Title: Prevention and mitigation of high‐voltage direct current commutation failures: a review and future directions
Abstract: IET Generation, Transmission & DistributionVolume 13, Issue 24 p. 5449-5456 Review ArticleFree Access Prevention and mitigation of high-voltage direct current commutation failures: a review and future directions Yanan Zhu, Yanan Zhu orcid.org/0000-0002-2722-1538 Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorShuqing Zhang, Corresponding Author Shuqing Zhang [email protected] Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorDong Liu, Dong Liu State Grid Global Energy Interconnection Research Institute, Beijing, People's Republic of ChinaSearch for more papers by this authorLin Zhu, Lin Zhu State Grid Global Energy Interconnection Research Institute, Beijing, People's Republic of ChinaSearch for more papers by this authorSheng Zou, Sheng Zou State Grid Jiangsu Economic Research Institute, Nanjing, People's Republic of ChinaSearch for more papers by this authorSiqi Yu, Siqi Yu Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorYubo Sun, Yubo Sun Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this author Yanan Zhu, Yanan Zhu orcid.org/0000-0002-2722-1538 Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorShuqing Zhang, Corresponding Author Shuqing Zhang [email protected] Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorDong Liu, Dong Liu State Grid Global Energy Interconnection Research Institute, Beijing, People's Republic of ChinaSearch for more papers by this authorLin Zhu, Lin Zhu State Grid Global Energy Interconnection Research Institute, Beijing, People's Republic of ChinaSearch for more papers by this authorSheng Zou, Sheng Zou State Grid Jiangsu Economic Research Institute, Nanjing, People's Republic of ChinaSearch for more papers by this authorSiqi Yu, Siqi Yu Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this authorYubo Sun, Yubo Sun Department of Electrical Engineering, Tsinghua University, Beijing, People's Republic of ChinaSearch for more papers by this author First published: 13 November 2019 https://doi.org/10.1049/iet-gtd.2019.0874Citations: 5AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract With the increasing applications of high-voltage direct current inverters in heavy-load grids, commutation failures (CFs) pose a severe threat to the safe and stable operation of power systems. This study first sorts methods of CF inhibition into different categories and then investigates their effectiveness, adaptability and limitations. Considering the economic benefits and applicability, CF inhibition calls for integrated methods involving supplementary power devices, control strategies and prediction techniques. Finally, this study makes clear potential future research directions, suggesting first control potential is exploited by machine learning-based control, and then auxiliary devices to be added if necessary. 1 Introduction High-voltage direct current (HVDC) links outperform alternating current (AC) transmission systems for long-distance electrical power transmission. Therefore, HVDC links have gradually become a significant part of the modern power grid. HVDC techniques fall into two categories: the line commuted converter HVDC (LCC-HVDC) and the voltage source converter HVDC (VSC-HVDC). Compared with thyristors in LCC-HVDC, the insulated gate bipolar transistors (IGBT) in VSC-HVDC systems operate at a high frequency and relatively low voltage rating, so the VSC-HVDC systems have high power losses and their transmission capacity is limited. Therefore, despite the lack of flexibility, LCC-HVDC may still play a pivotal role in the foreseeable future because of its high economic efficiency and transmission capacity [1]. However, in LCC-HVDC systems, the transient process caused by commutation failures (CFs) poses a great threat to system security and stability by interrupting power transmission and threatening equipment safety [2]. A single CF can lead to consecutive or recurrent CFs and cause stability problems throughout the entire grid, such as angle stability issues at the sending end and voltage stability issues at the receiving end. In addition, CFs can decrease the reliability and sensitivity of relay protection systems or further initiate power shortages [3–5]. In heavy-load areas, such as northeastern Brazil and those served by the China Southern Power Grid and East China Power Grid, multiple LCC-HVDC links feed the system. In these systems, one CF may lead to concurrent CFs in several HVDC systems and increase the disturbance in the system, thereby further deteriorating the commutation process. Therefore, research on the prevention and mitigation of CFs has received increased attention in recent years. To date, CF prevention and mitigation methods fall into three categories: (i) improving primary equipment, including valves, converters and converter transformers; (ii) improving control strategies by CF prediction and parameter optimisation and (iii) creating system-level coordination strategies involving several HVDC converters and receiving AC systems. However, a CF can be triggered by many factors occurring in the entire power grid. Although many attempts were made considering subsets of influential factors, they still failed to exactly identify various CF patterns and avoid all of CFs. To better guide research on the countermeasures for CFs and their harmful effects, this paper introduces the basic concept of CFs in Section 1. Sections 2 and 3 summarise recent studies on CFs from the perspective of the corresponding mechanism, and analyse the effectiveness, adaptability and the limitations of all the methods for CF inhibition in three categories as well. Additionally, to overcome the aforementioned challenges, this paper notes future directions, suggesting first control potential be exploited by machine learning-based prediction control, and then auxiliary compensation devices be added if necessary to minimise CFs. 2 Mechanism of CFs 2.1 Definition of commutation and CFs As a half-controlled electronic switching component, a thyristor requires both a triggering pulse to the gate terminal and the forward voltage to turn on, while it only requires the reverse voltage lasting a period to turn off. After recombination of current carriers during the turn-off period, the blocking capability can be restored. Otherwise, the thyristor will stay on when the voltage shifts forward, regardless of the triggering pulse. The turn-off time of a thyristor is determined by its physical characteristics and is normally 400–500 μs. For a thyristor converter, the AC line voltage provides the voltage for commutation. If the forward voltage blocking capability of the valve is not fully recovered after commutation, the forward voltage may turn on the valve by mistake [6, 7]. In this case, the current cannot be transferred to the next valve in the firing sequence. This phenomenon is called a CF. Fig. 1 demonstrates the CF process between valve 1 and valve 3. In Fig. 1, β refers to leading firing angle, μ is the commutation overlap angle, γ is the extinction angle and (1)where α is the firing angle. If the extinction angle is smaller than the minimum extinction angle, then four two-valve-on stages may occur. During stage A, because valve 1 falsely stays on, the direct line voltage decreases compared with a normal commutation process. During stage B, the valve 5 cannot turn on in the absence of the forward voltage, and a direct line-side short circuit occurs. This stage may cause a surge in direct line current. In stage C, valve 1 stays on, so the triggering pulse of valve 1 does not make a difference. In addition, the direct line voltage at the converter drops or even reverses. After stage D, if valve 1 successfully turns off, a complete CF process ends. It should be noted that these four stages do not necessarily occur because of the distortion of the commutation voltage. Fig. 1Open in figure viewerPowerPoint CF process between valve 1 and valve 3 2.2 Inducements to CFs Generally, a commutation voltage deficiency or distortion caused by AC faults, uncoordinated multi-HVDC control strategies and control failures are the main causes of CFs [8, 9]. AC faults cause disturbances by increasing the direct line current, decreasing the AC voltage and increasing the dynamic reactive power demand of the converter station near the faults. These issues can increase the difficulty of commutation or even lead to a CF. For example, when an AC short-circuit fault occurs close to the inverter, the inverter direct line voltage can drop to zero or even reverse, and the direct line current can increase to 1.5 times the rated current in a short period. At the same time, the interaction between the HVDC system and dynamic loads can lead to voltage stability problems or even repeated CFs after the fault is cleared [10]. In addition, because the same fault occurring at different times has different effects on the commutation area, the result of the commutation process can also be different [11, 12]. Concurrent CFs associated with uncoordinated control strategies exist in multi-infeed HVDC systems and become an important issue threatening system security and stability. If HVDC inverters are close to each other, and recovery from the disturbances caused by a single HVDC may cause voltage drops at other HVDC inverter stations. Without coordinated control strategies, the demand for a large amount of reactive power at several HVDC systems can result in concurrent CFs. Thyristor valve faults and the loss of triggering pulses can also lead to CFs. Since such failures cannot be prevented or mitigated in advance and the occurrence probabilities are low, this paper mainly focuses on the prevention of the CFs caused by AC faults and uncoordinated strategy problems. 3 Prevention and mitigation of CFs According to the physical process of a CF, there are four technical routes for CF prevention and mitigation, as shown in Table 1. Table 1. Comparison of the four technical routes for CF prevention and mitigation Technical routes Main methods Optimised objects (1) increase commutation margins (a) improving the performance of valves 3.1.1 valve (b) reducing the advanced firing angle based on an extinction angle/direct line current calculation/commutation area 3.2 control strategy (2) temporarily reduce the equivalent commutation impedance during the fault (a) installing a fully controlled module in the converter to aid in the commutation process; replace parts or all the half-controlled valves to create fully controlled valves 3.1.2 converter (b) improving the structure of converter transformers 3.1.3 converter transformers (3) eliminate specific harmonics (a) improving the performance of converter transformers 3.1.3 converter transformers (4) temporarily decrease the direct line current during the fault (a) installing superconducting fault current limiters/insert controllable reactance 3.1.4 direct line circuits (b) adjusting the VDCOL parameters 3.2 control strategy These CF prevention methods can also be classified into three categories according to the implemented objects: primary equivalent, HVDC control and system-level coordination, as discussed in the following three sections. 3.1 Improvements to primary equipment 3.1.1 Valve-level improvements During the transient valve reverse recovery process, the carrier concentration decreases by migration, diffusion and recombination. The turn-off time is related to the current rate of change, temperature and the characteristics of the valves [13]. Reducing the turn-off time of valves can increase the margins for commutation and improve commutation performance. However, because the turn-off time is negatively correlated with the on-state voltage drop, a shorter turn-off time is accompanied by a larger on-state loss [14]. 3.1.2 Converter-level improvements The improvements to converters focus on the topological structure of bridge arms and the entire converter. As shown in Fig. 2, improving bridge arms refers to adding flexible blocks in/between conventional bridge arms. Capacitor-commutated converters can make the commutation progress easier and faster with the help of capacitors in block A [15]. However, the capacitors result in additional reactive power consumption, harmonics and overvoltage issues that should be eliminated by additional filters and lightning arresters [16]. Adding thyristor-based full-bridge blocks or IGBT-based full-bridge blocks in block B can accelerate the commutation process by providing forward and reverse voltages in different commutation stages [17, 18]. In the literature works [19, 20], authors added additional modules at block C instead of destroying the original structure of the bridge arms. This method can ensure commutation during fault conditions, whether or not there are AC commutation voltages. However, the above three methods all add additional modules to the original inverter, which inevitably increases the difficulty of coordination control between blocks and valves, and results in higher costs. Fig. 2Open in figure viewerPowerPoint Improvements in/between bridge arms Improving an entire converter refers to replacing the LCC converter with a flexible converter. For example, in [21], it was proposed to replace the LCC inverters with MMCs to suppress CFs. Authors of [22, 23] suggested that applying VSC inverters instead of LCC inverters can fundamentally avoid CFs. Guo et al. [24] created a new inverter structure by combining LCC inverters and VSC inverters to reduce the probability of CFs. These improvements required a large number of sub-blocks and high-performance devices, increased costs and reduced reliability. In addition, the overcurrent and overload capacities of MCCs and VSCs are smaller than LCCs. Finally, different types of converters and blocks should be coordinated by system-level strategies, making HVDC control more difficult. Moreover, the transmission power capacity of these systems is not as good as that of conventional LCC-HVDC systems. In conclusion, improving the converter structure or replacing valve groups can help improve the commutation process. However, this process generally reduces the reliability of the system and substantially increases the construction and operational costs. 3.1.3 Converter transformer improvements The commutation impedance mainly consists of the leakage impedance of commutation transformer. A smaller commutation impedance yields a wider safety margin and reduces the risk of a CF. However, the leakage impedance of the converter transformer should satisfy the demand for short-circuit limitation. Thus, we cannot significantly reduce this factor at will. There are two methods for optimising the transformer impedance: (i) balancing the CF risk and the demand for limiting the short-circuit capacity and (ii) redesigning the winding connection type and improving the transformer structure. Based on the second method, Luo et al. [25] proposed a new transformer structure that can eliminate the negative effects of leakage impedance and satisfy the demand for limiting the short-circuit capacity at the same time. Li et al. [26] proposed a similar transformer structure by applying inductive filtering technology. This structure can adjust the commutation impedance with fully tuned branches and improve the reliability of the LCC commutation process. Modifications to the transformer structure to regulate the commutation impedance are theoretically effective for CF suppression. These methods target certain harmonics and cannot prevent CFs caused by voltage drops or phase shifts in an AC system. In addition, in this structure, the filter is designed to connect to the middle winding of the three windings in the transformer, which is difficult to implement. Therefore, this approach cannot be practically applied in actual systems. 3.1.4 Installing superconducting fault current limiters (SFCLs) Lowering the direct line current can help prevent CFs, and the installation locations of SFCLs were discussed in the literature [27]. In addition, Lee et al. [28] suggested that SFCLs could improve the restoration time characteristics of converters after CF. However, SFCLs are always installed at high-voltage level buses or on direct lines, which can result in application problems, such as high operational costs, heat production by the SFCL during the transient process, insulation design issues, and other problems [29–31]. 3.2 Improvements to the HVDC control A dynamic shift in the firing angle can prevent certain types of CFs, so the key to CF prevention and mitigation includes the prediction or diagnosis of the detailed forms of CFs and the creation of a targeted strategy. 3.2.1 CF diagnosis The implementation of the predictive method should be fast. Therefore, a high-risk situation can be immediately identified to allow time to adjust the control strategy. In addition, such methods should be feasible in practical projects as well. (i) Minimum voltage drop methods (voltage distortion methods): The minimum voltage drop methods assess the commutation process by comparing the commutation voltage drop to the critical value. The challenge lies in the fact that it is difficult to determine the critical value. The critical value based on the quasi-steady state model only considers the voltage magnitude and ignores the influences of the phase shift and waveform distortion. Therefore, based on this critical value, a CF occurring in other scenarios cannot be accurately predicted. In addition, the method is subject to the fault time and is insensitive to faults occurring around the zero-crossing point of the voltage [32]. As one of the widely used CF prevention modules, CFPPREV also determines whether CF occurs by comparing the voltage RMS value to the critical value, and determines the change value of the firing angle according to the severity of the fault [33–38]. (ii) Extinction angle methods (a) Based on measurements: Measurements of the extinction angle are obtained by comparing the zero-crossing point of the valve currents and the phase-locked voltage. Therefore, the sensitivity of this method depends on the judgment of the zero-crossing point of valve currents, which is difficult to obtain under certain conditions. In addition, this method is not strictly a predictive method, so it has poor real-time performance. Another type is based on the measurement of PMU [39]. However, this method relies on PMU data measurement and needs to know the Thévenin equivalent of the network. (b) Based on prediction: The extinction angle is predicted based on the quasi-steady-state model. CFs can be identified by comparing the reference value and the predicted value of the extinction angle. The reference extinction angle is usually set large than the inherent minimum extinction angle of a valve. Taking the six-pulse inverter as an example, the extinction angle γ can be calculated by (2)where k is the converter to transformer ratio, Id is the direct line current, ωLc is the commutation impedance, E is the commutation voltage, and β is the leading firing angle of the inverter. If the extinction angle is smaller than the reference extinction angle, then a CF is predicted to occur. This method is superior to that based on measurements in terms of predictive performance. However, the predictive value may yield serious errors during the transient process, which can lead to violations of the quasi-steady-state condition. Therefore, accurate predictions are difficult to obtain in some cases. Thus, the extinction angle formula optimisation does not fully meet the analysis needs. Wang et al. [40] calculated the extinction angle by fitting the historical data and predicting the commutation voltage and DC current during the transient process. However, prediction of commutation voltage is not guaranteed to be accurate during diverse events with this method. Therefore, a new scheme, such as a data-driven prediction scheme, is expected for the CF-related quantity prediction [41, 42]. (iii) Waveform diagnosis methods: By analysing the voltages and currents of valves and the AC side of the converter, some characteristics, including the decomposed DC and sin-cos components, involved in the transient process can be summarised and used as criteria to calculate the extinction angles and diagnose CFs [43–45]. Zhu et al. [46] used the wavelet transform technology to analyse the CF waveforms caused by different factors and identify the causes of the CFs in the actual system [47, 48]. Research on waveform diagnoses can aid in summarising the electrical characteristics of a CF, but in fact, this approach assumes that a CF lasts for a given period. Therefore, it belongs to detection technology rather than prediction technology. Such methods cannot predict whether and how a CF will occur. They provide limited guidance for control strategy development and system recovery from faults. 3.2.2 Control strategies against CFs Based on CF prediction or detection, some control strategies and optimisation solutions are proposed. (i) Voltage-dependent current order limiter (VDCOL): Currently, VDCOLs are the main CF elimination tools applied in actual projects. When the voltage near the inverter station drops, the rectifier side of the VDCOL is used to reduce the reactive power demand of the inverter station by reducing the direct line current and adjusting the recovery speed of the HVDC system [49]. Studies of VDCOL control optimisation fall into three categories, as shown in Table 2. Table 2. Optimisation control based on VDCOLs Types Purposes Main methods input optimisation to reflect the influence of AC systems and the interactions among HVDC systems adding AC voltage to the input signal of the VDCOL module [49] structure optimisation to reduce fluctuations in the direct line current during a fault setting a virtual resistance [23] designing a PI/fuzzy controller [50–52] parameter optimisation to speed up the recovery from CFs reducing the time constant of the voltage measurement module [49] to help multi-infeed HVDC systems sequentially recover optimising the current settings with predicted values [53–57] dividing the voltage drop and recovery process into stages and setting different power recovery rates [58] It is more economical and easier to prevent CFs with VDCOLs compared to using other methods. However, VDCOLs only work when the AC bus voltage on the other side of the HVDC lines decreases, which leads to a transmission delay. Therefore, this approach responds more slowly than local measurement and local control strategies. Second, measuring the AC voltage and direct line current can result in a millisecond-scale delay. Third, the direct line current order of VDCOLs is linearly related to the direct line voltage, and the direct line current order cannot quickly decrease [59]. Thus, this approach can be used as a supplement to the CF prevention scheme rather than the main strategy. (ii) Constant extinction angle (CEA) control: The CEA control strategy is applied at the inverter side. However, if the extinction angle is too large, the power factor of the AC side will drop, which may inhibit voltage recovery. Therefore, the CF recovery and voltage stabilisation processes should be balanced when setting the reference extinction angle. CEA control includes measured-type and predictive-type control. Since the measured-type control is a type of closed-loop control, it is not ideal for CF prediction. Therefore, predictive-type CEA control strategies are mainly used. Taking the six-pulse converter as an example, in CEA control, the original calculation formula of the firing angle α is as follows: (3)where k is the converter to transformer ratio, Id is the direct line current, ωLc is the commutation impedance, E is the commutation voltage, and γ0 is the reference extinction angle. There are two ways to optimise predictive-type CEA control strategies. (a) Reset the reference extinction angle γ0 : In actual operations, the turn-off time changes according to the operating conditions. Therefore, the reference extinction angle γ0 is not constant [60]. Son et al. [2] adopted an adaptive value instead of a constant value as the reference extinction angle according to the commutation voltage [61]. Thus, the control adjustment margin is expanded. The effectiveness of this method has been confirmed, but the proposed optimisation schemes are based on analyses of the direct line current and the amplitude and phase of the fundamental frequency component of the commutation voltage. The influence of harmonics is not considered. Thus, the reference extinction angle can be further optimised in consideration of harmonics. (b) Taking the transient process into consideration: In a transient process, E and Id change. Since the current at the end of commutation is unknown, the following equation is used (4)where Kd is the scale factor and is set according to experience. The meanings of other symbols are the same as those in (3). This formula uses an approximate direct line current to balance efficiency and accuracy. However, in this method, E is also influenced by harmonics during the transient process. To date, no control optimisation for the commutation voltage correction has been reported. (iii) Firing sequence control: In the above control strategies, the same firing angle control signal is sent to all bridge arms. In fact, under specific faults, the firing angle can be formulated for different bridge arms. Luo et al. [62] proposed changing the triggering sequence of bridge arms during CFs. However, because the order of the trigger pulse cannot be adjusted in an actual project, this method is not practical in application. 3.3 Improvements in system-level coordination The recovery process from a CF for a single HVDC line depends on the reactive power provided by the AC system. As the number and concentration of LCC inverters in multi-infeed HVDC systems increase, the CF recovery process is related not only to the local HVDC control strategy but also to the strategies of other HVDC lines. If multi-infeed HVDC lines synchronously recover too fast and the system fails to supply sufficient reactive power, voltage stability problems can occur with follow-up CFs. In conclusion, both HVDC control systems and terminal AC systems can affect recovery progress. Thus, a coordinated control scheme including both AC control and HVDC control should be established [36]. 3.3.1 Multi-infeed HVDC link coordination As shown in Fig. 3, HVDC system coordination falls into three categories: HVDC terminal location selection before faults occur, local converter control during/just after faults, and multi-converter control after faults. Fig. 3Open in figure viewerPowerPoint Coordinated control scheme As an active prevention method, HVDC terminal location selection approach can help reduce the probability of concurrent CFs in different HVDC systems during transient processes. However, the HVDC terminal location is selected mainly considering the load distribution and other economic factors. In addition, this method has poor adaptability to diverse operating conditions and the subsequent grid changes. During and just after a fault, the inverter control near the fault can minimise the power loss across the transmission line and the duration of the CF. The inverter control responds quickly because it is based on local measurements and there is no communication delay. The strategies for multi-infeed HVDC systems should be coordinated among different HVDC lines to find a balance between the power recovery speed of a single HVDC line and the stability of the entire system. Coordinated control strategies of multiple HVDC lines consider the interactions among converters. They can make the restoration process of the entire system smoother and eliminate the adverse interactions among HVDC lines, instead of blindly attempting to increase the recovery speed of a certain HVDC line. Most of them are based on VDCOL control to implement staggered progressive recovery [49, 63, 64]. However, there are still questions remained, such as parameter design and optimisation methods considering the stability and economy of different HVDC links and the design of communication systems to efficiently implement coordination strategies. 3.3.2 Coordinated AC control strategies AC control strategies for CFs include active prevention strategies and reactive power compensation during and after faults, as shown in Fig. 3.
Publication Year: 2019
Publication Date: 2019-11-13
Language: en
Type: review
Indexed In: ['crossref']
Access and Citation
Cited By Count: 30
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot