Diagnosis method of open-circuit fault for T-type three-level inverter based on residual voltage

Abstract

Multilevel inverter topologies have been widely used in applications of wide-power range and fault diagnosis of such circuits is becoming more and more important. T-type three-level inverters have advantages in efficiency compared to common neutral-point-clamped inverters. This paper proposes a new diagnosis method of an open-circuit (OC) fault for a T-type three-level inverter. The method is implemented by measuring the phase voltage of the circuit and comparing it with the reference voltage of the circuit to obtain a residual voltage. Firstly, according to the analysis of different current conduction paths of the circuit, the voltage level changes of the circuit bridge voltage and phase voltage in different device open modes are studied. Then, based on the above analysis, a device open-circuit fault diagnosis scheme is proposed based on the residual voltage. Finally, simulation and experimental results show that the proposed method can identify the faulty switch exactly and quickly.

Keywords

Fault diagnosis T-type inverter three-level open-circuit fault phase voltage residual voltage

1. Introduction

Multilevel inverters have been increasingly used in a wide range of power applications. Compared to traditional two-level inverters, multilevel inverters can get more levels, have fewer harmonics, and reduce the voltage stress requirements of the device [1, 2, 3, 4, 5]. However, these advantages are mainly obtained by series or parallel connection of switching devices, which leads to an increase in the number of switching devices in the circuit, thereby reducing the reliability of the entire inverter. As long as one of these switching devices fails, the entire system will fail or even cause serious consequences [6]. Therefore, fault diagnosis research for multi-level inverters is very necessary now.

The topology of multilevel inverters can usually be classified into a neutral-point-clamped type, a cascade type, and a flying-capacitor type. In general, these multi-level inverters are always used in high-power applications, but in order to obtain the advantages of multi-level inverters, applications in small and medium-power are more and more popular [7, 8, 9].

In recent years, a new inverter topology named T-type neutral-point-clamped inverter (TNPC) is proposed [10]. The topology is mainly used in high-efficiency applications such as solar and electric vehicle inverters. The main circuit topology is shown in Fig. 1. This topology is based on the traditional two-level inverter, adding three pairs of bidirectional active switching devices connected to the input voltage midpoint “o”. As shown in Fig. 1, $S_{x1}$ / $D_{x1}$ ( $x=a$ , $b$ , and $c$ ) is named high-side switching device, and $S_{x4}$ / $D_{x4}$ ( $x=a$ , $b$ , and $c$ ) is low-side switching device. Compared with the diode neutral-point-clamped (DNPC) topology [1], since one switching device is reduced on the high-side and the low-side, the conduction loss of the device is reduced. Hence, TNPC has the advantages of both traditional two-level inverters and the characteristics of the multi-level inverters mentioned above [10].

Figure 1.

Schematic of the three-level T-type inverter topology.

In order to improve the reliability of the circuit system, many fault diagnosis and fault tolerance methods have been proposed for various types of converters [11, 12, 13, 14]. Switching device faults in the converter can be divided into two major categories: short-circuit and open-circuit. Among them, the short-circuit fault will generate a large current in an instant, so it is more difficult to diagnose in practice and the main fault diagnosis methods include adding a fuse or other hardware circuit in the circuit [14]. Relatively, the open-circuit fault has no such serious consequences of short-circuit fault. When an open-circuit fault occurs, the output voltage or current will be distorted. The circuit can still work, but the performance will be reduced, so the current fault diagnosis research of the converter is mainly focus on open-circuit faults [11, 12, 13].

At present, some research results of fault diagnosis for T-type converter have been proposed [15, 16, 17, 18, 19]. In [15, 16] diagnose faulty switching devices by analyzing the average of the phase currents and the midpoint potential of the input voltage. In [17], the phase angle of the input AC voltage when the phase current crossing zero is used to determine the device in which an open-circuit fault has occurred. In [18], the output voltage is converted into a histogram according to a certain algorithm, and then the center of mass is calculated, and the faulty device is judged according to the calculated value. In [19], switching fault of a single-phase T-type inverter is diagnosed by analyzing the level of the output voltage and the corresponding switch state. In the paper [20], the proposed fault diagnosis method is achieved by measuring the bridge voltage and its duration time. An on-line nonintrusive diagnostic method for detecting open-circuit switch faults in T-Type multilevel converters is introduced in [21] and the principle of this method is based on monitoring the abnormal variations of the dc-bus neutral-point current in combination with the existing information on instantaneous switching states and phase currents. However, there are not many research results on the fault diagnosis of the T-type three-level inverter. In addition, some of the proposed methods need to simultaneously detect both voltage and current [15, 16, 17], or two kind of currents in different point [21]. An additional fault diagnosis circuit is needed for each phase of the T-type inverter in the paper [20]. Complex algorithms are required in other fault detecting method [18].

In this paper, a fault diagnosis method focus on open-circuit switching device of a T-type three-level inverter is proposed. The principle of the method is based on phase voltage residual. The output phase voltage of the circuit is detected and compared with the reference phase voltage. The open-circuit faulty device of the circuit is detected and located by analyzing the difference between these two voltages. The second section introduces the simple working principle of the T-type three-level inverter. The third section analyzes the output voltage of the T-type three-level inverter under normal fault working conditions and proposes a method of fault diagnosis. The forth section gives the experimental verification results. A comparison with other proposed methods is given in section five. Finally, the conclusions of the paper are given.

2. Principle of the T-type three-level inverter

As shown in Fig. 1, the voltage $V_{\rm xo}$ ( $x=a$ , $b$ , and $c$ ) between the midpoint of each phase bridge and the midpoint of the input DC voltage is named bridge voltage. The switching state of the circuit, the switching commands for each device and the output bridge voltage are shown in Table 1 [10]. The switch states include an output positive voltage state [P], an output zero voltage state [O], and an output negative voltage state [N]. In each state, four switching devices $S_{\rm x1}\sim S_{\rm x4}$ of each phase are controlled to be in “ $o n$ ” or “off”, and a corresponding output bridge voltage will be obtained.

Table 1
Switching state, switching command and bridge voltage of T-type three-level inverter

Switching state	Switching commands ( $x=a$ , $b$ , and $c$ )				Bridge voltage
	$S_{\rm x1}$	$S_{\rm x2}$	$S_{\rm x3}$	$S_{\rm x4}$	$V_{\rm xo}$
P	On	On	Off	Off	$+V_{\rm dc}/2$
O	Off	On	On	Off	0
N	Off	Off	On	On	$-V_{\rm dc}/2$

Taking phase “ $a$ ” as an example, the typical output bridge voltage and the corresponding reference sine wave voltage are shown in Fig. 2. Here, $V_{\text{refa}}$ is the reference voltage, $V_{\rm c1}$ and $V_{\rm c2}$ are carrier voltages for phase “ $a$ ” of the inverter. When the reference voltage $V_{\text{refa}}>0$ , the switching states of the bridge voltage $V_{\rm ao}$ include [P] and [O], that is, the bridge voltage should be switched between “ $+V_{\rm dc}/2$ ” and “0”. When the reference voltage $V_{\text{refa}}<0$ , the switching states of the bridge voltage include [N] and [O], that is, the bridge voltage should be switched between “ $-V_{\rm dc}/2$ ” and “0”.

Figure 2.

PWM modulation strategy of phase “ $a$ ”.

After obtaining three bridge voltages $V_{\rm ao}$ , $V_{\rm bo}$ and $V_{\rm co}$ , according to the basic principle of the three-phase inverter, the line voltage and the phase voltage of the T-type three-level inverter can be easily obtained. The line voltages are shown in Eq. (1) and the phase voltages are shown in Eq. (2).

$\displaystyle\begin{cases}V_{\rm ab}=V_{\rm ao}-V_{\rm bo}\\ V_{\rm bc}=V_{\rm bo}-V_{\rm co}\\ V_{\rm ca}=V_{\rm co}-V_{\rm ao}\end{cases}$ (1) $\displaystyle\begin{cases}V_{\rm an}={\displaystyle\frac{2}{3}}V_{\rm ao}-{% \displaystyle\frac{1}{3}}(V_{\rm bo}+V_{\rm co})\\ V_{\rm bn}={\displaystyle\frac{2}{3}}V_{\rm bo}-{\displaystyle\frac{1}{3}}(V_{% \rm ao}+V_{\rm co})\\ V_{\rm cn}={\displaystyle\frac{2}{3}}V_{\rm co}-{\displaystyle\frac{1}{3}}(V_{% \rm ao}+V_{\rm bo})\end{cases}$ (2)

3. Proposed fault diagnosis method

When an open-circuit fault occurs in any device in the inverter, it will directly affect current paths of the inverter and the bridge voltage of the corresponding phase. According to Eqs (1) and (2), the phase voltage and line voltage of the inverter can be expressed by three bridge voltages. Hence, it is necessary to perform detailed analysis of the bridge voltage when an open-circuit fault occurs. This paper only analyzes the open-circuit fault of a single switching device, and assumes that when the switching device has an open-circuit fault, its anti-parallel diode still work normally. Taking the symmetry of the TNPC into account, the following analysis only takes phase “ $a$ ” as an example.

Figures 3 and 4 show current paths of a T-type three-level inverter under different phase current directions. In normal condition, when the phase output current $i_{\rm a}$ is positive (the current direction in Fig. 3 is the reference direction), Fig. 3a–c are the switching states [P], [O], and [N], respectively. And when the phase current $i_{\rm a}$ is negative, Fig. 4a–c show the switch states [P], [O], and [N], respectively.

Figure 3.

Current paths under different switching states when $i_{\rm a}>0$ . (a) [P] state in normal; [O] state when $S_{\rm a1}$ open-circuit. (b) [O] state in normal; [N] state when $S_{\rm a2}$ open-circuit. (c) [N] state.

Figure 4.

Current paths under different switching states when $i_{\rm a}<0$ . (a) [P] state. (b) [O] state in normal; [P] state when $S_{\rm a3}$ open-circuit. (c) [N] state in normal; [O] state when $S_{\rm a4}$ open-circuit.

When the open-circuit fault of $S_{\rm a1}$ occurs, the normal current path of Fig. 3a will not be obtained. Therefore, when the phase current is positive, the [P] state cannot be obtained and the switching state will be changed to [O]. When the phase current is negative, there is no change in any switching states just shown in Fig. 4. As a result, the bridge voltage $V_{\rm ao}$ cannot obtain “ $+V_{\rm dc}$ /2” level. According to Eq. (2), three phase voltages will be affected and the voltage of the phase “ $a$ ” $V_{\rm an}$ will be affected most seriously. Figure 5a shows the simulation waveform of the bridge voltage and phase voltage when the open fault occurs in $S_{\rm a1}$ . It can be seen that the bridge voltage is a standard three-level waveform before the fault, and the [P] state is not available when $i_{\rm a}$ is positive after the open-circuit fault occurs.

Figure 5.

Bridge voltage and phase voltage when open-circuit fault of different switching devices. (a) $S_{\rm a1}$ open-circuit; (b) $S_{\rm a2}$ open-circuit; (c) $S_{\rm a3}$ open-circuit; (d) $S_{\rm a4}$ open-circuit.

If an open circuit fault occurs in $S_{\rm a2}$ , according to Fig. 3b, the [O] state when the current is positive will not be obtained and it is replaced by state [N]. As a result, the bridge voltage will change from a three-level waveform to a two-level waveform in this case. The simulation waveforms of $S_{\rm a2}$ open-circuit fault are shown in Fig. 5b. It can be seen that the bridge voltage becomes a two-level waveform when phase current is positive after the fault occurs, and the [O] state is missing.

If the open fault occurs in $S_{\rm a3}$ , the situation is similar to the open fault in $S_{\rm a2}$ . The switching state will change from [O] to [P] if the phase current is negative as shown in Fig. 4b. And the bridge voltage will become a two-level waveform too. The simulation waveforms include bridge voltage and phase voltage are shown in Fig. 5c.

In the last case, when an open circuit fault occurs in $S_{\rm a4}$ , it can be seen from Fig. 4c that the switching state [N] will be replaced by [O] if the phase current is negative. There will be no “ $-V_{\rm dc}/2$ ” level in the bridge voltage and it will keep in “0” voltage level. The bridge voltage and phase voltage waveforms when open-circuit fault of $S_{\rm a4}$ occurring are shown in Fig. 5d.

It can be seen that when the open-circuit fault occurs, bridge voltage and phase voltage are both affected. Under open-circuit fault occurs in different devices, the switching state of TNPC and the level of the bridge voltage are shown in Table 2.

Table 2

Switch state and bridge voltage under different fault switch

Faulty switch	Switching state		Bridge voltage level
	$i_{\rm a}>0$	$i_{\rm a}<0$	$i_{\rm a}>0$	$i_{\rm a}<0$
None	[P, o, n]	[P, o, n]	$+V_{\rm dc}/2$ , 0, $-V_{\rm dc}/2$	$+V_{\rm dc}/2$ , 0, ${\rm g}-v_{\rm dc}/2$
$S_{\rm a1}$	[O, n]	[P, o, n]	0, ${\rm g}-v_{\rm dc}/2$	$+V_{\rm dc}/2$ , 0, ${\rm g}-v_{\rm dc}/2$
$S_{\rm a2}$	[P, n]	[P, o, n]	$+V_{\rm dc}/2$ , ${\rm g}-v_{\rm dc}/2$	$+V_{\rm dc}/2$ , 0, ${\rm g}-v_{\rm dc}/2$
$S_{\rm a3}$	[P, o, n]	[P, n]	$+V_{\rm dc}/2$ , 0, ${\rm g}-v_{\rm dc}/2$	$+V_{\rm dc}/2$ , ${\rm g}-v_{\rm dc}/2$
$S_{\rm a4}$	[P, o, n]	[P, o]	$+V_{\rm dc}/2$ , 0, ${\rm g}-v_{\rm dc}/2$	$+V_{\rm dc}/2$ , 0

Taking phase voltage as the research object, the phase voltage is measured and compared with the reference voltage. Then the voltage residual between these two voltages is analyzed to detect and locate the faulty device. In order to eliminate the interference caused by the sudden change of voltage level, both voltages are obtained after filtering. The block diagram of the proposed fault diagnosis method is shown in Fig. 6. In the figure, $V_{\rm an}$ and $V_{\text{refa}}$ are the phase voltage and reference voltage of phase “ $a$ ” respectively. Both voltages are normalized and the residual voltage $\Delta V_{\rm an}$ which could be calculated by Eq. (3) is obtained after comparison. Under normal working conditions, this residual voltage should be zero. If there is an open-circuit fault occurring in TNPC, the residual voltage will not keep in zero and the diagnosis result can be acquired based on some rules.

$\displaystyle\Delta V_{\rm an}=V_{\text{refa}}-V_{\rm an}$ (3)

Figure 6.

Block diagram of fault diagnosis method for open-circuit fault.

After filter and amplifier, amplitude of the reference voltage and phase voltage are both adjusted to 5 V under normal working condition. The residual voltages when different switching devices under open-circuit fault can be obtained, as shown in Fig. 10. According to the average value of the waveform, these four residual voltages can be divided into two groups, wherein the average values of Fig. 7a and b are positive, and the average values of Fig. 7c and d are negative.

Table 3

Fault diagnosis rules for TNPC under switching device open-circuit fault

Faulty device	Average value of residual voltage	$\Delta V_{\textit{PN}}$
$S_{\rm a1}$	$>$ 0	$>V_{\rm th}$
$S_{\rm a2}$	$>$ 0	$<V_{\rm th}$
$S_{\rm a3}$	$<$ ${\rm w}$	$>V_{\rm th}$
$S_{\rm a4}$	$<$ 0	$<V_{\rm th}$

Figure 7.

Residual voltages when different switching devices under open-circuit fault. (a) $S_{\rm a1}$ open-circuit; (b) $S_{\rm a2}$ open-circuit; (c) $S_{\rm a3}$ open-circuit; (d) $S_{\rm a4}$ open-circuit.

As can be seen from Fig. 7, the two residual voltages of each group can be distinguished according to the difference between their positive and negative peaks, as shown in the Eq. (4).

$\displaystyle\Delta V_{PN}=|{|{V_{\rm pos}}|-|{V_{\rm neg}}|}|$ (4)

where $V_{\rm pos}$ represents the positive peak of the residual voltage, $V_{\rm neg}$ represents the negative peak of the residual voltage. A voltage threshold $V_{\rm th}$ can be set to distinguish two residual voltages within one group.

Based on the above analysis, the following fault diagnosis rules can be obtained as shown in Table 3.

4. Experimental results

An experimental circuit of T-type three-level inverter is set up. The main power device of the inverter is IRG4PC40U and the inverter is controlled by a DSP board with TMS320F2812. The input DC voltage is 100 V. The load of each phase is a resistor of 8 $\Omega$ and an inductance of 20 mH connecting in series. The simulation of the open-circuit fault of each device is achieved by removing the drive signal from the device controlled by TMS320F2812. The reference voltage of each phase is acquired from the DSP and the phase voltage is sampled by the voltage sensor LV25-P from the inverter. With the open-circuit faults from $S_{\rm a1}$ to $S_{\rm a4}$ , the experimental waveforms are given in Fig. 8.

In Fig. 8, the waveform in channel-1 is the reference voltage of phase “ $a$ ”, channel-2 the phase voltage after amplifier and filter, channel-3 the residual voltage, and channel-4 the fault diagnosis result signal. Each waveform in the figure has a $y$ -axis of 5 V/div and a time axis of 10 ms/div. It can be seen from Fig. 8 that when the TNPC works normally, the residual voltage in channel-3 is almost equal to zero. When there is an open-circuit fault in any device, the residual voltage will have a relatively large fluctuation. Here, if the residual voltage between $V_{\text{refa}}$ and $V_{\rm anlc}$ exceeds $\pm$ 2 V, it indicates that an open-circuit fault has occurred. Then, one-cycle of the residual voltage is sampled, and the average value and of the $\Delta V_{\textit{PN}}$ in Eq. (4) are both calculated to determine which device is under fault according to the fault diagnosis rules of Table 3. In the figure, $T_{d}$ is the moment when an open-circuit fault of a specific device is detected.

Table 4 shows the experimental data and diagnosis results under different fault mode. The average value of the residual voltage is obtained as follows: after signal preprocessing (operational amplifier circuit, etc.) and high-speed AD circuit, the residual voltage is sampled point by point and then integral calculation is carried out in DSP chip to obtain the average value of voltage. In the experimental, the voltage threshold $V_{\rm th}$ is set to 2.5 V.

Table 4
Experimental data under different fault mode with the resistor of 8 $\Omega$

Fault mode	Average value of residual voltage	$\Delta v_{\textit{PN}}$	Fault diagnosis result
1	0.86 V	3.52 V	$S_{\rm a1}$ open-circuit
2	0.64 V	1.25 V	$S_{\rm a2}$ open-circuit
3	$-$ 0.64 V	1.27 V	$S_{\rm a3}$ open-circuit
4	$-$ 0.87 V	3.68 V	$S_{\rm a4}$ open-circuit

If the resistors of each phase are changed, the diagnosis results are shown in Tables 5 and 6. The resistor in Table 5 is 16 $\Omega$ and the resistor in Table 6 is 32 $\Omega$ .

Table 5

Experimental data under different fault mode with the resistor of 16 $\Omega$

Fault mode	Average value of residual voltage	$\Delta v_{\textit{PN}}$	Fault diagnosis result
1	0.91 V	3.58 V	$S_{\rm a1}$ open-circuit
2	0.59 V	1.04 V	$S_{\rm a2}$ open-circuit
3	$-$ 0.60 V	1.04 V	$S_{\rm a3}$ open-circuit
4	$-$ 0.91 V	3.64 V	$S_{\rm a4}$ open-circuit

Table 6

Experimental data under different fault mode with the resistor of 32 $\Omega$

Fault mode	Average value of residual voltage	$\Delta v_{\textit{PN}}$	Fault diagnosis result
1	0.93 V	3.50 V	$S_{\rm a1}$ open-circuit
2	0.58 V	0.76 V	$S_{\rm a2}$ open-circuit
3	$-$ 0.59 V	0.80 V	$S_{\rm a3}$ open-circuit
4	$-$ 0.94 V	3.39 V	$S_{\rm a4}$ open-circuit

5. Comparison of diagnosis methods

Table 7 shows the main features of the proposed fault diagnosis method and the other diagnosis methods presented in other literatures. In column of complexity, ‘S’ means simple and ‘C’ means complex.

Table 7
Comparison of the diagnosis methods

Literature	Method proposed	Complexity	Cost
Refs [15, 16]	Average of the phase currents and the midpoint potential of the input voltages	Hardware: C software: S	High
Ref [17]	Phase angle of the input AC voltages when the phase currents crossing zero	Hardware: C software: S	High
Ref [18]	Histogram converted from output voltages and calculation of its mass center	Hardware: S software: C	Low
Ref [19]	Analyzing the level of the output voltage and the corresponding switch state	Hardware: S software: S	Low
Ref [20]	Measuring the bridge voltage and its duration time	Hardware: C software: S	High
Ref [21]	Monitoring the dc-bus neutral-point current in combination with the switching states and phase currents	Hardware: C software: C	High
This paper	Analyzing the residual voltage between of the referent voltage and phase voltage	Hardware: S software: S	Low

Figure 8.

Experimental waveforms when open-circuit fault of different switching devices. (a) $S_{\rm a1}$ open-circuit; (b) $S_{\rm a2}$ open-circuit; (c) $S_{\rm a3}$ open-circuit; (d) $S_{\rm a4}$ open-circuit.

There is one voltage sensor and three current sensors are needed in the papers [15, 16]. In paper [17], three voltage sensors and three current sensors are needed in the proposed method. In paper [20], three voltage sensors and three additional fault diagnosis circuits are needed for fault diagnosis. In paper [21], four current sensors are needed in the paper and these currents must be analyzed with the switching states. As a result, hardware systems of the papers [15, 16, 17, 20, 21] are complex and their costs are high. However, only three voltage sensors are needed in the papers [18, 19] and this paper because the reference voltages can be obtained from the control system. So, hardware systems of the papers [18, 19] and this paper are simple and their costs are low.

6. Conclusions

In this paper, a method for fault diagnosis of T-type three-level inverter is proposed, which is verified by an experimental system. The main conclusions are as follows.

1) 1)

When there is an open-circuit fault occurs in TNPC, some current path of the inverter cannot be realized. As a result, the bridge voltage will miss some voltage level and the corresponding phase voltage will be affected too.

Phase voltage of the TNPC contains enough fault information of the inverter. Through theory analysis and experimental verification, the method based on the residual voltage between the reference voltage and the phase voltage can locate the specific faulty device.

The reference voltage of the TNPC can be directly obtained from the control system. Therefore, only three voltage sensors are needed in the method proposed in this paper to acquire phase voltages of the TNPC and it is easy to implement.

Footnotes

Acknowledgments

This work was supported in part by the Zhejiang Province Public Welfare Technology Application Research Project under Grant LGF20F010004, Ningbo Municipal Natural Science Foundation under Grant 2018A610189 and Zhejiang Wanli University Initial Scientific Research Fund for PhD.

References

Wang

Zhang

and Xie

, A novel SVPWM for three-level NPC inverter based on m-mode controllability, IEEE Transactions on Industrial Electronics 65(8) (2018), 6055–6065.

Chen

Liu

Zhang

Zheng

and Zhang

, Detection of IGBT degradation in NPC inverter based on infrared thermography, Journal of Computational Methods in Sciences and Engineering 18(2) (2018), 459–468.

Karwatzki

and Mertens

, Generalized control approach for a class of modular multilevel converter topologies, IEEE Transactions on Power Electronics 33(4) (2018), 2888–2900.

Dekka

and Zargari

N.R.

, Start-up operation of a modular multilevel converter with flying capacitor submodules, IEEE Transactions on Power Electronics 32(8) (2017), 5873–5877.

Akagi

, Multilevel converters: Fundamental circuits and systems, Proceedings of the IEEE 105(11) (2017), 2048–2065.

Ceballos

Pou

Robles

Zaragoza

and Martin

J.L.

, Performance evaluation of fault-tolerant neutral-point-clamped converters, IEEE Transactions on Industrial Electronics 57(8) (2010), 2709–2718.

Yang

Wang

Cecati

and Xu

, Asymmetric control with hybrid/full bridge modular multilevel converter for low frequency low voltage drive application, 42nd Annual Conference of the IEEE Industrial Electronics Society (IECON), Florence, Italy (2016), 6465–6469.

Beig

A.R.

and Dekka

, Experimental verification of multilevel inverter-based standalone power supply for low-voltage and low-power applications, IET Power Electronics 5(6) (2012), 635–643.

Mukherjee

Giri

S.K.

and Banerjee

, A flexible discontinuous modulation scheme with hybrid capacitor voltage balancing strategy for three-level npc traction inverter, IEEE Transactions on Industrial Electronics 66(5) (2019), 3333–3343.

10.

Schweizer

and Kolar

J.W.

, Design and implementation of a highly efficient three-level T-type converter for low-voltage applications, IEEE Transactions on Power Electronics 28(2) (2013), 899–907.

11.

Shi

Wang

and Xu

, Fault diagnosis and tolerant control of single IGBT open-circuit failure in modular multilevel converters, IEEE Transactions on Power Electronics 31(4) (2016), 3165–3176.

12.

Amini

and Moallem

, A fault-diagnosis and fault-tolerant control scheme for flying capacitor multilevel inverters, IEEE Transactions on Industrial Electronics 64(3) (2017), 1818–1826.

13.

Yang

Tang

and Wang

, Seamless fault-tolerant operation of a modular multilevel converter with switch open-circuit fault diagnosis in a distributed control architecture, IEEE Transactions on Power Electronics 33(8) (2018), 7058–7070.

14.

Alavi

Wang

and Luo

, Short-circuit fault diagnosis for three-phase inverters based on voltage-space patterns, IEEE Transactions on Industrial Electronics 61(10) (2014), 5558–5569.

15.

Choi

U.M.

Lee

K.B.

and Blaabjerg

, Diagnosis method of an open-switch fault for a grid-connected T-type three-level inverter system, IEEE 3rd International Symposium on Power Electronics for Distributed Generation Systems (PEDG), Aalborg, Denmark (2012), 470–475.

16.

Choi

U.M.

Lee

K.B.

and Blaabjerg

, Diagnosis and tolerant strategy of an open-switch fault for T-type three-level inverter systems, IEEE Transactions on Industry Applications 50(1) (2014), 495–508.

17.

Lee

J.S.

and Lee

K.B.

, An open-switch fault detection method and tolerance controls based on SVM in a grid-connected T-type rectifier with unity power factor, IEEE Transactions on Industrial Electronics 61(12) (2014), 7092–7104.

18.

Pires

V.F.

Foito

and Amaral

T.G.

, Fault detection and diagnosis in a PV grid-connected T-type three level inverter, International Conference on Renewable Energy Research and Applications (ICRERA), Palermo, Italy (2015), 933–937.

19.

Becker

Poure

Jamshidpour

and Saadate

, Open-switch fault diagnosis for five-level H-bridge neutral point piloted or T-type converters, IEEE 16th International Conference on Environment and Electrical Engineering (EEEIC), Florence, Italy (2016), 1–6.

20.

Chen

Liu

and Zheang

, Open-circuit fault diagnosis method for the T-type inverter based on analysis of the switched bridge voltage, IET Power Electronics 12(2) (2019), 295–302.

21.

Demerdash

N.A.O.

Weise

and Katebi

, A fast on-Line diagnostic method for open-circuit switch faults in SiC-MOSFET-based T-type multilevel inverters, IEEE Transactions on Industry Applications 53(3) (2017), 2948–2958.

Diagnosis method of open-circuit fault for T-type three-level inverter based on residual voltage

Abstract

Keywords

1. Introduction

Table 1 Switching state, switching command and bridge voltage of T-type three-level inverter

Table 4 Experimental data under different fault mode with the resistor of 8 Ω

Table 7 Comparison of the diagnosis methods

Footnotes

Acknowledgments

References

Table 1
Switching state, switching command and bridge voltage of T-type three-level inverter

Table 4
Experimental data under different fault mode with the resistor of 8 $\Omega$

Table 7
Comparison of the diagnosis methods