Internal fault analysis and detection method of the ‘unit-form’ microgrid

Abstract

The microgrids can provide sustainable supply to the important power users. However, the internal fault detection methods are not mature yet. A kind of microgrid topology is defined to decide the protection configuration. For a microgrid with inverter-based distributed generation (IBDG), the transient characteristics of typical faults are studied. And a fault detection method is put forward based on statistical analysis. All above may contribute to the research on the protection strategy for the microgrid.

Keywords

Microgrid protection fault detection statistical analysis microgrid topology

1. Introduction

Due to developments of distributed generation (DG) technology, the microgrid is considered as the important part of power distribution. It can not only promote the deep applications of the clean energy, but also make it possible to ensure continuous power supply for important loads in power system [1].

There are two operation modes for a microgrid, namely grid-connected and islanding. IEEE Std.1547-2003 [2] notes: it is an urgent task that DGs are involved in a planned islanding operation. At present, the main power sources in a microgrid include wind turbines, solar photovoltaic batteries, fuel cells, micro turbines, and flywheel storage devices, etc. Most of them are based on IBDGs. Moreover, the maximum Ampere of output current of IBDG is limited, which is usually 2 times the rated current. Therefore it is important to study the transient fault characteristics in a microgrid based on IBDG and the protection strategy [3, 4].

2. Theoretical analysis

2.1 The protection situation and strategy of mircogrid

From the perspective of communication, the protection can be divided into using only local information [5, 6, 7, 8] and using the information obtained through the communication equipment [9, 10, 11, 12, 13, 14]. From the point of protected object, there are the single component protection [8, 9, 10, 11, 12, 13, 14] and the wide area protection [15, 16, 17]. In this paper, from the perspective of whether the use of the pattern recognition, artificial intelligence and other tools, there are two kinds of protection scheme. One is mainly based on the electrical mechanism [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17], such as traditional overcurrent protection, low-voltage protection, zero sequence current protection, and distance protection, etc; The other is mostly based on the machine learning [14, 18, 19, 20], such as expert system, neural networks, random forests, etc.

The former scheme is mainly based on the electrical mechanism. The existing typical microgrid protection strategies are low-voltage protection and current differential protection. If the electrical parameters are appropriately set, the basic requirements of protection can be satisfied as much as possible. However there is a correlation between these electrical parameters. So those choices based on experience are often subjective. If using some statistical analysis tools, the optimal parameters would be selected, and the protection would be more perfect.

The latter scheme is mainly based on pattern recognition, at the same time, using certain mathematical algorithm, such as graph theory [21], a theory of root tree, neural network, etc. This protection treats the fault detection as a nonlinear system to process. It requires a large number of training and testing samples. And it is difficult to meet the four basic requirements of the relaying protection, especially the speed.

The advantages of both of the schemes are considered in this paper.

2.2 Microgrid model

2.2.1 Microgrid topology definition

No matter from the perspective of the physical connection or the electrical connections, the load is the end of the microgrid even the power system. So, it’s similar to the traditional power grid that the power flow has one direction for a load. And it remains the same whenever in normal or fault status. Each of the loads is equipped with protection devices. Therefore the fault of load and its protection performance are not considered in this paper.

There are about three kinds of grounding ways in the microgrid: TN, TT, IT. From the perspective of research protection, and considering whether it can provide enough fault current or not, TN-C-S is adopted in order to start the overcurrent protection device as soon as possible [22].

Figure 1.

Topological diagram of ‘unit-form’ microgrid.

In order to study the fault transient characteristics of microgrid and its protection configuration, a microgrid topology structure is defined scientifically. A ‘unit-form’ microgrid topology is proposed, as shown in Fig. 1. It’s completed through three definitions.

Definition 1: The main feeder is considered as an independent unit (as shown in Fig. 1, expressed by ‘Unit 1’). There is a fast circuit breaker at the beginning of the main feeder, which is near the main grid. The main feeder is connected to main grid by PCC (Point of Common Coupling).

Definition 2: Each unit is constituted by a pair of micro source and load, and each unit has its own protection and corresponding circuit breaker, such as CB1, CB2 and CB3 shown in Fig. 1. It’s certainly allowed some unit to have only micro source or only load.

Definition 3: The modes of the unit accessing the main feeder are as follows:

(1)

There is directly paralleled a unit. It’s suggested that the important load should keep such connection to improve the reliability of power supply.

(2)

Two units are paralleled to the same location in the main feeder (as shown in Fig. 1, ‘Unit 2’ and ‘Unit 3’).

Such a topology definition is of the following advantages:

The microgrid structure is more clarity and easier to manage.

It is helpful with the protection configuration, the control of circuit breaker and the connection or disconnection of micro source, etc.

When islanding operation happens after microgrid fault in elsewhere but internal, it ensures the continuous supply of important load. It can improve the power supply reliability.

2.2.2 The micro-source control strategies and methods

The control of output voltage and frequency of IBDG power supply is completely different from the conventional synchronous generator. It is decided by the control strategies of power electronic interface. Therefore, the model requires a reasonable choice control strategy of interface inverter, which is very important for the normal operation of the micro-sources and microgrid.

Currently there are three kinds of classical control methods of IBDG [23]: constant power control, droop control, constant voltage and constant frequency control.

3. The simulation and experimental results

According to the experience of a large number of literatures, preliminarily five classic electric parameters are selected as the test object in this paper. By building microgrid model, the difference of five electric parameters is observed under normal and various kinds of faults.

Five electrical parameters were initially selected as phase voltage $\dot{V}_{abc}$ , phase current $\dot{I}_{abc}$ , zero-sequence current $\dot{I}_{0}$ , negative sequence current $\dot{I}_{2}$ and differential current $\dot{I}_{d}$ . Observations were its magnitude, that $V_{abc}$ , $I_{abc}$ , $I_{0}$ , $I_{2}$ , $I_{d}$ .

Among them, if the three-phase voltage are $\dot{V}_{a}$ , $\dot{V}_{b}$ , $\dot{V}_{c}$ , the three-phase current are $\dot{I}_{a}$ , $\dot{I}_{b}$ , $\dot{I}_{c}$ , and the neutral current is $\dot{I}_{n}$ , and

$\displaystyle\left[\begin{array}[]{c}\dot{I}_{1}\\ \dot{I}_{2}\\ \dot{I}_{0}\\ \end{array}\right]=\frac{1}{3}\left[\begin{array}[]{ccc}1&\alpha l&{\alpha^{2}% }\\ 1&{\alpha^{2}}\hfill&\alpha\\ 1&1\hfill&1\\ \end{array}\right]\left[\begin{array}[]{c}\dot{I}_{a}\\ \dot{I}_{b}\\ \dot{I}_{c}\\ \end{array}\right]，\text{and}\ \alpha=e^{j120^{\circ}}$

Then the five electrical parameters can be expressed as follows:

$\displaystyle V_{abc}=\max(|\dot{V}_{a}|,|\dot{V}_{b}|,|\dot{V}_{c}|)$ $\displaystyle I_{abc}=\max(|\dot{I}_{a}|,|\dot{I}_{b}|,|\dot{I}_{c}|)$ $\displaystyle I_{0}=\frac{1}{3}|\dot{I}_{a}+\dot{I}_{b}+\dot{I}_{c}|$ $\displaystyle I_{2}=\frac{1}{3}|\dot{I}_{a}+\alpha^{2}\dot{I}_{b}+\alpha\dot{I% }_{c}|$ $\displaystyle I_{d}=\left|\sum\limits_{k=a,b,c,n}\dot{I}_{k}\right|$

Figure 2.

Diagram of the ‘unit-form’ topology of a benchmark European microgrid.

3.1 The microgrid model

On the simulation platform of PSCAD/EMTDC, the two kinds of typical control strategies are involved with IBDG. And on this basis the model of a benchmark European 400 v microgrid is built. According to this microgrid topology definition of “unit-form” as previously mentioned, the standard six units are defined in this model, as shown in Fig. 2.

The parameters of load and micro source are shown in Tables 1 and 2 respectively.

Table 1
Values of the power of the loads in constant power mode in Fig. 2

Load	$L_{1}$	$L_{2}$	$L_{3}$	$L_{4}$	$L_{5}$	$L_{\Sigma}$
P (kW)	4.845	48.45	19.55	1.620	21.25	95.715
Q (kVar)	3.021	30.21	12.19	1.007	13.25	59.678

Table 2

Values of the power of the sources in Fig. 2

Souce	MG ${}_{1}$	MG ${}_{2}$	MG ${}_{3}$	MG ${}_{4}$	MG ${}_{5}$	MG ${}_{\Sigma}$
P (kW)	30	30	20	3	10	93

3.2 The waveform of various electric parameters under normal condition or while several typical faults occur

Simulation time: 0 s–0.5 s, fault during time: 0.3 s–0.36 s. Note that there are 3 fault locations set as F1, F2, F3 respectively as shown in Fig. 2. And there are 5 fault types: 1-Single Line-to-Ground (SLG) faults, 2-Line-to-Line (LL) faults, 3-three phase faults, 4-Line-to-Line-to-Ground (LLG) faults, 5-three phase to ground faults. And all the experiments were under the islanding operation mode for the microgrid model.

4. Statistical analysis and conclusion

Values of the 5 electric parameters under the conditions of grid-connecting are listed in the Table 3. And based on these the histograms are formed to analysis and compare the differences of the values in between normal and faulted condition. In addition, “0” denoted as normal operation. All voltage unit appeared in this section is kV, and the current unit is kA.

4.1 Phase voltage $V_{abc}$

Table 3
Values of $V_{abc}$ under 16 different conditions

Fault loction	0	1	2	3	4	5
F1	0.32	0.46	0.46	0.40	0.35	0.37
F2	0.32	0.40	0.40	0.37	0.32	0.34
F3	0.32	0.44	0.47	0.40	0.34	0.36

Figure 3.

Histograms based on Table 3.

As seen from Fig. 3 regardless of the fault location or the fault type, the maximum values of phase voltage measured near PCC are mostly greater than the voltage value in the normal condition. By this over-voltage protection can play a role. However due to the voltage different is not too large, the reliability and sensitivity of this protection principle is not so good. For that the voltage protection can be applied as back-up protection.

4.2 Phase current

I_{abc}

In general, values of fault currents have the maximum at F1 and the minimum at F2. This is related to the distances from fault location to PCC, and also related to the power source and control strategy. The minimum fault current is twice as large as the normal phase current, namely there is a SLG fault at F2. Due to the fault current is larger than normal current, the over-current protection based on the phase current is indeed appropriate. And reliability and sensitivity of this protection are preferred. The applicability, sensitivity and reliability of over-current protection are all excellent. Thus it can be considered as the main protection.

4.3 Zero sequence current $I_{0}$

When the SLG fault, LL fault, three phase fault occur at any fault location, the zero sequence fault current is far greater than zero sequence normal current. The smallest zero sequence fault current, namely the three phase fault current is 5 times as large as normal current. The protection of zero sequence current is the most applicable to SLG fault. And in its range, this protection is of high sensitivity and reliability.

4.4 Negative-sequence current $I_{2}$

The negative sequence current is the largest, namely 3.10 kA, when LLG fault occur. It is about 30 times than the negative sequence current in normal running. The negative sequence current is smallest when three-phase to ground fault occur, about 0.325 kA. It is about three times than the negative sequence current while normal operating. Regardless of the fault location or the fault type, the negative sequence current value is greater than or equal to the normal value measured near PCC.

4.5 Differential current $I_{d}$

When the SLG and LL faults occur, the current is much greater than normal current. The minimum of differential fault current, when LL faults occur at F2, is about 3 times than normal current. When other faults occur, the value of current is similar with the normal current. So the differential protection is absolutely cannot be used for those types of fault. The differential current protection is only applicable to SLG fault and LL fault. Although there are some limitations of differential current protection, it is of high sensitivity and reliability once it can be used. So cooperated with other protection principle, it can be considered as the main protection.

5. Analysis on the application

From the above analysis, A definition of minimum difference ratio $T_{min}$ is proposed in this paper. Making a ratio by two parameters’ values which are in fault and normal conditions respectively. Note that $T_{min}=\min(f_{(i.j)})/f_{0}$ , $i=$ F1, F2, F3, which is the fault location; $j=$ 1, 2, 3, 4, 5, which is the fault type. And $f_{(i.j)}$ is the measured value of electrical parameters in fault operation; $f_{0}$ is the normal value of corresponding parameters. If the five electric parameters can reflect the corresponding fault type, their minimum difference ratios are listed in Table 4.

Table 4
Distinguishing fault ability of the five fault features

Parameters	1	2	3	4	5	$T_{min}$
$V_{abc}$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	1
$I_{abc}$	$\bullet$	$\bullet$	$\bullet$	–	–	3
$I_{0}$	$\bullet$	$\bullet$	$\bullet$	–	–	5
$I_{2}$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	3
$I_{d}$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	$\bullet$	2

Note: for all kinds of fault feature: ‘ $\bullet$ ’ is denoted capable; ‘–’ is denoted incapable.

Table 4 shows that $I_{abc}$ , $V_{abc}$ and $I_{2}$ can completely reflect all faults, whatever the fault location and fault type are. So these protections based in them all can be used as the main protection at PCC. From the perspective of the minimum difference ratio, that of negative sequence current protection is the largest. But the sensitivity of negative sequence current protection for different fault types is different. So its reliability is poorer. Comprehensive comparison, no matter from the applicable scope or reliability, sensitivity of over-current protection is the first choice of main protection. Although zero sequence current protection for fault type is limited, but in its applicable range, it has the most significant difference, so the sensitivity and reliability is the best. And the measurement way of zero sequence current has been very mature. So it is suitable to be a backup protection. In conclusion, the protection configuration in PCC is as follows: the main protection is over-current protection, and backup protection is zero sequence current protection. That can basically meet the four basic requirements of relay protection. Changing the load of mircogrid to certify the above methods, the same conclusion can be drawn.

6. Summary

In order to study mocrogrid protection for further, the “unit-form” microgrid model is defined in this paper. Under this definition, a benchmark European low-voltage microgrid model is formatted for the future protection configurations. Also the corresponding simulation model is established.

The existing protection strategies are surveyed. It is clarified as schemes mainly based on electrical mechanism and that on machine learning. To take advantages of both the protection categories’, a fault detection methods using statistical analysis on some certain typical electrical parameters is proposed. For that a great deal of samples data are to be studied. Tested by experiments, it is shown that the detection methods is simple, straightforward, and appropriate.

Footnotes

Acknowledgments

The research was sponsored by Zhejiang Province Natural Science Foundation for Youths (LQ17E070003) and Zhejiang Key Discipline of Instrument Science and Technology.

References

European Technology Smart Grid Platform. (2006). Smartgrids: Vision and strategy for European electricity networks of the future [Online]. Available: http://www/smartgrids.eu/documents/vision.pdf.

IEEE Standard for Interconnecting Distributed Resources With Electric Power Systems, IEEE Std 1547-2003.2003, pp. 1–28.

Lasseter

R.H.

et al., CERTS microgrid laboratory test bed, IEEE Transactions on Power Delivery 26(1) (2011), 325–332.

Laaksonen

H.J.

, Protection principles for future microgrids, IEEE Transactions on Power Electronics 25(12) (2010), 2910–2918.

Mahat

et al., A simple adaptive overcurrent protection of distribution systems with distributed generation, IEEE Transactions on Smart Grid 2(3) (2011), 428–437.

Oureilidis

K.O.

Gkavanoudis

S.I.

and Demoulias

C.S.

, Fault Clearing In A Converter-Dominated Microgrid With Traditional Protection Means. Power Electronics for Distributed Generation Systems (PEDG), 2014 IEEE 5th International Symposium on DOI: 10.1109/PEDG.2014.6878646Publication Year: 2014, pp. 1–7.

Bui

D.M.

et al., Investigate dynamic and transient characteristics of microgrid operation and develop a fast-scalable-adaptable algorithm for fault protection system, Electric Power Systems Research 120 (2015), 214–233.

Nikkhajoei

and Lasseter

R.H.

, Microgrid fault protection based on symmetrical and differential current components, Public Interest Energy Research, California Energy Commission, Contact No. 500-03-024, 2006.

Sortomme

Venkata

S.S.

and Mitra

, Microgrid protection using communication-assisted digital relays, IEEE Trans. Power Del 25(4) (2010), 2789–2796.

10.

Ustun

Ozansoy

and Zayegh

, Modeling of a centralized microgrid protection system and distributed energy resources according to IEC 61850-7-420, IEEE Trans. Power Syst 27(3) (Aug. 2012), 1560–1567.

11.

Zamani

M.A.

Yazdani

and Sidhu

T.S.

, A communication-assisted protection strategy for inverter-based medium-voltage microgrids, IEEE Trans. Smart Grid 3(4) (Dec. 2012), 2088–2099.

12.

Southern

Crossley

Potts

Pickering

Caunce

and Weller

, A new type of differential feeder protection relay using the global positioning system for data synchronization, IEEE Trans.Power Del 12(3) (July 1997), 1090–1099.

13.

Villamagna

and Crossley

, A symmetrical component-based gps signal failure-detection algorithm for use in feeder current differential protection, IEEE Trans. Power Del 23(4) (Oct. 2008), 1821–1828.

14.

Casagrande

et al., A differential sequence component protection scheme for microgrids with inverter-based distributed generators, IEEE Transactions on Smart Grid 5(1) (2014), 29–37.

15.

Adamiak

M.G.

et al., Wide area protection:technology and infrastructures, IEEE Transactions on Power Delivery 21(2) (2006), 601–609.

16.

et al., Wide-area protection fault identification algorithm based on multi-information fusion, IEEE Transactions on Power Delivery 28(3) (2013), 1348–1355.

17.

Dai

Wang

and Jiao

, Reliability evaluation of the communication network in wide-area protection, IEEE Transactions on Power Delivery 26(4) (2011), 2523–2530.

18.

Zhang

Wen

and LI

, Power System Fault Diagnosis with an Enhanced Fuzzy Petri Net Accommodating Temporal Constraints[J], Automation of Electric Power Systems 38(5) (2014), 66–72.

19.

Tan

J.C.

et al., An expert system for the back-up protection of a transmission network, IEEE Transactions on Power Delivery 15(2) (2000), 508–514.

20.

Negnevitsky

and Pavlovsky

, Neural networks approach to online identification of multiple failures of protection systems, IEEE Transactions on Power Delivery 20(2) (2005), 588–594.

21.

Zhao

and Sun

, A novel algorithm for fault location for low-voltage microgrids, Electric Power Science and Technology 29(9) (2013), 24–28.

22.

Liu

, Power System Fault Analysis, China Electric Power Press, 2012.2. ISBN: 9787508396910.

23.

Collotta

Tirrito

and Bobovich

A.V.

, Trasmission amplitude management in power line communications through a fuzzy logic controller: a real case study[J], Journal of Computational Methods in Sciences & Engineering 16(1) (2016).