Legal forms

ABB 654 WPO GIC on Power Transformers

Description
ABB 654 WPO GIC on Power Transformers
Categories
Published
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
   1 R. Girgis, Fellow, IEEE  , K. Vedante, Senior Member IEEE  ABB Power Transformers St. Louis, MO, USA   Abstract: There is some misconception in the electric power industry today that geo – magnetically induced currents (GIC) have caused, and would cause, significant damage to the majority of large and medium power transformers installed on power grids in North America and around the world. The purpose of this paper is to present, both qualitatively and quantitatively, the true effect of GIC on power transformers, based on thorough knowledge of power transformer design and performance. The paper also provides concise explanation of the cases of transformer failure / overheating reported in the published literature to have been caused by GIC currents. The paper demonstrates that, because of the nature of the GIC currents, the great majority of power transformers would not experience much overheating or damage due to even high levels of GIC. Only some specific transformers could suffer some winding damage due to high winding circulating currents if exposed to high levels of GIC. The paper also presents the real issue with GIC; namely, the narrow pulse of the magnetizing current which results from part – cycle, semi – saturation of transformer cores when subjected to high levels of GIC currents. This one current pulse / cycle could cause relays and capacitive components in power systems, such as SVC (s), to trip, which may contribute to grid instability. Also, the current pulse is associated with 2 nd  and higher order harmonics. As a result, resonance may occur, differential relays may operate, and stability of the grid may be compromised.  Keywords:  Power transformers, Power Systems, Geo – magnetically  Induced Currents, GIC, DC. I. INTRODUCTION In order to understand the effect of GIC currents on power transformers and power systems, this paper starts with presenting the basic theory of the effect of DC on Power Transformers. This includes presenting the phenomenon of part – cycle, semi – saturation of transformer cores due to DC, the resulting magnetizing current pulse, and the effect of this core semi – saturation on temperature of transformer windings & structural parts. In the following section, the paper presents a detailed discussion, supported by measured data, of the nature of the GIC currents associated with GIC events. This is followed by deriving the effect of GIC on transformers considering the nature of the GIC currents. Section IV provides explanations of transformer failures / overheating reported in the published literature as caused by GIC currents. In section V, actual measurements, made on full size power transformers when subjected to high levels of DC, are presented. In section VI, a short discussion of the effect of GIC on power systems is presented. II. BASICS OF EFFECT OF DC ON POWER TRANSFORMERS    A. Phenomenon of Part – Cycle, Semi – Saturation of transformer cores when subjected to DC When a power transformer is subjected to DC, it results in a unidirectional DC flux in the core. The magnitude of this flux depends on the magnitude of the DC, number of turns in the windings carrying the DC current, and reluctance of the path of this DC flux. The result is that the DC flux adds to the AC flux in one half – cycle and subtracts from the AC flux in the other half – cycle; as illustrated in Figure 1 (a) below. When large enough, or when the design flux density is high, this leads to core peak flux densities in the magnetic core pre – saturation range in one half of the cycle. As shown in Figure 1 (b) below, the B – H characteristics of the transformer core materials is inherently very non-linear. For higher magnitudes of DC, the core provides a very high reluctance to the DC Ampere – Turns; resulting in a smaller flux density shift. Correspondingly, the magnetizing current becomes a short duration pulse with a high peak. Figure 2 presents the calculated magnitudes and wave – shapes of the magnetizing current which results as a consequence of a large 1 – Phase power transformer being subjected to 3 different levels of DC; namely, 6, 12, and 40 Amps. As can be seen from the figure, the magnitudes of the peaks of these magnetizing currents are 5, 11, and 47 percent of the full load current, respectively. Correspondingly, the magnitude of the percentage magnetizing current in the absence of DC is a fraction of a percent for this transformer. Also, the duration of this pulse is only in the range of 1/10 th  to 1/12 th  of the cycle. Flux Density vs Phase angle -2.0-1.5-1.0-0.50.00.51.01.52.02.5-90-60-300306090120150180210240270300330360 Phase Angle, Degrees    F   l  u  x   D  e  n  s   i   t  y ,   T  e  s   l  a Bm, ACBm, (AC+DC)   (a) Flux density shift in the core caused by DC Effects of GIC on Power Transformers and Power Systems _____________________________________________ Dr. Ramsis Girgis  is presently the R & D manager at the ABB Power Transformer plant, Saint Louis, MO. Mr. Kiran Vedante is presently a Senior R &D Engineer at the ABB Power Transformer plant, Saint Louis, MO.     2 (b) Part – Cycle, Semi – Saturation of Transformer cores Fig. 1 – Effect of DC on transformer core magnetization % Magnetizing Current vs Phase Angle 0%5%10%15%20%25%30%35%40%45%50%0306090120150180210240270300330360 Phase Angle, Degrees    %    M  a  g  n  e   t   i  z   i  n  g   C  u  r  r  e  n   t Neutral DC = 6 AmpsNeutral DC = 12 AmpsNeutral DC = 40 Amps  Fig. 2 – Magnetizing current pulse due to DC current  B. Factors affecting how much DC causes Semi – saturation Core – type, core geometry, and core design As mentioned earlier, the magnitude of the DC flux shift in the core depends on the magnetic reluctance of the DC flux path. Thus, the DC flux shift in a 3 – phase, 3 – limb core form transformers would be the lowest of all core – types. This is because this design offers an order of magnitude higher magnetic reluctance to the DC Ampere – turns in the Core – Tank magnetic circuit, refer to Figure 3 below. The DC flux has to pass through the very high reluctance path from the core top yoke to the tank cover, through the tank walls, and return to the bottom yoke through, again, the high reluctance path from the tank bottom. All other core – types offer much less reluctance to the DC Ampere – turns because the path of this DC flux is through the core; which has orders of magnitudes higher permeability, see Figure 3 below. Core material and type of core joint have some influence. However this influence is small and depends on the core – type and the operating flux density in the core. This influence decreases even further for high magnitudes of DC. Fig. 3 – Core DC flux path in various core types Figure 4 below presents the percentage magnetizing current drawn by two different types of transformers, one is a large 1 – phase power transformer with a 3 – limb core and the other is a large 3 – phase power transformer with a 3 – limb core. It can be seen that the magnetizing current drawn by the 3 – phase, 3 – limb core is much smaller than the magnetizing current drawn by the 1 – phase transformer for the same values of DC currents flowing through the windings of these transformers. % Magnetizing Current vs DC Current 012345678051015202530354045 DC Current/Phase    %    M   a   g   n   e   t   i   z   i   n   g   C   u   r   r   e   n   t 3-Phase, 3-Limb Transformer1-Phase Transformer  Fig. 4 – Peak magnetizing current for 2 different core – types C. Harmonic content of the magnetizing current pulse Typically, 3 – phase transformers with 3 – limb cores will have narrower magnetizing current pulses under the effect of DC compared to corresponding magnetizing current pulses of transformers with other core – types. Figure 5 compares the %   3 harmonic content for these two types of transformers. The magnetizing current pulse in the 3 – phase transformer has higher amplitudes of the lower order harmonics and much lower amplitudes of the high order harmonics. For other core – types, represented by the 1 – phase transformer data in Figure 5, the magnetizing current pulse has much more uniform amplitudes among low and high order harmonics. Therefore, these transformers would have a significant content of higher order harmonics. Another observation from Figure 5 is that transformers, with other than the 3 – phase, 3 – limb, core – type, have a much lower % of the 2 nd  order harmonic. This is significant as lower levels of 2 nd  order harmonics associated with the total primary current of the transformer could give an erroneous message to the differential relay set at a low value of 2 nd  harmonic content to differentiate between Inrush current events and a fault condition. Harmonics Spectrum of Magnetizing Current under DC Conditions0%2%4%6%8%10%12%14%16%60120180240300360420480540600660Harmonic Frequency, Hz    %    H  a  r  m  o  n   i  c  s   A  m  p   l   i   t  u   d  e  3 - Phase, 3 - Limb Transformer1 - Phase Transformer   Fig. 5 – Harmonic content of magnetizing current of transformers subjected to DC  D. Calculation of increase in winding hot spot temperatures The much higher magnetizing current, and the nature of its wave – shape, produce correspondingly higher magnitudes of leakage flux that is also rich in harmonics. This results in appreciably higher eddy and circulating current losses in the windings as well as the structural parts of the transformer but a small increase in the total load losses of the transformer. In Figure 6 below, calculated temperatures of the hot spot in the windings of a 1 – Phase large power transformer are presented when the transformer is subjected to 20, 30, and 50 Amps / phase DC currents for a 30 minute duration, while fully loaded. The figure shows that the hot spot temperature reached close to a final value within about 5 minutes from the application of the DC. This corresponds to the typical range of the time constant of transformer windings. The increase in the hot spot temperature is still not significant even after being subjected to a continuous 50 Amps DC for 30 minutes. This increase is only 12 º C. Winding Hot Spot Temperature vs Time, 1-Phase Transformer 0102030405060708090100110120130051015202530 Time, Minutes    W   d  g   H  o   t   S  p  o   t   T  e  m  p   t ,   D  e  g  r  e  e   C Idc = 20 AmpsIdc = 30 AmpsIdc = 50 Amps  Fig. 6 – Winding Hot spot temperatures under DC  E. Calculation of hot spot temperature of structural parts The magnitude of this temperature rise depends on the core construction, the operating / design flux density, and magnitude of DC current flowing through the windings. Once the core flux density reaches close to the saturation flux density level of the core steel, there will be spillage of the core flux outside the core. This results into additional flux linkages to the structural parts, such as tie plates, yoke clamps, tank walls, tank cover, tank bottom, etc. This adds to the existing leakage flux due to load current and the high peak magnetizing current pulse. The result is an additional increase in temperature of these parts. Figure 7 below presents the calculated hot spot temperature of a tie – plate for different values of DC applied continuously for 120 minutes. The figure shows that the hot spot temperature reached close to a final value within about 20 minutes from the application of the DC. This corresponds to the typical range of time constant of the metallic structural parts of transformers. The increase in the hot spot temperature is still not significant even after being subjected to a 50 Amps DC for 120 minutes continuously (32 C). Under actual GIC, however, this temperature rise will be much lower than shown below due to the very short duration nature of GIC (1 – 2 minutes). This is demonstrated in the following section of this paper. Flitch-Plate Temperature vs Time 020406080100120140160020406080100120 Time, Minutes    F   l   i   t  c   h  -   P   l  a   t  e   T  e  m  p   t ,   D  e  g  r  e  e   C Idc = 20 AmpsIdc = 30 AmpsIdc = 50 Amps   Fig. 7 – Tie – plate hot spot temperature under DC   4 III. EFFECTS OF GIC ON POWER TRANSFORMERS   In order to accurately determine the capability of a transformer to GIC, one needs to consider the nature of the signature of the GIC current as well as magnitudes and time duration involved with a GIC event. Typical signature of GIC Currents Figs. 8 & 9, below, present two different examples of GIC signatures; one is estimated [1] based on the srcinal measured VAR consumption at the PSE&G Salem Generating station during the K9 GIC event on March 13, 1989. The other is a recent signature obtained at a generating station in southern Manitoba, Canada during a GIC event in February, 2011. Fig. 8 – GIC current at Salem Generating station in 1989 Fig. 9 – Measured GIC at a Generating station in Southern Manitoba, Canada in February 2011 As it is demonstrated in these two figures, a Solar magnetic event is associated with GIC which is quasi DC characterized by high levels of DC as well as a few much higher peaks that have very short time duration. For example, in Fig. 8 above, the GIC signature shows mainly several consecutive narrow (a few minutes duration) peaks of 20 amps magnitude over a period of 90 minutes along with two high peaks of 80 & 100 Amps of about 2 minute duration each. Similarly in Fig. 9, the GIC event is characterized by a number of narrow peaks of < 5 amps magnitude for a total accumulative duration of 5 minutes and one 8 amps peak of < 1 minute duration. Calculation of the effect of GIC currents Due to the short duration of the high peaks of GIC (1 – 2 minutes) and the fact that the duration of the resulting magnetizing current pulse is only in the range of 1/10 th  to 1/12 th  of a cycle, the actual duration of the resulting core part – cycle, semi – saturation, and associated high peak pulses of the core magnetizing current, is only a few seconds. Hence, the increase in temperature rises in the transformer windings due to GIC would be expected to be at least an order of magnitude lower than that estimated based on continuous duration DC currents. The same is true for temperature rises in the core and structural parts. This is because the thermal time constants of windings and metallic structural parts are much higher than the application time of the current pulse and associated leakage flux in the transformer. In order to illustrate the above, a profile of GIC per Figure 10 (a) below was assumed and winding hot spot temperature was calculated for a fully loaded large 1 – phase power transformer when subjected to this assumed GIC current profile. The assumed GIC profile is a base level of 100 Amps / phase (300 Amps for a 3 – phase bank) followed by 2 minute duration very high level GIC pulses of 400 Amps / phase (1200 Amps for a 3 – phase bank) every 30 minutes. This GIC profile / magnitudes is almost 5 times that experience at the Salem Generating Station caused by the K9 GIC event of March 13, 1989. The calculated winding temperatures due to the GIC profile of Figure 10 (a) are presented in Figure 10 (b) below. It can be seen from the figure that: 1.   The temperature rise of the winding hot – spot due to the base 100 Amps is about 7º C; which when added to the winding hot spot temperature of 110 º C at full load results in a total hot spot temperature of 118 º C. 2.   The temperature of the winding hot – spot, due to the 2 minute duration of 400 Amps / phase of GIC, increases by about 35 C; resulting in a hot spot total temperature of about 152 C. The temperatures will be much lower when the transformer is not fully loaded and / or ambient temperature is lower than 30 C. Such winding hot spot temperatures for such a short duration would not cause any appreciable damage of the windings, structural parts, or loss of insulation life of the transformer. In fact. Industry Standards allow much higher winding hot spot temperature levels for much longer times under emergency loading conditions. 3.   After the 2 – minute duration, the hot spot temperature of the windings goes back down to the srcinal temperature existing before that pulse of GIC. In fact, it takes only about 3 minutes for this to happen as the time constant of the windings is only several minutes. 4.   The duration of that pulse of winding temperature rise is only a few minutes; which would not result in any damage to the windings of this transformer anyway.   5 The same is true for structural parts, except that the temperature rise will be even smaller for these structural parts as their time constant is higher than that of the windings.                  (a)   Simulated GIC current Winding Hot Spot Temperature vs Time 02040608010012014016001020304050607080 Time, Minutes    W   i  n   d   i  n  g   H  o   t   S  p  o   t   T  e  m  p   t ,   D  e  g  r  e  e   C  (b) Calculated Winding Hot Spot Temperature Fig. 10 – Winding Hot Spot Temperature for GIC profile above IV.  REPORTED T RANSFORMER DAMAGE  /   OVER –  HEATING CONTRIBUTED TO GIC Reference [1] reports several of those cases. Figure 11 below shows a picture of the one shell form transformer that experienced significant overheating of the windings of a GSU transformer at a PSE&G Generating power station at Salem, NJ during the March 13, 1989 K9 GIC event that injected extremely high GIC currents shown in Figure 8 of this paper. The transformer did not actually fail but was taken out of service a week later because of significant gassing. Fig. 11 – Winding series – connection overheating in PSE&G Transformer caused by the March 13, 1989 GIC Event This is an old Shell form transformer that had an old winding design that made it susceptible to overheating caused by high circulating current when subjected to high levels of GIC currents. This winding design was srcinally optimized for the leakage flux pattern typically associated with normal loading and operating conditions. During this high GIC event, the high GIC current caused part – cycle, semi – saturation of the core; during which the leakage flux pattern changed and resulted in very high circulating currents in the series connection of the LV windings of this transformer [2, 3] . Same design transformers at PSE&G experienced similar overheating, although to a lesser degree. This old generic design of the LV winding was already changed since the early seventies. The newer design minimizes circulating currents in this winding in general as well as under the conditions of core semi – saturation associated with high levels of GIC. Another case reported in Reference [1] is a Shell – form Transformer that had tank wall heating during the same GIC event. In the construction of such shell – form transformers, wood slabs are placed between the core and tank walls. These areas of the tank walls are, therefore, cooled by air only on one side. Under GIC conditions, when the core goes through the short periods of semi – saturation, part of the main flux in the core and the windings leakage flux travel to the tank wall causing localized eddy losses in the tank. Being blanketed by the wood slabs these regions of the tank start to overheat. Additionally, it should be noted that this temperature increase had not resulted into any real consequences to transformer except discoloration of the tank paint at the tank regions opposite to the wood slabs on the inside of the tank walls [4] . Improvements made to this type of construction allow better cooling of the tank walls.   Another case of overheating damage reported to have been caused by GIC is the case of several EHV core form power transformers in South Africa; where overheating of the HV leads was observed. At that time, the manufacturer of these transformers investigated the cause of the lead overheating and it was realized at the time that these leads were over – insulated which did not allow proper oil flow in these leads resulting in the overheating. The high pulses of magnetizing current that these transformers experienced during the very high GIC event there could have contributed some to the overheating. It was concluded that overheating could have also been already there before the GIC event. Also, Reference [1] included a statement that within 2 years of the 1989 GIC event, 11 nuclear plants experienced failures of several Generator Step – Up transformers (GSUs) but no details were given. During that period, a number of GSU failures have been studied in detail and were found to be caused by back – feed mode operation. In this mode of operation, the generator step – up transformers are not sufficiently protected from switching and lightning surges. Also, the generator is not connected and, hence, the electrical damping in the electric circuit is very low making it vulnerable to electrical circuit and winding resonances.
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks