Dealing with interactions in modern power electronics dominated power systems
Introduction
We are experiencing a unique time in history whereby, driven by climate change, most governments have committed to reduce carbon emissions and many aim to reach Net Zero within the next couple of decades. These ambitious and binding goals require, amongst many other measures, the replacement of fossil fuel generation with carbon-free renewable generation. Underpinning this unprecedented transition is the use of Power Electronic Interfaced Devices (PEIDs), not only on the generation and demand sides, but also integrated as system assets, such as FACTS and HVDC systems.
The dynamic behaviour of PEIDs, and the ways that they interact with the rest of the power system, are more complex than conventional generators, and less understood. PEIDs can bring significant benefits to power systems, but their controls can also introduce harmful interactions if not considered carefully. Recent incidents worldwide have exhibited unwanted oscillatory behaviour of PEIDs that have caused loss of supply, economic costs and blackouts. Therefore, the transition to a clean energy future necessitates a detailed understanding of these increasingly complex interactions in modern power systems.
Within the remit of SC C4, two Working Groups (WGs) were created to better understand interactions in PEID-dominated systems and advanced methods of analysis: JWG C4/B4.52 and WG C4.49. Both WGs have completed their tasks and delivered two excellent Technical Brochures (TB 909 and TB 928), which are available at https://www.e-cigre.org/.
Dealing with Sub-Synchronous Oscillations (JWG C4/B4.52)
Subsynchronous oscillations (SSO) have been identified as a major concern in modern power systems. Conventionally, large thermal generating units such as nuclear and steam turbine plants were identified as vulnerable for SSO. This was mainly due to the torsional interactions with the long shafts of the devices. The recent field experiences and the research have shown that there is a significant risk of SSO associated with other generator types and power electronic based devices such as renewable generation, HVDC and FACTs.
JWG C4/B4.52 (Guidelines for Sub-synchronous Oscillation Studies in Power Electronics Dominated Power Systems) has investigated sub-synchronous oscillations (SSO) in modern power systems and the outcome has been published in TB 909. Considering current industry needs, the TB provides a general guideline for understanding relevant SSO phenomenon, study methodology, mitigation/prevention techniques and protection mechanisms.
- Classification of SSO
Having reviewed recent SSO events and keeping in-line with a recent classification proposed by the IEEE, an extended classification of SSO is proposed in TB 909, as shown in Figure 1.
SSO is divided into two main categories: Sub-synchronous Resonance (SSR) and Power Electronic Device Interactions (PEDI). SSR is further divided into Electrical and Torsional. This is to differentiate between the SSR that is purely electrical, where torsional systems are not involved, and those where a torsional system is involved. The Electrical type SSR could be due to resonance in the network when there are series compensation devices present, or it could be due to the negative resistance offered by a generator at a network resonance frequency. In both cases shaft systems are not involved. The Torsional type SSR is divided into three types: Shaft Torque Amplification, Torsional Interactions with the Network (TI-N) and Torsional Interactions with another Device (TI-D). The new category of SSO introduced in this TB is the PEDI. This category includes control interactions of two types: Control Interactions with the Network (CI-N) and Control Interactions with another Device (CI-D).
Figure 1: New classification of SSO.
- Systematic Approach for Analyzing SSO Issues
TB 909 proposes a systematic approach to investigate and resolve SSO issues in power systems. The proposed SSO study procedure has been divided into four sections:
- Screening of Potential SSO Risk
When a new project is planned or an existing SSO issue is analysed, the first step is to perform screening studies. At this stage, the possible SSO phenomenon is not clearly identified; devices contributing to the SSO are unknown; many operating scenarios and contingencies need to be considered; and detailed simulation models may not be available. Therefore, the main objective of screening studies is to evaluate the SSO risk while allowing false alarms and avoiding false dismissals. The TB presents an overview of the different methods available for screening the potential risk of various SSO types. The methodology and theoretical background of each method is explained along with the limitations, assumptions, modeling requirements and evaluation criteria. The TB also provides the reader with practical advice regarding the use of each screening method together with guidelines for the interpretation of results.
- Detailed Evaluation of SSO
The detailed study procedures include electromagnetic transient (EMT) simulations and small-signal stability analysis (eigen analysis). EMT simulation is the most popular study procedure due to its flexibility, such as the use of black-boxed device models. However, the root cause of SSO phenomenon is difficult to determine by inspection of time-domain simulation plots. In contrast, the eigen analysis (frequency domain) provides more insight into SSO phenomenon. The oscillations and damping in the entire study case can be identified from the eigenvalues and the devices contributing to the oscillations can be evaluated using eigen properties such as participation factors and mode shapes. TB 909 provides clear guidelines for conducting these analyses.
Once detailed studies have confirmed the risk of SSO, suitable mitigation measures need to be implemented. The TB identifies and describes a range of mitigation options. Some mitigation measures can be used as short-term temporary measures until a permanent solution is implemented.
In addition to the mitigation measures, suitable backup protection mechanisms need to be implemented to ultimately avoid damage to devices. Furthermore, power systems are rapidly evolving, and their dynamic behavior may change unknowingly. This means that great attention must be paid to the behavior of all devices involved, and sophisticated condition monitoring is therefore needed.
The TB also provides a review of industry practices for SSO evaluation and reported SSO issues. In summary, TB 909 provides a general guideline for SSO studies in power electronic dominated power systems.
Multi-frequency stability of PEIDs (WG C4.49)
WG C4.49 (Multi-frequency Stability of Converter-based Modern Power Systems) has examined new forms of instability and interactions driven by power electronic converters in modern power systems over a wide range of frequencies. The outcomes of the WG have been published in TB 928.
The TB presents an overview, status and outline of oscillatory instability and interaction analysis in converter-based systems. It covers definitions, classification, literature review and industry experience, including state-of-the-art stability and interaction analysis in power systems dominated by converters. The TB elaborates on small-signal instability phenomena, consolidates definitions, and explains available methods for analyses with their advantages and disadvantages to provide a common understanding for academia and industry. General guidelines for studies are provided, considering available information at different stages of an asset’s lifetime, as well as choice of tools and mitigation methods. A new benchmark system has been developed by the WG to aid academia and industry in the analysis of multi-frequency stability.
- Overview of instability phenomena and frequency range
In PEID-dominated power systems, the small-signal dynamics of converters can introduce negative damping in different frequency ranges, the severity of which depends on the controller types, the topologies, and the power system conditions. The root cause is the interaction of a converter controller (or several) with a grid resonance. Figure 2 shows the indicative frequency range considered for multi-frequency stability of power electronic control systems. It covers oscillations within the sub-synchronous (i.e. below the fundamental frequency) and harmonic (super-synchronous) frequency ranges.
Figure 2: Typical time frame for various dynamic phenomena in power systems.
- Stability analysis process for converter-based power systems
Figure 3 shows the proposed stability analysis workflow which can be easily adapted to any power system with a high level of PEID penetration such as photovoltaic power plants or wind power plants.
Figure 3: Proposed stability analysis workflow for converter-based power systems.
Step 1:Consider the short-circuit ratio (SCR). SCR is an indicator of p.u. impedance at fundamental frequency, and therefore does not consider resonances that may be present in the grid. However, it is indicative of how “strong” the network is at the point of interconnection and can be used (with caution) for an initial screening. Low SCR indicates possible stability challenges for grid-following converters due to weak grid connections. A high SCR can indicate challenges for converters with grid-forming control.
Step 2:Investigate small-signal stability. Small-signal stability is a necessary but not sufficient condition for entire system stability (small and larger disturbances). However, any system unable to withstand small-signal disturbances will not be able to operate in a stable manner. Thus, the first and principal task is to perform small-signal studies to investigate power system stability under a range of operating conditions. Various analytical methods are presented and discussed in the TB. These methods can yield insights into the nature of potential instability and enable optimised mitigation methods to be determined. Analytical methods are based on linearisation around a specific operating point. Therefore, multiple operating points, and therefore multiple linearised representations, are required to gain an understanding of the system’s behaviour as a whole.
Step 3:Validate small-signal stability studies. The linearisation conducted in Step 2 is a necessary simplification for analytical methods. Therefore, it is important to perform validation using a non-linear method, such as time-domain numerical simulations using detailed models of the power electronic converters with their controls.
Step 4:Perform large-signal stability studies. Power systems should remain stable not only when affected by small-signal disturbances but also by large-signal disturbances such as voltage drop and excessive inrush current due to transients, symmetrical and asymmetrical faults, frequency variation and phase jumps due to imbalance between power production and consumption, excessive harmonic distortion from power electronics or transformer energisation. Guidelines for conducting time-domain simulation analysis are included in the TB.
Step 5:Implement mitigation methods. If any possibly unstable events are identified in the previous steps, there is a need to apply suitable instability mitigation measures. There are multiple ways to avoid unstable operation, especially because, in comparison to classical power system components, converter systems are flexible due to their control and are therefore capable of changing their behaviour. However, any mitigation measure comes with a cost and impact of asset operational philosophy. Figure 4 provides an overview of various mitigation options, which are described and discussed in detail in TB 928.
Figure 4: Mitigation methods.