Wide area monitoring system and its application on power system low-frequency oscillation suppression

Abstract

As modern electric power systems are transforming into smart grids, real-time wide area monitoring system (WAMS) has become an essential tool for power system operation and control. Though many WAMS based stability analysis and control schemes have been proposed in recent years, their practicality is often limited since important WAMS characteristics such as synchronization accuracy and signal time delay are often not considered in the design. Now, there is an urgent and necessary need for a comprehensive, systematic and quantitative analysis of WAMS characteristics. This thesis not only embarks on the analysis of WAMS characteristics but also investigates their influence on the power system stability analysis and control. Synchronization accuracy and signal time delay are two of the inherent WAMS characteristics which have to be considered in any WAMS based real-time applications. The control effectiveness of a well-designed wide-area controller which performs well in an ideal system could be badly deteriorated once these two WAMS characteristics are taken into account. In this thesis, synchronization accuracy and signal time delay are systematically studied and quantitatively calculated for the first time. Recommendations have been made to improve the synchronization accuracy and reduce the signal time delay of synchronized phasors supplied by WAMS. With the increasing applications of WAMS for on-line stability analysis and control in smart grids, it is also imperative to quantitatively evaluate the reliability of WAMS so as to identify the critical components for ensuring the secure and reliable operation of a smart grid. This thesis proposes a comprehensive reliability evaluation scheme for WAMS and its components based on Monte Carlo fault tree (MCFT) analysis. WAMS is a complex system consisting of many component devices and special communication networks which can be divided to five functional sub-systems for PMUs, PDC, local and wide area communication networks, and control center. A reliability model for each sub-system will first be constructed using the fault tree (FT) modeling method and then be analyzed using the Monte Carlo simulation approach to evaluate a set of reliability indices. Lastly, the FT modeling method will be applied again to construct the reliability model of WAMS and evaluate its reliability using the sub-system results. The validity and advantages of this MCFT reliability evaluation method applied on WAMS have been verified with simulation and comparison studies. A simple example based on an adaptive wide-area damping control scheme has also been given to show the application of the proposed WAMS reliability evaluation method. The restructuring of the modern electric power systems has had profound effects on the operation of the power grid. Traditional control strategies could become ineffective under the new system structure and new control strategies would be needed as a result. As an important information platform in modern power grid, WAMS with different topology structures would have large impacts on the design of these new control strategies. In this thesis, an in-depth investigation on the relative merits of centralized and distributed WAMS has been conducted to find the most suitable WAMS topology for on-line monitoring and control of the Shandong power grid. Various aspects including investment, signal time delay, operation reliability and risk have been systematically analyzed and compared for the centralized and distributed WAMS. An improved Minimum Spanning Tree (MST) algorithm has been proposed for the construction of communication networks with minimum investment in both centralized and distributed WAMS of Shandong power grid. Though the investment needed for a centralized or distributed WAMS in Shandong power grid is almost the same, the distributed WAMS has shorter signal time delay, higher reliability, and lower risk than the centralized one. Last but not least, a novel WAMS based adaptive wide-area low frequency oscillation damping control scheme has been proposed as a rational and logical application of the WAMS because of its relative long operation time frame. Practical issues including signal transmission time delay, changes in system operation conditions and uncertainties in system configuration and parameters have been fully considered in the design to ensure the damping effectiveness and the practicality of the control scheme. The core control algorithm consists of stochastic subspace identification (SSI), wide-area input signal formulation, and signal time delay compensation. Frequency deviation in each generator will be measured and collected via the WAMS to identify the low frequency oscillation modes using SSI and generate a wide-area control signal for each oscillation mode. The superiority of SSI over other identification methods lies in its ability to identify the oscillation modes collectively from all the measured signals instead of individual ones and high resistance to measurement noise. For each identified mode, the corresponding generator cluster will be identified to produce a wide-area control signal to combine with the local input signal as the new input signal of PSSs installed in the generators participating in this oscillation mode. Signal time delays are measured and compensated locally at the generators using GPS time services and adaptive time delay compensators. The effectiveness and robustness of the proposed adaptive wide-area damping control scheme have been verified with simulation studies on the IEEE 4-generator 2-area and IEEE 16-generator 5-area test systems under a number of disturbance scenarios.