

required, and will generate learning on infrastructure requirements, issues and implementation. This learning will feed future work as wide area monitoring, protection and control play a bigger role in GB grid operation [2]. The VISOR applications include: •

Wideband Oscillation Monitoring and Source Location (0.002–46Hz)



Area-Angle Based Constraint Definition for the Scotland-England Transmission Boundary



Validation of Dynamic Models



Real-time and off-line disturbance reporting



Line Parameter Estimation



Hybrid State Estimation

Related study work is also being undertaken as part of VISOR. This includes annual System Performance Reports covering dynamic behavior of the GB power system and performance of the WAMS itself, laboratory performance testing of WAMS infrastructure and applications, and assessment of the impact of uncertainty on system security margins. II. OSCILLATION MONITORING AND SOURCE LOCATION Monitoring of “low frequency” (LF) electromechanical oscillations, typically 0.1–2Hz, using synchronized measurements from Phasor Measurement Units (PMUs) is already established in GB and elsewhere [3]. The reducing and variable inertia of the GB system, together with the recent and imminent closure of large synchronous generation units with a power system stabilizing role, the deployment of series compensation and the proliferation of Power Electronic systems, necessitate extended and enhanced monitoring of power system oscillations in GB. In particular, there is need for: •

New tools to aid in identifying the sources of oscillations – in both real time and study domains.



Renewed focus on oscillations in the 0.002-0.1Hz governor or “Very Low Frequency” (VLF) range, driven by changing system inertia.



New monitoring of the 4-46Hz range, commonly termed “sub-synchronous oscillations” (SSO) – though to be correct the VLF and LF ranges are also sub-synchronous. This is motivated by the risk of interaction between new series compensation, power electronic controls, and generator shafts.

VISOR is demonstrating new WAMS analyses, applications and infrastructure to meet these needs. A number of common principles are applied, in line with the VISOR goal of providing clear and actionable information to operators, and valuable and usable data for long-term study: •

Analyses are performed in real-time, with results stored for later study e.g. model validation, baselining or operational review. Live and historical values can be accessed within the application, and can be exported for further analysis in third party tools as comma separated value (CSV) files or via a database driver.



Results are grouped into configurable frequency bands for independent alarming and visualization, enabling focus on modes of interest such as generator torsional frequencies. The analysis and results stored are unaffected, so as not to influence historical review.



Mode selector tabs allow users to switch between bands in charts and map views, and display the present dominant frequency and alarm state of each band.



Alarm and source location information is shown in map views through simple “traffic light” or color gradients respectively. Charts below display mode frequency, damping and amplitude for the user-selected location, alongside the worst-case system-wide value.

A. Oscillation Source Location (0. 002-4Hz) Whilst monitoring of electromechanical oscillations in using WAMS is well-established, identification of which generator(s) are contributing to an oscillation is less so. Currently, operational response to an oscillation will either require prior study of the mode, or else apply general guidance such as reducing power flow between participating regions. Both will usually impact the economic and efficient operation of the grid. Targeting action directly at the source ensures effectiveness whilst minimizing the impact on grid operation.

Other methods [5] use statistical analyses to find correlation between oscillation behavior and system conditions e.g. generator output. However this is most effective when applied over multiple events rather than in real-time. There is also a risk of highlighting correlative but noncausal links. 2) New VISOR Source Location Approach A new approach has been developed which uses PMU voltage angle measurements to locate an oscillation source to the nearest monitored bus(es) - allowing useful location information to be derived from sparse measurements. This is particularly valuable in large interconnected systems, since only a few voltage and frequency measurements, considered commercially nonsensitive, need be shared between operators. Contributions are identified through the relative phase of oscillations in voltage angle. A location with leading mode phase indicates a “source” contributing to poorer damping, whilst a lagging phase indicates a “sink” providing a more positive damping contribution. In the case of opposing-phase oscillations e.g. an inter-area mode, the participants must be first split into two opposing groups. The average phase of each group is determined, and individual phase differences or “contributions” derived for each signal relative to its group average. If one group leads the other by