東華大學圖書館 |

Modeling and Prevention of Cascading Failures in Power Systems.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Modeling and Prevention of Cascading Failures in Power Systems./
作者:	Gharebaghi, Sina.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2023,
面頁冊數:	181 p.
附註:	Source: Dissertations Abstracts International, Volume: 85-05, Section: A.
Contained By:	Dissertations Abstracts International85-05A.
標題:	Communications networks. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30720598
ISBN:	9798380727709

Modeling and Prevention of Cascading Failures in Power Systems.
Gharebaghi, Sina.

Modeling and Prevention of Cascading Failures in Power Systems. - Ann Arbor : ProQuest Dissertations & Theses, 2023 - 181 p.

Source: Dissertations Abstracts International, Volume: 85-05, Section: A.

Thesis (Ph.D.)--The Pennsylvania State University, 2023.

This item must not be sold to any third party vendors.

Cascading failure in the power system is a low-probability-high-risk event. It has significantly negative impact on the economy and society in general. Therefore, it is necessary to carefully study this phenomenon, which can help prevent its occurrence. Precise modeling of cascading failure in the power system is a very complicated hard-to-solve problem. Broadly, there are three types of cascading failure models reported in literature - DC-quasi-steady-state (QSS), AC-QSS, and dynamic models. This dissertation is focused on AC-QSS and dynamic models.As the name suggests, QSS-type models study cascading failures at the 'snapshots' of of a sequence of pre-and post-disturbance steady state conditions. Using DC-QSS models is the most simple way of simulating cascading failure. Although these models are easy to implement and computationally inexpensive, they cannot capture phenomena including voltage stability/collapse and reactive power in the power system. AC-QSS models on the other hand can represent such issues during cascade propagation and possess higher accuracy compared to DC-QSS models - albeit at a higher computational cost.A challenging problem facing AC-QSS cascading failure models of power system is the divergence issue primarily stemming from voltage collapse phenomena. In reality, there are undervoltage load shedding (UVLS) relays, which aim to prevent such a collapse by shedding a pre-specified fraction of load at buses where the corresponding voltages fall below a threshold. However, capturing the UVLS action in QSS models is very difficult, because most of the time the model cannot generate an equilibrium below the voltage threshold due to divergence. To address this problem, current models have applied different variants of uniform load shedding (ULS) till convergence is achieved, which differ from the ground truth. In order to solve this, we propose a methodology that leverages the post-ULS load flow as a starting point when divergence occurs. In this condition, a sensitivity index coupled with the voltage magnitudes of buses is used to recognize the buses that are most prone to voltage collapse. The UVLS scheme is then applied to these buses. To verify the accuracy of the results, we also present a suitable dynamic cascade model with appropriate limits and protection details that can selectively capture UVLS action, thereby revealing the proximate ground truth. Predictions of the proposed AC-QSS model are validated against those of the dynamic model for representative cases in IEEE 118-bus system. In addition, results of the proposed model are contrasted with two ULS schemes on the 2, 383-bus Polish system.Next, we consolidate the proposed AC-QSS cascading failure model with a centralized AC optimal preventive control approach to alleviate cascade propagation. We use a simple AC-QSS model with ULS as benchmark to contrast results of UVLS model with preventive control and further demonstrate the effectiveness of preventive control on IEEE 118-bus system and a 2383-bus Polish network.It is worth noting that, the ground truth for cascading failure in power system is only achievable through a detailed dynamic model involving nonlinear differential and algebraic equations that need to be solved by computationally expensive numerical integration methods. This has prohibited adoption of such models for cascading simulation and lead to commonly used QSS models that are inaccurate, but allow statistical analyses.To solve this, we propose a fast cascading failure simulation approach based on implicit Backward Euler method (BEM) with stiff decay property. Unfortunately, BEM suffers from hyperstability issue in case of oscillatory instability and converges to the unstable equilibrium. We propose a predictor-corrector approach to fully address the hyperstability issue in BEM and we call the model as BEM-PC. The predictor identifies oscillatory instability based on eigendecomposition of the system matrix at the post-disturbance unstable equilibrium obtained as a byproduct of BEM. The corrector uses right eigenvectors to identify the group of machines participating in the unstable mode. This helps in applying appropriate protection schemes as in ground truth. We use Trapezoidal method (TM)-based simulation as the benchmark to validate the results of the proposed approach on the IEEE 118-bus network, 2, 383-bus Polish grid, and IEEE 68-bus system. The proposed approach is able to track the cascade path and replicate the end results of TM-based simulation with very high accuracy while reducing the average simulation time by {acute}{89}{88} 10 {acute}{88}{92} 35 fold. The proposed approach was also compared with the partitioned method, which led to similar conclusions.Next, we demonstrate that a further speedup can be achieved by a parallelized version of BEM-PC, which we call BEM-PC-parallel (BEM-PCP). In this version, the predictor subprocess of BEM-PC is run in multiple parallel processors for identification of oscillatory instability using eigen decomposition of the system matrix at post-disturbance unstable equilibria. Monte-Carlo studies on a 2, 383-bus Polish system confirm that BEM-PCP is on average 17% faster than BEM-PC and {acute}{89}{88} 40 times as fast as TM while maintaining the same accuracy as BEM-PC.Our next contribution is regarding the cascade mitigation in the context of dynamic models. To this end, we focus on the dynamic security assessment (DSA), which is a proactive strategy that helps avoid initiation of cascading failure. A comprehensive DSA entails that system operators are capable of inspecting hundreds of most probable contingencies every 15 {acute}{88}{92} 30 minutes using online state estimator data and system dynamic model. Ideally, one would get the most accurate sense of system security by running nonlinear time-domain simulations for {acute}{89}{88} 20 {acute}{88}{92} 30 s, which is computationally unmanageable for large-scale systems. This has led to application of Lyapunov-based direct methods, trajectory sensitivity-based approaches, and machine learning-based heuristics, all of which suffer from different drawbacks. To address this, we propose a fast detailed dynamic simulation approach based on BEM with stiff decay property. Unfortunately, BEM may converge to the unstable equilibrium in case of oscillatory instability in power systems. We propose an augmented BEM approach to solve this issue. It performs eigen decomposition on the post-disturbance system matrix of the linearized model obtained as a byproduct of BEM to detect oscillatory instability. Results from the IEEE 68-bus NE-NY network and 2, 383-bus Polish system confirm that the proposed method can maintain the accuracy of the traditional TM-based simulation while demonstrating {acute}{89}{88} 25 {acute}{88}{92} 31 times average speedup.Finally, this dissertation focuses on developing a comprehensive cyber-physical cascading failure model of power systems along with a cascade mitigation strategy. We propose a dynamic cascading failure model that considers realistic interdependencies between power and fiber-optic communication networks used for system monitoring and control in power grids. In this model, power line outages do not immediately disconnect communication links, whereas communication nodes have battery backup that starts depleting after considerable load shedding in the collocated bus or bus outage. When a communication node's battery is fully depleted, the node disconnects from the cyber layer, potentially reducing the observability and controllability of the power grid. A centralized optimal preventive controller (OPC) algorithm to minimize load shedding is proposed for cascade mitigation, which is applied selectively on fully observable and controllable islands. OPC considers AC power flow equations, multiple hard constraints, and treats overloading of lines as soft constraints. The results of Monte-Carlo simulations on the IEEE 118-bus and 2, 383-bus Polish systems demonstrate that the proposed OPC algorithm is effective in mitigating cascading failures. Finally, we demonstrate that our recently proposed BEM-PC can reduce the average simulation time by approximately 9 {acute}{88}{92} 26-folds compared to the TM with acceptable accuracy.{A0}

ISBN: 9798380727709Subjects--Topical Terms:

3564745
Communications networks.
Subjects--Index Terms:

Cascade mitigation

Modeling and Prevention of Cascading Failures in Power Systems.
LDR:09562nmm a2200397 4500 001 2395790
005 20240517105018.5
006 m o d
007 cr#unu||||||||
008 251215s2023 ||||||||||||||||| ||eng d
020 $a 9798380727709
035 $a (MiAaPQ)AAI30720598
035 $a (MiAaPQ)PennState_23366svg5765
035 $a AAI30720598
040 $a MiAaPQ $c MiAaPQ
100 1 $a Gharebaghi, Sina. $3 3765294
245 1 0 $a Modeling and Prevention of Cascading Failures in Power Systems.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2023
300 $a 181 p.
500 $a Source: Dissertations Abstracts International, Volume: 85-05, Section: A.
500 $a Advisor: Chaudhuri, Nilanjan Ray.
502 $a Thesis (Ph.D.)--The Pennsylvania State University, 2023.
506 $a This item must not be sold to any third party vendors.
520 $a Cascading failure in the power system is a low-probability-high-risk event. It has significantly negative impact on the economy and society in general. Therefore, it is necessary to carefully study this phenomenon, which can help prevent its occurrence. Precise modeling of cascading failure in the power system is a very complicated hard-to-solve problem. Broadly, there are three types of cascading failure models reported in literature - DC-quasi-steady-state (QSS), AC-QSS, and dynamic models. This dissertation is focused on AC-QSS and dynamic models.As the name suggests, QSS-type models study cascading failures at the 'snapshots' of of a sequence of pre-and post-disturbance steady state conditions. Using DC-QSS models is the most simple way of simulating cascading failure. Although these models are easy to implement and computationally inexpensive, they cannot capture phenomena including voltage stability/collapse and reactive power in the power system. AC-QSS models on the other hand can represent such issues during cascade propagation and possess higher accuracy compared to DC-QSS models - albeit at a higher computational cost.A challenging problem facing AC-QSS cascading failure models of power system is the divergence issue primarily stemming from voltage collapse phenomena. In reality, there are undervoltage load shedding (UVLS) relays, which aim to prevent such a collapse by shedding a pre-specified fraction of load at buses where the corresponding voltages fall below a threshold. However, capturing the UVLS action in QSS models is very difficult, because most of the time the model cannot generate an equilibrium below the voltage threshold due to divergence. To address this problem, current models have applied different variants of uniform load shedding (ULS) till convergence is achieved, which differ from the ground truth. In order to solve this, we propose a methodology that leverages the post-ULS load flow as a starting point when divergence occurs. In this condition, a sensitivity index coupled with the voltage magnitudes of buses is used to recognize the buses that are most prone to voltage collapse. The UVLS scheme is then applied to these buses. To verify the accuracy of the results, we also present a suitable dynamic cascade model with appropriate limits and protection details that can selectively capture UVLS action, thereby revealing the proximate ground truth. Predictions of the proposed AC-QSS model are validated against those of the dynamic model for representative cases in IEEE 118-bus system. In addition, results of the proposed model are contrasted with two ULS schemes on the 2, 383-bus Polish system.Next, we consolidate the proposed AC-QSS cascading failure model with a centralized AC optimal preventive control approach to alleviate cascade propagation. We use a simple AC-QSS model with ULS as benchmark to contrast results of UVLS model with preventive control and further demonstrate the effectiveness of preventive control on IEEE 118-bus system and a 2383-bus Polish network.It is worth noting that, the ground truth for cascading failure in power system is only achievable through a detailed dynamic model involving nonlinear differential and algebraic equations that need to be solved by computationally expensive numerical integration methods. This has prohibited adoption of such models for cascading simulation and lead to commonly used QSS models that are inaccurate, but allow statistical analyses.To solve this, we propose a fast cascading failure simulation approach based on implicit Backward Euler method (BEM) with stiff decay property. Unfortunately, BEM suffers from hyperstability issue in case of oscillatory instability and converges to the unstable equilibrium. We propose a predictor-corrector approach to fully address the hyperstability issue in BEM and we call the model as BEM-PC. The predictor identifies oscillatory instability based on eigendecomposition of the system matrix at the post-disturbance unstable equilibrium obtained as a byproduct of BEM. The corrector uses right eigenvectors to identify the group of machines participating in the unstable mode. This helps in applying appropriate protection schemes as in ground truth. We use Trapezoidal method (TM)-based simulation as the benchmark to validate the results of the proposed approach on the IEEE 118-bus network, 2, 383-bus Polish grid, and IEEE 68-bus system. The proposed approach is able to track the cascade path and replicate the end results of TM-based simulation with very high accuracy while reducing the average simulation time by {acute}{89}{88} 10 {acute}{88}{92} 35 fold. The proposed approach was also compared with the partitioned method, which led to similar conclusions.Next, we demonstrate that a further speedup can be achieved by a parallelized version of BEM-PC, which we call BEM-PC-parallel (BEM-PCP). In this version, the predictor subprocess of BEM-PC is run in multiple parallel processors for identification of oscillatory instability using eigen decomposition of the system matrix at post-disturbance unstable equilibria. Monte-Carlo studies on a 2, 383-bus Polish system confirm that BEM-PCP is on average 17% faster than BEM-PC and {acute}{89}{88} 40 times as fast as TM while maintaining the same accuracy as BEM-PC.Our next contribution is regarding the cascade mitigation in the context of dynamic models. To this end, we focus on the dynamic security assessment (DSA), which is a proactive strategy that helps avoid initiation of cascading failure. A comprehensive DSA entails that system operators are capable of inspecting hundreds of most probable contingencies every 15 {acute}{88}{92} 30 minutes using online state estimator data and system dynamic model. Ideally, one would get the most accurate sense of system security by running nonlinear time-domain simulations for {acute}{89}{88} 20 {acute}{88}{92} 30 s, which is computationally unmanageable for large-scale systems. This has led to application of Lyapunov-based direct methods, trajectory sensitivity-based approaches, and machine learning-based heuristics, all of which suffer from different drawbacks. To address this, we propose a fast detailed dynamic simulation approach based on BEM with stiff decay property. Unfortunately, BEM may converge to the unstable equilibrium in case of oscillatory instability in power systems. We propose an augmented BEM approach to solve this issue. It performs eigen decomposition on the post-disturbance system matrix of the linearized model obtained as a byproduct of BEM to detect oscillatory instability. Results from the IEEE 68-bus NE-NY network and 2, 383-bus Polish system confirm that the proposed method can maintain the accuracy of the traditional TM-based simulation while demonstrating {acute}{89}{88} 25 {acute}{88}{92} 31 times average speedup.Finally, this dissertation focuses on developing a comprehensive cyber-physical cascading failure model of power systems along with a cascade mitigation strategy. We propose a dynamic cascading failure model that considers realistic interdependencies between power and fiber-optic communication networks used for system monitoring and control in power grids. In this model, power line outages do not immediately disconnect communication links, whereas communication nodes have battery backup that starts depleting after considerable load shedding in the collocated bus or bus outage. When a communication node's battery is fully depleted, the node disconnects from the cyber layer, potentially reducing the observability and controllability of the power grid. A centralized optimal preventive controller (OPC) algorithm to minimize load shedding is proposed for cascade mitigation, which is applied selectively on fully observable and controllable islands. OPC considers AC power flow equations, multiple hard constraints, and treats overloading of lines as soft constraints. The results of Monte-Carlo simulations on the IEEE 118-bus and 2, 383-bus Polish systems demonstrate that the proposed OPC algorithm is effective in mitigating cascading failures. Finally, we demonstrate that our recently proposed BEM-PC can reduce the average simulation time by approximately 9 {acute}{88}{92} 26-folds compared to the TM with acceptable accuracy.{A0}
590 $a School code: 0176.
650 4 $a Communications networks. $3 3564745
650 4 $a Failure. $3 3561225
650 4 $a Communication. $3 524709
650 4 $a Buses. $3 870578
650 4 $a Technical communication. $3 3172863
653 $a Cascade mitigation
653 $a Cyber layer
653 $a Dynamic models
653 $a Machine learning
653 $a Uniform load shedding
690 $a 0459
690 $a 0643
710 2 $a The Pennsylvania State University. $3 699896
773 0 $t Dissertations Abstracts International $g 85-05A.
790 $a 0176
791 $a Ph.D.
792 $a 2023
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30720598