Enhancing power system stability with adaptive under frequency load shedding using synchrophasor measurements and empirical mode decomposition

In recent years several catastrophic power systems blackouts have occurred worldwide. Various reasons have been declared for these failures. Economical limitations due to power system restructuring restrictions, inadvertent operation of protective relays and inefficient design of conventional load shedding schemes are of the most important reasons causing these blackouts. In fact, due to both economical and technological restrictions, it is not possible to completely prevent these blackouts. However, with the aid of some protection and control strategies, frequency and intensity of these blackouts may be reduced. One of the important protection strategies used for this purpose is a class of protection schemes known as ‘System Protection Schemes’ or ‘Wide Area Protection Schemes’. One of the most commonly used types of system protection schemes, generally accepted after the north-eastern blackout of 1965, is Under Frequency Load Shedding (UFLS) scheme. Conventional under frequency load shedding scheme is designed to retrieve the balance of generation and consumption following a disturbance. In the conventional load shedding method frequency settings, time delay settings and the amount of load to be shed in each step are constant values. The loads to be shed by this scheme are also constant load feeders and are not selected adaptively. Using this constant non-adaptive load shedding algorithm is not the most efficient scheme for all power system disturbances. In some combinational disturbances, events causing frequency to drop are followed by other events causing voltage drop. In these cases, since loads are voltage dependent, total system load is reduced and system frequency might not decrease so much to activate UFLS relays. However, the system could eventually collapse due to voltage instability. In many such cases the system would survive if the load shedding relays operate adaptively and appropriately. For example if for large disturbances, higher frequency settings and lower time delays are used adaptively, a faster load shedding response is obtained and as a result system collapses may be prevented. In this paper a new UFLS algorithm is proposed. The purpose of this algorithm is to adaptively adjust speed of load shedding based on the magnitude of disturbance. In this method rate of frequency decline is used as a criterion to determine intensity of disturbance. Thereby, for large disturbances higher frequency settings and lower time delays are used. Application of the proposed algorithm to the simulated model of Khorasan network in Iran confirms its satisfactory performance. As the results of simulations show, several voltage collapse instabilities may be prevented by using the proposed adaptive UFLS method.

Adaptive underfrequency load shedding in systems with renewable energy sources and storage capability

A novel adaptive control strategy for under frequency protection and load shedding associated with synchronous phase measurement unit (PMU) is proposed in this paper. With the help of the PMU, the magnitude of active power deficit is estimated by using the simplest expression of the generator swing equation and static load model since the frequency, voltages and their rate of change can be obtained by means of measurements in real-time. Then, a robust control strategy with consideration of the uncertainties in system is presented. In the proposed robust control strategy, the under frequency load shedding (UFLS) schemes are used to prevent system frequency collapse. And minimal amounts of remote information are required. The effectiveness and robustness of the proposed control strategy investigated carrying out on 17-generator 69-bus test system under various disturbance scenarios

Wind energy is becoming a major source of generation in many countries because of its zero fuel cost and no air pollution. Due to the integration of large-scale wind power in conventional grids, synchronous generators are being economically replaced. Modern wind turbine generators (WTGs) are based on power electronic interfaces. Hence, unlike synchronous generators, they do not usually provide inertia and governor response. Therefore, from a power system security point of view, wind penetration can be limited by frequency response criteria. Up to now, several methodologies have been proposed to estimate the maximum wind penetration level. However, none of them suggest how a system operator could be immediately informed about a secured wind penetration limit as soon as a generation profile is known. Therefore, a tool to estimate the maximum wind penetration level by maintaining an adequate frequency response is proposed in this thesis. The proposed tool is able to instantly provide the highest margin of wind power that could be dispatched at a particular load condition without violating frequency response constraints. In some power systems, the combined generation capacity of wind and distributed photovoltaic (PV) sources is close to reaching the level of average demands (e.g. South Australia). Under high availability of wind and PV, a few synchronous machines may be required to be committed for frequency regulation and hence the system can be viewed as a low inertia grid. Such a grid could be potentially at risk of experiencing an excessive rate of change of frequency (ROCOF) after a contingency. A high ROCOF may initiate tripping of other synchronous generators. As a result, network frequency response may become more vulnerable and the system may be subject to significant under frequency load shedding (UFLS) or even at risk of a blackout. Furthermore, some existing distributed PV units may have a default under frequency protection setting, which is higher than the UFLS threshold. Hence, a contingency may cause a subsequent trip of those PV units. It may result in an unacceptable low frequency, triggering further load shedding. To explore the above issues, the effects of cascading contingencies triggered by high ROCOF and PV tripping on the frequency response of a low inertia grid are investigated in this thesis. Remedial measures using synchronous condenser and WTG frequency support to prevent the cascading contingencies are also suggested. Any mismatch between load and generation is intended to be balanced by contingency reserve, which is also known as contingency Frequency Control Ancillary Services (FCAS) requirement. Load frequency relief (LFR), which represents the effect of frequency dependent loads on power system frequency excursion, is crucial for correctly evaluating contingency FCAS requirement during a generation dispatch to ensure an adequate frequency response. Conventionally, the LFR is considered as a fixed quantity during the estimation of frequency response and FCAS requirement. However, recent experience in the Australian power system suggests that such an assumption may lead to an inaccurate outcome. To explore the above issue, the LFR is investigated based on field measurement data. Data were captured from different locations in the Southern states of Australia. An approach is developed to identify the predominating factors affecting the LFR. A technique is then proposed to appropriately determine the LFR and contingency FCAS requirement at any operating conditions. Along with frequency response, short-circuit performance is another vital index regarding power system security. Modern WTGs have limited fault current contributions, which may cause unacceptable short-circuit ratios at the grid connection point of those wind power plants. Traditionally, frequency response and short-circuit performance are individually improved. However, both of them are simultaneously affected by high wind penetration. Hence, a common approach to concurrently enhance both indices is essential. Therefore, an idea of operating some of the retired synchronous generators in the synchronous condenser mode, which is termed as ‘post-retirement scheme’, is introduced in this thesis. Such a second use of the retired synchronous generators can jointly improve frequency response and short-circuit strength during high wind penetration. A methodology is proposed to determine when and how much post-retirement scheme should be deployed for ensuring adequate security performances in a wind dominated power system.

This paper explains a fast acting Under Frequency Load Shedding and restoration scheme. This load shedding technique has an added advantage of selecting the amount of load which is to be dropped at a particular frequency threshold. The amount of load dropping is selected based on the rate of change of the frequency. Automatic restoration of loads is discussed with engaging a command signal from the system control centre. On/Off control of the power factor correction capacitor is also explained based on the magnitude of its terminal voltage. A three phase voltage vector Phase Lock Loop (PLL) is developed in EMTDC/PSCAD environment to measure the frequency of the power system. The PLL is tuned together with a filter to smooth the frequency measurement. This PLL measures the frequency online and based on phase angle deviation. Therefore it has a fast initial response. A load shedding factor is selected based on the rate of change of frequency between the specified thresholds. According to the load shedding factor the amount of load is selected. Load shedding and automatic load restoration operations are triggered accordingly when the frequency passes its relevant thresholds. The proposed distributed load shedding and restoration control technique was tested for IEEE 12 bus system at two different grid substations. The simulation results confirmed the fast, proper and smooth operations of the proposed technique.

Under frequency load shedding is implemented to restore power system frequency stability if system frequency drops below the operational set point during major disturbance such as lost of generation. Different countries/utility companies have their own philosophies in implementing the under frequency load shedding scheme. Generally, it is based on country/utility requirements, e.g. the overall power system network and the country's demographic. This paper presents the principles and implementation of the under frequency load shedding (UFLS) and presented using simulations of 56 test bus-system. The performance of the developed schemes under various conditions of disturbance were compared and analyzed. All the simulation works were performed using Siemens PTI software PSS®E.

This paper proposes a scheme that integrates rate of change of frequency (ROCOF) and under frequency (UF) elements to design more effective load shedding response. A new load block activation algorithm is also introduced that improves the overall load shedding (LS) performance. The combination of ROCOF and UF criteria is used to overcome the LS block discrimination problem and to provide a flexible and decentralised solution to mitigation. The proposed method is verified through simulations on the IEEE 39-bus test system.

Droop control, inertia control, and PD control are three typical frequency response (FR) strategies of grid-tied inverters interfaced renewable energy sources (RES) to restrain the power system frequency indicators, i.e., frequency deviation ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Delta \omega $ </tex-math></inline-formula> ) and the rate of change of frequency (RoCoF). Due to the highly limited inverter capacity, the FR performance in power systems with high RES penetration should be optimized. To this end, this paper first analyzes the impact of high-penetration renewable energy injection on power system frequency stability, and derives the system frequency expressions in the absence of an FR scheme, and in the presence of droop/inertia/PD controls that are representative for various virtual synchronous generator-based FR schemes. Under the same constraint of available FR capacity, control parameter ranges of these FR schemes are obtained, and their optimal conditions for restraining the maxima of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\Delta \omega $ </tex-math></inline-formula> and RoCoF are identified and compared analytically. Comparison conclusions are finally proved by the provided experiment and simulation results.

A dynamic-adaptive load shedding methodology to improve frequency resilience of power systems

Modern power systems are pushed to operate close to their secure limits due to increased system complexity and high penetration of renewable energy sources (RES). Conventional stability assessment (SA) and control methods face significant adaptiveness and robustness issues under highly uncertain and variable modern power systems. In recent years, with advanced real-time system monitoring and communication techniques like phasor measurement unit (PMU) devices, huge volumes of network-wide time-synchronized measurements can be collected and utilized. Under this new reality, data-driven methods have been identified as powerful tools for online and real-time stability assessment and control. This Ph.D. research aims to develop data-analytic methodologies online and real-time assess the system stability status and provide intelligent control actions when the power system is subject to a severe contingency. Specifically, this research covers a series of new methods in pre-fault SA, post-fault real-time SA and power system emergency control. Special efforts are devoted to enhancing the accuracy and robustness of data-driven methods under some extreme scenarios and practical problems, and a confidence-aware data-driven control scheme to manage the potential prediction error better. In data-driven SA, the point and consecutive data missing at both offline and online stages are studied, and two missing data-robust methods are developed, respectively for pre-fault SA and post-fault SA. For the pre-fault method, to enhance the robustness of SA against incomplete PMU measurements, a hybrid ensemble learning-based online SA method is proposed with feature-validation-based missing data estimators. Considering the possibility that some features may have a low correlation with other measurements, a feature validation process detects the potentially inaccurately estimated features to ensure the model performance and data restoration quality. Further, a novel spatial-temporal recurrent imputation network, SRIN, is proposed for time-series data missing at both sensor and PMU levels. The proposed SRIN method jointly learns the imputation for missing values and prediction output through an integrated multi-task recurrent network model, where the imputed values and the SA output can be simultaneously and effectively updated during backpropagation. In this way, the proposed method can be effectively trained by incompletely recorded datasets, which is immensely valuable for practical applications without high-quality historical data. Moreover, by learning shared representations or features across tasks, multi-task models can better exploit the information present in the data and improve the overall performance of both tasks. In addition, considering the time-varying and location-varying missing data patterns, the imputations of SRIN are derived by an intelligent combination of history-based and feature-based estimations through a trainable weighting component. Thereby, the proposed method is highly adaptive to various data missing scenarios, including both spatial and temporal consecutive missing patterns. In emergency control, data-driven methods for event-based load shedding (ELS) and response-based load shedding (RLS) are developed. Specifically, a hierarchical data-driven method is proposed for real-time prediction of ELS to mitigate fault-induced delayed voltage recovery (FIDVR). First, a novel adaptive trajectory sensitivity analysis is developed to optimize the ELS solutions offline and construct the ELS database. Then, by hierarchically modelling the ELS problem as a multi-output classification subproblem and a regression subproblem, the critical load buses mainly responsible for the delayed voltage recovery can be identified, and the over-shedding buses can be effectively reduced. Moreover, based on an improved weighted kernel extreme learning machine (WKELM) method, the direct mapping between system pre-fault operating conditions and corresponding control variables is constructed. Meanwhile, the established weighted learning scheme can extract the decision-making knowledge from the highly sparse and imbalanced dataset and emphasize on the rare yet critical cases. Therefore, the risk of insufficient load shedding is reduced, and better ELS effectiveness can be achieved. Subsequently, a probabilistic confidence-aware method for real-time RLS is developed to manage the potential prediction error better. The method is based on a two-stage process where a scalable Gaussian process (SGP) method determines the average load shedding amount at the first stage. The upper bound of the prediction interval is considered supplementary in the second stage. In parallel, if an RLS case is identified as unconfident, the conservative rule-based load shedding is taken as a backup plan rather than adopting LS estimations that cannot be trusted. Guided by the probabilistic information, the potential prediction error can be fully considered, and the confidence information can be incorporated into the control process, thereby more accurate and reliable load shedding actions can be taken. Leveraging confident predictions-based 2-stage RLS process, the excessive load loss can be reduced while ensuring high control effectiveness. All the proposed methods have been well verified on simulation experiments over standard power system models.

Power system blackouts have become a serious problem for electric utilities especially in recent years. Different forms of system instability have emerged in recent blackouts, such as voltage instability and frequency instability. To counteract each form of system instability, special algorithms have been designed in the protection system, e.g. Under Frequency Load Shedding (UFLS) and Under Voltage Load Shedding (UVLS) schemes. One of the major weaknesses of these algorithms is that combination of different forms of instability is not considered in their design, while any one form of instability may not occur in its pure form. This is particularly true in highly stressed systems and for cascading events. This paper presents an adaptive algorithm to combine UVLS and UFLS schemes. The purpose of this method is to enhance the flexibility of under frequency relays and increase the security of power system during large disturbances by improving system voltage stability margins. In the proposed algorithm, loads with greater voltage decay are shed sooner. In this way, locations of load shedding become dependent to the location of disturbance and voltage stability margins of the system are enhanced. Indeed, load shedding is performed faster for severe events accompanying large voltage or frequency declines. Using this load shedding method, faster reactions could be obtained for major system failures. This way, system blackouts could be better controlled. Performance of the proposed method is compared with the conventional UFLS method. In this paper, dynamic model of the Khorasan HV network of Iran national grid is used for simulation. This network includes 75 high voltage 400 kV and 132 kV buses with about 2700 MW of generation. Performance of the schemes has been evaluated using dynamic simulations as well as the P-V curves, and reactive power margins for a number of events. Simulation results show that conventional UFLS algorithm will not necessarily result in acceptable voltage stability margins for the system following a disturbance. Meanwhile, the proposed algorithm improves voltage stability margins. The proposed algorithm is suitable for saving system stability after occurrence of major system disturbances.

Load shedding is performed, as last option, for emergency measure when system frequency drifts below the allowed limit. Utilities use underfrequency load shedding schemes to bring back frequency within limit. In this article, an adaptive underfrequency load shedding scheme is proposed for the power system in the presence of solar photovoltaic plants using wide-area measurements. Stepwise load shedding is performed at selected buses with consideration of any change in photovoltaic power output and load/generation at each step. Share of load to be shed at a bus is based on the voltage stability index of the buses at that instant considering predefined load priority. The voltage stability index at a bus is obtained using the Thevenin equivalent of the system. For a solar photovoltaic plant integrated system, the Thevenin equivalent changes with solar power variation, where available methods become inaccurate. A Thevenin equivalent calculation approach is proposed in this article using an equivalent model of synchronous generators and photovoltaic plants. The proposed load shedding method is tested on the New England 39-bus system and the 233-bus North Eastern Indian grid with integrated photovoltaic plants. It is compared with available methods and found to be a better technique in bringing system frequency to normalcy with adaptive load shedding.

The identification of voltage collapse (VC) is crucial to prevent a possible collapse of the preventive control actions and voltage security evaluation. Selecting the critical buses in the network can assist in determining the locations where voltage collapses may occur. VC initially manifests at the weakest bus in a serious emergency case. So, for operational and planning purposes, it is essential to detect critical buses in an electric power system. The analysis of the power system in view of voltage stability (VS) was performed under different loading conditions using voltage stability indices (VSIs). The line voltage stability indices in terms of the fast voltage stability index (FVSI) and line stability index (LMN) are utilized in this paper to identify the weakest bus in the network. The suggested procedure is tested on a standard network (IEEE 14-bus system) at different overloading conditions (25 – 100 % overload) utilizing power system analysis toolbox (PSAT) software. The results proved that this index has a higher standard for predicting the point of voltage collapse and identification the weakest bus in the network, so consider more reliable and accurate relative to other indices.

Power systems with high penetration of inverter-based renewable generation units result in reduced inertia. In such power systems, the frequency events due to plant outage or large sudden increase/decrease in load power cause an increased rate of change of frequency (RoCoF) and larger frequency deviations. The higher RoCoF and larger frequency nadir can cause frequency instability which leads to blackout. Energy storage units may be deployed in such systems to support the frequency response. In this paper, Supercapacitor (SC) energy storage is used to provide the inertial response in power system with high penetration of intermittent renewable energy sources. Derivative control is employed for inertia emulation of SC. An additional control input signal based on the frequency deviation is added to improve the inertial response. The robustness of the proposed control is tested with different penetration levels of renewable generation. A comparison between the derivative control and the proposed control is presented to show the effectiveness of the proposed scheme.

When a disturbance occurs in a power system which leads to a decline in frequency, the power grid voltage stability is also at risk. In this research, we propose an improved under frequency load shedding technique that incorporates both the estimated disturbance and the voltage index of load buses. Based on the swing equation the disturbance is estimated in real time and the value is used for load shedding. Instead of shedding prefixed load-buses, all load-buses participate in the shedding process. When a disturbance occurs the voltage drop of each bus is expressed as a fraction of the total calculated voltage drops, and this gives the bus voltage index of the load buses. The greater the bus voltage index the more the load that is to be shed from the respective bus. It is also shown that the use of phasor measurements reduces the complexities of using the Jacobian to calculate voltage stability indicators.