Portal de Periódicos da CAPES

Flux‐based turn‐to‐turn fault protection for power transformers

2016; Institution of Engineering and Technology; Volume: 10; Issue: 5 Linguagem: Inglês

10.1049/iet-gtd.2015.0738

1751-8695

Mohsen Mostafaei, Farhad Haghjoo,

Electric Power Systems and Control

IET Generation, Transmission & DistributionVolume 10, Issue 5 p. 1154-1163 Research ArticleFree Access Flux-based turn-to-turn fault protection for power transformers Mohsen Mostafaei, Mohsen Mostafaei Department of Electrical Engineering, Abbaspour School of Engineering, Shahid Beheshti University, Tehran, IranSearch for more papers by this authorFarhad Haghjoo, Corresponding Author Farhad Haghjoo f_haghjoo@sbu.ac.ir Department of Electrical Engineering, Abbaspour School of Engineering, Shahid Beheshti University, Tehran, IranSearch for more papers by this author Mohsen Mostafaei, Mohsen Mostafaei Department of Electrical Engineering, Abbaspour School of Engineering, Shahid Beheshti University, Tehran, IranSearch for more papers by this authorFarhad Haghjoo, Corresponding Author Farhad Haghjoo f_haghjoo@sbu.ac.ir Department of Electrical Engineering, Abbaspour School of Engineering, Shahid Beheshti University, Tehran, IranSearch for more papers by this author First published: 01 April 2016 https://doi.org/10.1049/iet-gtd.2015.0738Citations: 32AboutSectionsPDF 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 Since the internal turn-to-turn faults (TTFs), as a common cause of the transformer failures, result in minor changes on terminal currents/voltages, they maybe undetected by the conventional protective relays. In this study, a simple, sensitive and robust linkage flux-based technique is proposed to protect the power transformers against TTFs. In this approach, some separated multi-turn windings as search coils (SCs) have to be wrapped around the transformer core legs to sense the related passing flux. Since passing equal flux through a transformer core leg in normal condition induces equal voltages in the related SCs, variation of the induced voltages indicates the fault occurrence in that phase. Against the traditional differential protection schemes, current transformer error and saturation, tap changer operation, energising inrush current, over-fluxing etc. cannot impact the proposed technique performance. Therefore, it can be introduced as a comprehensive protection technique for power transformers.To confirm the proposed technique performance, it is tested on a real 50 kVA, 20 kV/400 V three-phase distribution transformer, in various conditions, too. 1 Introduction Power transformers are one of the most important and expensive apparatus in the power systems and short circuits are an unavoidable event in power transformers that can be happened. Turn-to-turn faults (TTFs) are the most common events inside transformers. A brief review on the records of the modern transformer breakdowns shows that around 70–80% of the failures are caused by TTFs [1]. Early stages of TTF may often have negligible effects on the transformer performance; however, such faults may rapidly lead to more serious permanent forms such as phase-to-phase or phase-to-ground faults [2]. TTFs can be started by [3–5]: Metallic contact resulting from mechanical forces due to external short circuit. Severe insulation deterioration as a result of excessive overloading. Insulating oil degradation due to contamination by moisture. Applying a huge impulse voltage due to a switching/lightening overvoltage. The most common protective device for power transformers is differential relay. Though it can protect the power transformer against the severe TTFs, it is unable to detect the minor ones because the terminal currents are not affected significantly by them. Though a TTF will give rise to a heavy fault current in the short-circuited loop, the terminal currents will be very small, because of the high ratio of transformation between the whole winding and the short-circuited turns [5]. Since the current-based differential relays use the terminal current waveforms, their performance and setting are depended on some variables, including: Magnetising current. Current transformers (CTs) turn ratio error. CT saturation in heavy external fault. Tap changer (TC) operation. which all of above items reduce the protection sensitivity to detect the internal faults. Therefore, an operate-restrain characteristic [5, 6] must be considered to activate the differential relays. Indeed, this characteristic reduces the differential relay sensitivity, and the relay performance will be limited by it, while a TTF occurs. Indeed, the minor TTFs (that are located in the restrain region of the above-mentioned characteristics) cannot be detected by differential relays. Energising inrush current is another phenomenon that affects the differential relays security and causes a mal-operation. The inrush current, though generally resembling an in zone fault current, differs greatly when the waveforms are compared. The difference in the waveforms can be used to distinguish between the conditions. The inrush current contains all harmonic orders but in practice only the second harmonic is used as an attractive basis for a stabilising bias against inrush effects. To distinguish the inrush current from the faulty current, the differential current is passed through a filter that extracts the second harmonic; this component is then applied to produce a restraining quantity sufficient to overcome the operating tendency due to the whole of the inrush current that flows in the operating circuit [5]. In addition, to modify the differential relay performance, some algorithms had been presented to block or restrain the relays such as: Harmonic restraint [7, 8]. Wave-shape recognition [9, 10]. Signal processing methods such as wavelet transform [11]. The over-fluxing phenomenon as another trouble for differential relays arises principally from the following system conditions: High system voltage. Low system frequency. Geomagnetic disturbances. which the latter results in low-frequency earth currents circulating through a transmission system. Since momentary system disturbances can cause transient over-fluxing that is not dangerous, time delayed tripping is required. The normal protection is an inverse definite minimum time (IDMT) or definite time characteristic, initiated if a defined V/f threshold is exceeded. Some relays provide a fifth harmonic detection feature, which can be used to detect such a condition, as levels of this harmonic rise under over-fluxing conditions [5]. Though various current and/or voltage-based algorithms [12–19] had been presented to modify the differential relay performance for TTF detection, there are some problems yet, as below: They are insensitive to detect minor TTFs. The most of them have mal-operation in the face of inrush current, over-fluxing, CT saturation or TC operation. Some of them cannot operate properly when the related transformer feeds an unbalanced load or while the transformer is fed by an unbalanced voltage source. They cannot operate accurately due to the instrument transformers errors (because they use the instrument transformers to obtain the necessary data). Most of them cannot specify the faulty phase and/or the TTF region on the faulty winding. A brief review on the above-mentioned methods including their operation bases, advantages, disadvantages and ambiguities are listed in Table 1. Table 1. Brief review on the various methods to detect TTF in transformers on the basis of terminal current or voltage signals [12–19] Methods Ref. Operation Base Advantages Common Disadvantages Individual Disadvantages Ambiguities Negative Sequence Component [12–15] Measuring the magnitude and the phase of negative sequence component of currents. – Sensitive to detect minor TTFs. Unable to locate the fault / localize the faulty region on faulty phase Impressed by the instrument transformers error – Unable to operate in no load conditions/ unbalanced loads [20, 21]. The impact of: – TC operation – Over-fluxing – Inrush current [16] Comparing the ratio of the negative sequence components of the line currents with the transformer turns ratio. – Operates in imbalanced load/voltage source. – Needs to have TC position. The impact of: – Over-fluxing – Inrush current Zero Sequence Component [17] Measuring the zero sequence currents and voltage. – Stable against the inrush current. – Usable only for transformer banks. The impact of: – TC operation – Over-fluxing – Unbalanced load/voltage source V-I locus diagram [18] Locus diagram of input current versus differential voltage between input and output voltages. – Online technique to detect the TTFs and mechanical faults. – Needs to have the healthy condition data. The impact of: – TC operation – Over-fluxing – Inrush current – Unbalanced load/voltage source On-load exciting current Extended Park's Vector [19] Comparing between three phase exciting currents using Park's transform. – Discriminate between unbalanced load and internal fault. – Insensitive to the load Recently, a few flux-based techniques have been presented to detect the TTFs in transformers [22–26], which their operation bases, advantages, disadvantages and ambiguities are listed in Table 2. As can be seen, only the presented technique in [25, 26] by current authors can be introduced as a comprehensive technique to achieve the aim. Table 2. Brief review on the various flux-based techniques to detect TTF in transformers [22–26] Methods Ref. Operation Base Advantages Disadvantages Ambiguities Leakage Flux [22] Some sensors are located near the HV winding to sense the leakage flux during faulty conditions. – Detects the winding deformation. – Needs to change the transformer design and increase the tank dimensions to install the sensors to maintain permissible HV clearances. – Unable to localize the fault region on the faulty phase. – Unable to detect the TTFs located in the middle point of each winding. – The effect of the Over-fluxing, Inrush current and unbalanced load/voltage source on the algorithm performance is not investigated. [23,24] Operates by calculating the increments of the linkage flux according to transformer equations. – Stable against the inrush current and over-flux. – Detect the TTF in the transformer winding. – Identify the faulty phase. – Cannot detect the TTF under 10% of the winding due to measuring transformers error. – Needs to have the TC position. – The effect of the Over-fluxing, Inrush current and unbalanced load/voltage source on the algorithm performance is not investigated Linkage Flux [25,26] Linkage core flux value is sensed through some installed SCs on transformer legs, and TTFs are detected by measuring the differential value of the SCs induced voltages. – Simple and accurate; – Identify the faulty phase; – Determine the fault location in the faulty phase; – Stable and robust in the face of Inrush current, Over-fluxing, TC operation and unbalanced load. – Needs to install SCs. All of the related ambiguities are cleared in this paper. The linkage flux-based (LFB) technique, which is described and discussed in this paper for online TTF detection, can also be used to identify the faulty phase and faulty region in transformer winding as an offline repairing method [26]. The LFB technique can detect TTFs by measuring the linkage flux in various points along the core legs, and compare them to assess the flux distribution symmetricity overall the legs. The LFB technique uses some search coils (SCs), which each one is made by a few turns of tiny wire wrapped around the core, with specific intervals to measure the flux in different places. In normal and healthy condition (NHC), the core flux passing each transformer leg induces equal voltages in the all corresponding SCs (installed on the corresponding leg). TTF occurrence in each transformer winding decreases the core flux in the faulty region and reversely increases the leakage flux in that region, i.e. the flux distribution along the faulty leg will be disturbed, and accordingly unequal voltages will be induced in the related SCs. Comparing the output voltage of the opposite SCs on any phase reveals the flux linkage asymmetrically as an essential effect of TTF occurrence. Since the installed SCs in the regular places cannot detect TTFs in the middle points (MTTFs) of the phases, a modification is applied on the mentioned technique by using two additional irregular installed SCs. In this paper, LFB protection technique is explained, simulated and implemented on a real distribution transformer (Fig. 1). In addition, the effect of the magnetising inrush current and over-fluxing phenomenon is shown, experimentally. Fig. 1Open in figure viewerPowerPoint Sample transformer and installed SCs on the transformer LV winding Since the TC operation cannot change the flux symmetricity along the core leg, the performance of the proposed technique cannot be affected by it. Moreover, the LFB technique cannot be affected by the unbalanced load and/or unbalanced voltage source, because they are unable to change the flux distribution along the core leg. Details of the tested transformer (including core, windings, connections, SCs and related dimensions) are shown in Fig. 2. Fig. 2Open in figure viewerPowerPoint Dimensions of the sample transformer (core, winding and SCs) and windings connections It is notable that the proposed technique can also identify the faulty phase and the TTF location on that phase during repairing process, which is presented in the other paper [26]. 2 Fundamental of the LFB technique The proposed method in this paper is based on the magnetic flux continuity law [27] that means the magnetic fluxes are continuous and enclosed lines, which do not originate nor terminate at a point. Moreover, flux is determined by the induced or applied voltage, and when some turns are shorted, the linkage flux in them tends to reduce. In that time, the linkage flux changes to leakage flux and closes its path with air (or oil). Linkage flux in power transformers passes through both the primary and secondary windings. This sinusoidal flux (φlink) induces voltage (Eind) in any windings according to Faraday's law as (1), which N is the number of the related winding (1)In addition, there is a miniature flux in NHC that passes through the air and called the leakage flux. This flux does not create any linkage between the transformer windings, and only it reduces the voltage in the primary and secondary windings proportional to the load [28]. In fault condition, the leakage flux will have the higher magnitude in comparison with NHC. Against the NHC (that the flux lines close own paths through the core), flux lines in the faulty region tend to change own pathway and escape from the core, i.e.: The symmetrical flux lines will be disturbed. The linkage flux reduces, and inversely the leakage flux increases. Flux density varies along the faulty transformer leg. The flux density increases in the transformer tank (that is about zero in NHC). which these variations are detected in this paper via inserting some SCs on the transformer legs, and their induced voltages are used to detect TTFs and transformer protection. A three-phase transformer core is considered to fulfil the proposed technique, as shown in Fig. 1, and six SCs by regular interval distances are installed on each phase (due to existing five high voltage (HV) winding sections). For more investigation, the mentioned transformer (its technical specifications are shown in Table 3) is considered to simulate by finite element method (FEM) in ANSOFT MAXWELL [29]. The two-dimensional transient solver is used for this purpose to simulate the transformer behaviour during the TTF occurrence according to the transformer dimensions. The core characteristics such as non-linear B–H curve and its permeability are chosen from MAXWELL package materials list. These simulations are done just to find out the flux behaviour during TTFs. Table 3. Technical specifications of the sample transformer Company SIEMENS Year 1967 Rating power 50 kVA Vector group Yzn5 Voltage ratio 20,000(±4%)/400 V SC impedance 4.1% Frequency 50 Hz HV turns + tap 4410 + 2 × 185 LV turns (zigzag) 2 × 53 HV nominal current 1.44 A LV nominal current 72 A Fig. 3 (left) shows the flux lines distribution in the transformer during healthy and no-load condition, while the rated voltage is applied to the low voltage (LV) winding. The applied voltage waveform on the phase U is as the reference phase (V and W lags and leads U by 120°) and the flux lines are shown at t = 28 ms for healthy case. As can be seen, sum of the three-phase fluxes is equal to zero and no flux is escaped from the transformer core. Fig. 3Open in figure viewerPowerPoint Simulated flux lines in healthy (left) and faulty (right) conditions and the related following current in the shorted turns before and after occurrences of the TTF40V1 (top) Now, a TTF on HV winding including 40 shorted turns (TTF40) is simulated on the first winding section of the middle phase (V) at t = 30 ms, while the core flux distribution is shown at t = 48 ms in Fig. 3 (right). As can be seen, the core flux in the faulty region is decreased while the leakage flux is increased, inversely. In NHC, SCs on each leg sense simultaneously equal and similar variation of linkage flux, i.e. equal voltages induce in all of them (similar fluxes with 120° phase angle shifts will pass through the other SCs installed on the adjacent legs). Accordingly, the output voltages of the opposite SCs (such as: SC1U and SC6U, SC2U and SC5U or SC3U and SC4U) on each phase are exactly equal. Therefore, in NHC (2)Therefore, by summation of the above differential signals and definition of a cumulative signal for each phase (ΔEx), it can be concluded that for NHC (3)or (4)Equation (4) indicates that summation of the induced voltages in the upper SCs and summation of the induced voltages in the lower SCs on each phase are equal in NHC. Indeed, ΔEx is a signal that is made from difference between summation of induced voltages in the upper SCs and summation of induced voltages in the lower SCs , i.e. . Finally, an overall differential-based signal ∑ΔE as (5) can be obtained by summation of ΔEx|x = U, V, W, to present an overall TTF detector signal. This sinusoidal signal ∑ΔE (with some distortions in abnormal conditions such as over-fluxing and inrush current) has amplitude near to zero in NHC, while experiences the higher values in the faulty conditions (5) 3 Preliminary experimental results To validate the theatrical idea, it is implemented on the aforementioned transformer. SCs are installed on the LV winding (instead of the core) to sense and measure the core flux. Each SC is made by five-turns enamel insulated copper wire with 0.4 mm diameter. Such SCs can be installed on the core during the manufacturing process without significant changes in the transformer design; because they are made by a few turns of very thin wires that are wrapped around the core on an insulation paper to protect the sensors. In fact, SCs should be installed on the core (under LV winding) and the connections should be gone out through the core corners, as shown in Fig. 4 (left). However, on the basis of limitation to do this scheme on a manufactured transformer, they are located between LV and HV windings (on the LV winding) and the related connections are inserted on the LV winding by using an insulating paper and left from space between LV and HV windings, as shown in Fig. 4 (right). Fig. 4Open in figure viewerPowerPoint Windings positions on the transformer core to install during manufacturing (left) and mounted on the sample transformer (right) To validate the proposed technique performance, two classes of TTFs are experimentally tested by applying 15 V (6.5% of Vn) on the LV winding, including: TTFs with various number of the shorted turns in a specified location. TTFs located on the various places with a specified number of the shorted turns. 3.1 TTFs with various number of shorted turns To explain the effect of the number of shorted turns on the ∑ΔΕ magnitude, three TTFs including: TTF40, TTF100 TTF400 are created artificially on the first winding section of phase V. Fig. 5 shows the test results for the mentioned TTFs. As can be seen, the TTF with the more shorted turns produces: Bigger ∑ΔΕ magnitude. Larger ΔE in the faulty phases. Fig. 5Open in figure viewerPowerPoint Effect of TTFs severity on ∑ΔE (left column) and ΔEx for x = U, V, W (right column) In addition, it can be seen that ΔE in the adjacent phases has a phase displacement of 180°. Indeed, the escaped flux from the faulty phase links the adjacent phase SCs and induces a voltage with opposite polarity in each of them. However, the biggest ΔE appears in the faulty phase, which can be used to discriminate the faulty phase from the healthy ones. 3.2 TTFs located on the various places To clarify the fault location effect on ∑ΔΕ magnitude, a series of TTF100 are created artificially on some locations (winding sections), as: the first winding section of the phase V (TTF100V1), the fourth winding section of the phase U (TTF100U4) and the middle point (third winding section) of the phase U (MTTF100U). Fig. 6 (left) shows the test results for the above-mentioned TTFs. As can be seen, MTTF100U (against the TTF100V1 and TTF100U4) cannot be detected by the introduced ∑ΔΕ. Indeed, MTTFs cannot be detected by using such ∑ΔΕ, which is made by the opposite and regular installed SCs. In other words, MTTFs create symmetrical distortions in the passing fluxes through the opposite SCs (, and are about zero), and they deactivate the ∑ΔΕ. Fig. 6Open in figure viewerPowerPoint Effect of TTF location on the ∑ΔE and ΔEx for x = U, V, W before (left two columns) and after modification (right two columns) 4 Modification process To modify the proposed technique, an asymmetrical component such as (or or ) have to be added to ΔEx in (4) as (6)or (7)Such modification can be done by: Practical method, i.e. inserting duplex SC1 and SC4, as (6). Numerical method, i.e. the SC1 and SC4 induced voltages are multiplied numerically by 2, as (7). The modified technique results are shown in Fig. 6 (right columns). As can be seen, the MTTF can be detected by the modified technique. Though a multiple of such additional component can be added to ΔEx (e.g. or ) to make a more sensitive criterion, considering the bigger multiple of generates the larger voltage in the ∑ΔΕ during faulty conditions, which should be controlled and limited to prevent the protective device (∑ΔΕ meter) against the employ of impermissible voltage amplitude. 5 Stability in abnormal conditions A suitable protection algorithm has to be secure in the face of abnormal conditions. Accordingly, the introduced technique performance has to be tested during transformer energising (inrush) current and over-fluxing condition. 5.1 Transformer energising Since the inrush current is not a faulty condition, transformer protection must remain stable during the inrush transient. The typical inrush currents contain substantial amounts of the second, third and fifth harmonics [5]. Three-phase currents and ∑ΔΕ output voltage are recorded by an oscilloscope (Tektronix TDS 2024C) in normal trigger mode to record the transient energising waveforms, as shown in Fig. 7. As can be seen, inrush current amplitude can be reached to about 250 A in the worst cases for this sample transformer. Fig. 7Open in figure viewerPowerPoint Three-phase current waveforms (top row), ∑ΔE (middle row) and the peak values of the fundamental frequency and harmonic components (bottom row) for inrush currents during three worst cases Three suitable CTs are used to measure the inrush currents in three phases, which their secondary windings are connected to suitable resistive elements to convert the currents to voltages. As shown, all of inrush currents and ∑ΔΕ output voltage are non-sinusoidal and transient waveforms. Accordingly, two criteria can be considered to discuss the inrush current effect on the ∑ΔΕ, as follows: ∑ΔΕ total amplitude-based algorithm As shown in Fig. 7, the maximum amplitude of ∑ΔΕ is about 3 V, i.e. a threshold of 4 V can be considered as a suitable value to secure the protection algorithm against the inrush current miss-operation. By a linear estimation to find the ∑ΔΕ amplitude in faulty conditions, it can be concluded that even MTTF100U as a worst case can be detected easily by considering such threshold (4 V). Indeed, while MTTF can create a peak value of 0.4 V in ∑ΔE by applying 6.5% of the rated voltage, it can be estimated (by linear estimation) to have 6.15 V in the rated voltage. ∑ΔΕ fundamental component amplitude-based algorithm Using a digital (numerical) or analogue filter, the higher harmonics can be filtered to obtain more sensitivity to protect transformer. Since the inrush currents in the various phases are non-sinusoidal and non-stationary/transient signals, the peak values of their decomposition components are shown in Fig. 7. As shown, the maximum value of the inrush current fundamental frequency does not exceed from 1.32 V, i.e. considering a threshold as small as 1.35 V can be suited to access a sensitive and reliable protection. Using such threshold, TTFs including a few tens of shorted turns are detectable, as well. It is worth noting that the above-determined thresholds (on the basis of either the fundamental component or the total amplitude) depend on the various parameters such as transformer core material and network short-circuit capacity; therefore, the above thresholds cannot be considered in all cases and for all transformers. 5.2 Over-flux condition Though over-fluxing is usually encountered in transformers which are directly connected to the generators, a suitable protection algorithm has to become robust in the face of such phenomenon. Since the core flux symmetricity cannot be deformed due to over-flux phenomenon, it is anticipated that it cannot impact the proposed technique performance, while whole of the core leg experiences the more flux density. In such conditions, the leakage flux increases and the primary current (magnetisation) waveform is distorted due to core saturation property, without any distortion in the flux symmetricity. To show the proposed technique performance in such condition, 110% of the nominal voltage (440 V) is applied on the LV windings. The no-load currents, ∑ΔΕ and peak values of them are shown in Fig. 8. As can be seen, the second harmonic is dominant component in over-fluxing state. Moreover, ∑ΔΕ is approximately equal to zero, as expected. Fig. 8Open in figure viewerPowerPoint Three-phase currents waveforms (up), ∑ΔE (mid) and the peak values of the fundamental frequency and harmonics components (down) in over-fluxing condition 6 Final experimental results As the final test, an artificial TTF10V1 is created on the sample transformer for a short time and the recorded signals are shown in Fig. 9. The fundamental component magnitude of this waveform is 1.4 V, which is bigger than the obtained threshold (1.35 V using the ∑ΔΕ fundamental component amplitude-based algorithm, which is obtained in Section 5.1). This result shows that the proposed algorithm is a success to detect the minor TTFs as small as ten turns (lower than 0.25% of the winding). In other words, it can be introduced as a sensitive technique to identify TTFs in transformers. Fig. 9Open in figure viewerPowerPoint ∑ΔE (left) and ΔEx for x = U, V, W (right) during TTF10V1 in the rated voltage 7 Discussion It is notable that not only the internal TTFs can be detected by the proposed technique but also other types of internal faults are diagnosable by this technique. For example, the Earth faults near the star point are detectable whenever the star connected winding is grounded solidly. Such faults are similar to TTFs near the star point, and then they are detectable by the proposed technique, as well. Otherwise, while the star point is grounded through an impedance, the proposed technique sensitivity will be reduced, so that it maybe not successful to detect such Earth faults. TC operation is another issue that can influence the transformer protection performance and reduce the TTF detection sensitivity. This problem cannot impact on the proposed technique performance; because the TC operation changes the winding turns while the flux symmetricity along the core leg remains stable (symmetrical SCs sense similar flux in each leg). Therefore, ∑ΔE remains zero and the proposed technique can be introduced as a stable and secure method in the face of such conditions. The presented technique can discriminate the external faults from the internal ones, too. When an external fault occurs, transformer winding is shorted entirely. Therefore, the corresponding SCs measure the equal reduced linkage fluxes overall the faulty phase, and accordingly the ∑ΔΕ on the faulty phase cannot be activated, as it should be. As a brief review, the ∑ΔΕ can detect any event that creates a distortion in the flux symmetricity along each core leg. It senses only the flux uniformity along the core legs and reacts whenever this regularity is distorted. Since the inrush current, over-fluxing, TC operation and even unbalanced load or unbalanced voltage source cannot disturb significantly the flux homogeneity in each leg, the proposed technique is not affected by the mentioned events. The proposed technique can be used for all power transformers with different rating voltages and powers. Moreover, type of windings connections cannot impact the performance of the proposed technique, while it is demonstrated on the basis of the flux homogeneity in the healthy transformer core. 8 Conclusion In this paper, a novel LFB protection technique is proposed to detect the TTFs in power transformer windings, through installing some thin SCs on the transformer legs to measure the linkage flux passing through them. Using series symmetrical SCs with opposite polarities on each core leg, a signal ΔE is obtained to identify TTF on the related winding, and a cumulative signal as ∑ΔΕ (that is the summation of ΔE on three legs) is introduced as an overall TTF detector in online conditions. As a brief description, the ∑ΔΕ can detect any event that creates a distortion in the flux homogeneity along each core leg. It senses only the flux non-uniformity along the core legs. Since inrush current, over-fluxing, TC operation and unbalanced load/voltage source cannot have any effect on the flux distribution symmetricity along the core legs, they cannot impact the LFB technique performance. It is shown experimentally that the proposed technique: can sensitively detect the minor TTFs; is stable during transformer energising process; and is stable in the face of over-fluxing condition; and then, it can be introduced as a successful online technique to protect the expensive power transformers. 9 References 1Wang, H., Butler, K.L.: ' Modeling transformer with internal winding faults by calculating leakage factors'. Proc. of the 31st North American Power Symp., San Luis Obispo, CA, USA, 1999 2Behjat, V., Vahedi, A.: 'An experimental approach for investigating low-level interturn winding faults in power transformers', Electr. Eng., 2013, 95, (2), pp. 135– 145 (doi: 10.1007/s00202-012-0248-4) 3Guzmán, A., Zocholl, S., Benmouyal, G., et al.: 'A current-based solution for transformer differential protection. part I: problem statement', IEEE Trans. Power Deliv., 2001, 16, (4), pp. 485– 491 (doi: 10.1109/61.956726) 4Wang, M., Vandermaar, A.J., Srivastava, K.D.: 'Review of condition assessment of power transformers in service', Electr. Insul. Mag., 2002, 18, (6), pp. 12– 25 (doi: 10.1109/MEI.2002.1161455) 5 ALSTOM: ' Network protection & automation guide' ( Alstom Grid, France, 2002, 1st edn.) 6Ziegler, G.: ' Numerical differential relay' ( Publicis KommunikationsAgentur GmbH, GWA, Berlin, Erlangen, 2005) 7Hamilton, R.: 'Analysis of transformer inrush current and comparison of harmonic restraint methods in transformer protection', IEEE Trans. Ind. Appl., 2013, 49, (4), pp. 1890– 1899 (doi: 10.1109/TIA.2013.2257155) 8Hamedani Golshan, M.E., Saghaian-nejad, M., Saha, A., et al.: 'A new method for recognizing internal faults from inrush current conditions in digital differential protection of power transformers', Electr. Power Syst. Res., 2004, 71, pp. 61– 71 (doi: 10.1016/j.epsr.2003.11.015) 9Alomar, A.M., Shamseldin, B.E.: 'Alternative approaches for distinguishing between faults and inrush current in power transformers', Energy Power Eng., 2014, 6, pp. 143– 160 (doi: 10.4236/epe.2014.67014) 10Xiang-ning, L., Liu, P., Malik, O.P.: 'Studies for identification of the inrush based on improved correlation algorithm', IEEE Trans. Power Deliv., 2002, 17, (4), pp. 901– 907 (doi: 10.1109/TPWRD.2002.803730) 11Ramesh, K., Sushama, M.: 'Power transformer protection using wavelet packet transform'. Int. Conf. on Computer Communication and Informatics (ICCCI), 2014, pp. 1– 7 12Babiy, M., Gokaraju, R., Garcia, J.C.: ' Turn-to-turn fault detection in transformers using negative sequence currents'. IEEE Electrical Power and Energy Conf., Canada, 2011 13Gajić, Z., Brnčić, I., Hillström, B., et al.: ' Sensitive turn-to-turn fault protection for power transformers'. CIGRÉ Study Committee B5 Colloquium, Calgary, Canada, 2005 14Abi, M., Mohammad, M.: ' Detection and localization of internal turn-to-turn short circuits in transformer windings by means of negative sequence analysis'. 29th Int. Power System Conf. (PSC), Tehran, Iran, 2014 15Venikar, P.A., Ballai, M.S., Umre, B.S.: 'Negative sequence based relay operate area approach to detect transformer inter-turn fault', The 40th Annual Conference of Ind. Electron. Soc. (IECON), 2014, pp. 749– 754 16Vahedi, A., Behjat, V.: 'Online monitoring of power transformers for detection of internal winding short circuit faults using negative sequence analysis', Eur. Trans. Electr. Power, 2011, 21, (1), pp. 196– 211 (doi: 10.1002/etep.431) 17Wiszniewski, A., Rebizant, W., Schiel, L.: 'New algorithms for power transformer inter-turn fault protection', Electr. Power Syst. Res., 2009, 79, pp. 1454– 1461 (doi: 10.1016/j.epsr.2009.04.021) 18Abu-Siada, A., Islam, S.: 'A novel online technique to detect power transformer winding faults', IEEE Trans. Power Deliv., 2012, 27, (2), pp. 849– 857 (doi: 10.1109/TPWRD.2011.2180932) 19Oliveira, L., Cardoso, A.J., Cruz, M.A.: 'Power transformers winding fault diagnosis by the on-load exciting current extended Park's vector approach', Electr. Power Syst. Res., 2011, 81, (6), pp. 1206– 1214 (doi: 10.1016/j.epsr.2011.01.009) 20Joshi, N.C., Sood, Y.R., Jarial, R.K., et al.: 'Transformer internal winding faults diagnosis methods: a review', MIT Int. J. Electr. Instrum. Eng., 2012, 2, (2), pp. 77– 81 21Bhide, R.S., Srinivas, M.S.S., Banerjee, A., et al.: ' Analysis of winding inter-turn fault in transformer: a review and transformer models'. IEEE ICSET, Kandy, Sri Lanka, 2010 22Cabanas, M.F., Melero, M.G., Pedrayes, F., et al.: 'A new online method based on leakage flux analysis for the early detection and location of insulating failures in power transformers: application to remote condition monitoring', IEEE Trans. Power Deliv., 2007, 22, (3), pp. 1591– 1602 (doi: 10.1109/TPWRD.2006.881620) 23Kang, Y.C., Lee, B.E., Kim, Y.G., et al.: 'A transformer protective relaying algorithm using the ratio of induced voltages'. Power Engineering Society Summer Meeting, 2001 24Kang, Y.C., Lee, B.E., Zheng, T.Y., et al.: 'Protection, faulted phase and winding identification for the three-winding transformer using the increments of flux linkages', IET Gener. Transm. Distrib. 2010, 4, (9), pp. 1060– 1068 (doi: 10.1049/iet-gtd.2010.0094) 25Haghjoo, F., Mostafaei, M., Mohammadzadeh, M.: ' A novel flux-based protection scheme for power transformers'. The Ninth Power Systems Protection and Control Conf. (PSPC), Tehran, 2015 26Haghjoo, F., Mostafaei, M.: 'A novel flux-based method to diagnose and identify the location of turn-turn faults in transformers', IET Gener. Transm. Distrib., doi: 10.1049/iet-gtd.2015.1180 27Lehner, G.: ' Electromagnetic field theory for engineers and physicists' ( Springer, Berlin, 2010) 28Chapman, S.J.: ' Electric machinery fundamentals' ( McGraw-Hill, New York, Fourth edn.) 29Software Maxwell 2D Field Simulator, ANSOFT Corporation, February 2002 Citing Literature Volume10, Issue5April 2016Pages 1154-1163 FiguresReferencesRelatedInformation