CHAPTER POWER-FLOW SOLUTIONS Power-fow studies are of great importance in planning and Gesigningthe future expansion of power systems as weil as in determining the best operation of existing systems. The principal information obtained from a power-flow study is the magnitude and phase angle of the voltage at each bus and the real and reactive power flowing in each line. However, much additional information of value is provided by the printout of the solution from computer programs used by the electric utility companies. Most of these features are made evident in our discussion of power-How studies in this chapter We shal. examine some of the methods upon which solutions to the power-flow problem are based. The great value of the powerflow computer program in power system cesign and operation wil! become apparent 9.1 THE POWER-FLOW PROBLEM Either the bu self- and mutual admittances which compose the bus admittance matrix Y¥,,, or the driving-point and transfer impedances which compose Zyys may be used im solving the power-flow problem. We confine our study to methods using admittances. The starting point in obtaining the data which must be furnished to the computer is the single-line diagram of the system. Transmis- sion lines are represented by their per-phase nominal-7 equivalent circuits like that shown in Fig, 6.7. For each line numerical values for the series impedance Z and the total line-charging admittance Y (usually in terms of line-charging megavars at nominal voltage of the system) are necessary so that the computer can Cetermine all the elements of the NN bus admittance matrix of which 329330 CHAPTER 9 POWER.FLOW SOLUTIONS the typical elemeat ¥,, is IZ, = 1¥gleos 8; + JI¥ylsin 6,, = Gi; + 1B, (9.1) Other essential information includes transformer ratings and impedances, capacitor ratings, and transformer tap settings. In advance of each power-flow study certain bus voltages and power injections must be gwen known values, as discussed below. The voltage at a typical bus (@) of the system is given in polar coordinates by ¥,= HLL, = IMI(cos 4, + sin 3,) (92) and the voltage at another bus (7) is similady written by changing (he subscript from i to j. The net current injected into the nenwork at bus Co in terms of the elements ¥,,, of Y,,. is given by the summation Ybad 0 4 Yin (9.3) Let P, and Q, denote the net real and reactive power entering the network at the bus G). Then, the complex conjugate of the power injected at bus @ fs Fe - IQ (94) in which we substitute from Eqs. (9.1) and (9.2) to obtain & Pe IQ:= ViVi Bin + 8. = 5 (9.5) at fa — Expanding this equation and equating real and reactive parts, we obtain Nn LN, MMos, +8, - 8,) (9.6) maa Nv Q,= - LAIY,MV,lsin(@,, +8 - 4) (9.7) ast Equations (9.6) and (9.7) constitute the polar form of the power-flow equations; they provide calcwlated values for the net real power P; and reactive power Q, entering the network at typical bus @. Let P,; denote the scheduled power being generated at bus () and P,, denote the Scheduled power demand of the©) THEPOWER-FLOW PROBLEM 331 Pe Pha Pe On ian QE o—— — |p, ae == /@, © le —, eo Fon bo : (a) (b) FIGURE 9.1 Novation for (a) active and reactive power at a typical bus @) in power-fiow studies Joad at that bus Then, Py = is the net scheduled power being injected into the network at bus (@), as illustrated in Fig. 9.1(@). Denoting the calculated value of P. by P,gje leads to the definition of mismarch AP, as the scheduled value 2, ., minus the calculated value P..., BPS Pay — Proate = Beem Pad — Pe cate (9.8) Likewise, for reactive power at bus CG) we have = (Qe — Qui) ~ Qecate (99) as shown in Fig. 9.1(6). Mismatches occur in the course of solving a power-flow problem when calculated values of P, and Q, do not coincide with the sched uled values. If the calculated values P, .¢ and Q,.cay. match the scheduled values Pc, aNd O, .o, perfectly, then we sey that the mismatches AP, and AQ; are zero at bus @), and we write the power-balance ecuations AQ, = Qiu — Qc =P, Pisey =P (Pim Pu) = 0 (9.10) Q;— Qiun = Qi ~ (Oy: ~ Qa) = 9 (9.11) Aswe Shall sce in Sec, 9.3, the functions gy and gy are convenient for writing certain equations involving the mismatches AP, and AQ,. If bus @ has no generation or load, the appropriate terms are set equa! to zero in Eqs. (9.10) and (9.11), Each bus of the network has two such equations, and the power-flow problem is to solve Eas. (9.6) and (9.7) for values of the unknown bus voltages which cause Eqs. (9.10) and (9.11) to be numerically satisfied at each bus. If there is no scheduled value P,.,, for bus @, then the mismatch AP; = Py coy — e Cannot be defined and there is no requirement to satisfy the correspond- ing Eq. (9.10) in the course of solving the power-flow problem. Similarly, if Q, on is not specified at bus (, then Eq. (9.11) does not have to be satisfied. Four potentially unknown quantities associated with each bus @ are P,Q), voltage angle 3,, and voltage magnitude |I/|, At most, there are two equations like Eqs. (9.10) and (9.11) available for each node, and so we must consider how the number of unknown quantities can be reduced to agree with332 CHAPTERS POWER.FLOW SOLUTIONS the number of available equations before beginning to solve the power-fiow problem. The general practice in power-flow studies is to identify three types of buses in the network. Atcach bus @) two of the four quantities 6, |V;|, P and Q; are specified and the remaining two are calculated. Specified quantities are chosen according to the following discussion: 1. Load buses. At each nongenerator bus, called a foad bus, both P,, and O,, are zero and the real power P,, and reactive power Q,, drawn from the system by the load (negarive ayputs into the system) are known from histori- cal record, load forecast, or measurement. Quite often in practice only rea) power is known and the reactive power is then based on an assumed power factor such as 0.85 or higher. A load bus @ is often called a P-O bus because the scheduled values Po, = —Py and QO, = —Qy; are known and mismatches AP, and AQ, can be defined. ‘The corresponding Eqs. (9.10) and (9.11) are then explicitly included in the statement of the power-flow problem, and the two unknown quantities to be determined for the bus are 6; and (V1 2. Voltage-controlled buses. Any bus of the system at which the voltage magni- tude is kept constant is said to be voltage controlled, At each bus to which there is a generator connected the megawatt generation can be controlled by adjusting the prime mover, and the voltage magnitude can be controlled by adjusting the generator excitation. Therefore, at each generator bus (1) we may properly specify P,, and |V/|. With P,, also known, we can define mismatch A P, according to Eq, (9.8). Generator reactive power Q,; required to support the scheduled voltage |] cannot be known in advance, and so mismatch AQ, is not defined. ‘Pheretore, at a gencrator bus @) voltage angle 6, is the unknown quantity to be determined and Eq, (9.10) for P, is the available equation. Afier the power-flow problem is solved, Q, can be calculated from Eq, (9.7) For obvious reasons a generator bus is usually called a voltage-con trolled or PV bus. Certain buses without generators may have voltage control capability, such buses are also designated voltage-controlied buses at which the reai power generation is simply zero. 3. Slack bus. For convenience throughout this chapter bus (1) is almost always designated as the slack bus. The voltage angle of the slack bus serves as reference for the angles of all other bus voltages. The particular angle assigned to the slack bus voltage is not important because voltage-angle differences determine the calculated values of P, and Q; in Eqs. (9.6) and (9.7). The usual practice is to set 8, = 0°. Mismatches are not defined for the slack bus, as explained below, and so voltage magnitude |V;| is specified as the other known quantity along with 8, = 0°. Then, there is no require- ment to include either Eq. (9.10) or Eq, (9.11) for the slack bus in the power-flow problem.9) THEPOWER-FLOWPROBLEM 333, To understand why P, and Q, are not scheduled at the slack bus, consider that at each of the N buses of the system an equation similar to Eq. (9.10) can be written by letting @ range from 1 to N. When the resulting N equations are added together, we obtain N N Nv Py = Lhe Pee di (9.12) Real power loss Total generation Total load The term P, in this equation is evidently the total J?R loss in the transmission lines and transformers of the network. The individual currents in the various transmission !ines of the network cannot be calculated until after the voltage magnitude and angle are known at every bus of the system. Therefore, Py is initially unknown and it is not possible to prespecify ai’ the quantities in the summations of Eq. (912). In the formulation of the power-fow problem we choose one dus, the slack bus, at which P, is noe scheduled or otherwise prespecified. Aiter the power-flow prodlem has heen solved, the difference (slack) between the total specified P going into the system atall the other buses and the total outout P plus /?R losses are assigned to the slack bus. For this reason a generator bus must be selected as the slack bus. The difference betveen the total megavars supplied by the generators at the buses and the megavars received by the loads is g'ven by N ¥ LoO= LO, (9.13) int ‘This equation is satisfied on an individual bus basis by satisfying Eq. (9.11) at each bus @ in the course of solving the power-flow problem. Individual Q, can be evaluated from Ey. (9.7) after the power-flow solution becomes available. ‘Thus, the quantity on the ‘cft-hand side of Eq. (9.13) accounts for the combined megavirs assooted with bne charging, shunt capcitors and reactors installed at the buses, and the so-called 17X° loss in the series reactances of the transmission lines The unscheduled bus-voltage magnitudes and angles in the input data of the power-flow study are culled stare wariabtey or dependent variables since theit lues, which describe the svate of the system, depend on the quantities specified at all the buses, Hence, the power-fiow problem is to determine values for all state variables by solving an equal number of power-flow equations based on the input data specifications. If there are NV, voltage-controlled buses (aot counting the slack bus) in the system of N buses, there will be (2N — N, — 2) equations to be solved for (2N — N, — 2)state variables, as shown in Table 9.1. Once the state variables have been calculated, the complete state of the system is known and all other quantities which depend on the state variables can be3340 CHAPTER 9 POWER-FLOW SOLUTIONS TABLE 9.1 Summary of power-flow problem No. of Quantities available No.of 8.1% Bus type No.ofbuses specified equationt state variables Slack: # = 1 1 5,.1Mi1 0 0 Voltage controiied 7 . , (= 2 M1) ' fel Ns ' Load a , ; . : (FN de) NTMTE MQ UNAM UNE N=) Totals a aN IN=N-2 IN=-W—2 determined. Quantities such as Py and Q, at the slack bus, Q, at each voltage-controlled bus, and the power loss P; of the system are examples of dependent functions. The functions P, and Q; of Eqs. (9.6) and (9.7) are nonlinear junctions of the state variables 6, and |V,|. Hence, power-flow calculations usually employ iterative techniques such as the Gauss-Seide] and Newton-Raphson procedures, which are described in this chapter. The Newton-Raphson method solves the polar form of the power-flow equations until the AP and AQ mismatches at all buses fall within specified tolerances. The Gauss-Seidel method solves the power-flow equations in rectangular (complex variable) coordinates unt!l dif- ferences in bus voltages from one iteration to another are sufficiently small Both methods are based on bus admittance equations Example 9.1, Suppose that the P- load is known at cach of the nine buses of a small power system and that synchronous generators are connected to buses G)}, a power-flow study, identify the AP and AQ mismatche! associated with cach bus. Choose bus (1) as the slack bus and the state variabl Solution, Tac nine buses of the system are categorized as follows: PO doses: @, @,@, Wand @) V buses: ®,and @ Slack bu = The mismatches corresponding to the specified P and Q are At P-Q buses: BPs BOs APSO APSO. APs. AQy, AP A Qy AtP-V buses: AP,, APs, AP;92 THEGAUSSSEIDEL METHOD 335 and the state variables are P-Q buses: 85,1Vyl; S4,1Mls 86. 1%ls 5g, Vels So, Io! PV buses: 8), 85, 85 Since N= 9 and N, = 3, there are 2N'~N, ~ 2 = 13 equations to be solved for the 13 state variables shown, 9.2. THE GAUSS-SEIDEL METHOD ‘The complexity of obtaining a formal solution for power flow ina power system arises because of the differences in the type of data specified for the different kinds of buses. Atthough the formulation of sufficient equations to match the number of unknown state variables is net difficult, as we have seen, the closed form of solution is not practical. Digital solutions of the power-flow problems follow an iterative process by assigning estimated values to the unknown bus voltages and by calculating a new value for each bus voltage from the estimated values al the other buses and the real and reactive power specified. A new set of values for the voltage at each bus is thus obtained and used to calculate still another set of bus voltages. Each calculation of a new set of voltages if called an iteration. The iterative process is repeated until the changes at each bus are less thar a specified minimum value. : We derive equations for a four-bus system and write the general equations later. With the slack bus designated as number 1, computations start with bus COTE Py ya, and Oz .¢), are the schedulec real and reactive power, respectively, entering the network at bus @), it follows from Eg, (9.4) with é set ecual to 2 and N equal to 4 that Pa sen ~ 1Q2, Ve Mm Yay + Yaals + Yaas + Yaa (9.14) 2 Solving for gives Sh (YY, + Yas + rato} (9.15) For now let us assume that buses G) and @) are also load buses with real and reactive power specified, Expressions similar to that in Eg. (9.15) may be written > we have wn JO YY, 4 Yh + nal (9.16) vy If we were to equate real and imaginary parts of Egs. (9.15), (9.16), and the336 CHAPTER 9 POWER-FLOW SOLUTIONS similar equation of bus @), we would obtain six equations in the six state variables 8, to 8, and IV] to |V;|. However we solve for the complex voltages directly from the equations as they appear. The solution proceeds by iteration based on the scheduled real and reactive power at buses (2), G), and @), the scheduled slack bus voltage V; = 1Y4|/5, and initial voltage estimates V{, V, and VO at the other buses. Solution of Eq. @.15) gives the corrected voltage V$ calculated from ay) [Pas — oly Yyo= ~ (YM MO YOY) (9.07) in which all quantities in the right-hand-side expression are either fixed specifi- cations or initial estimates. The calculated value V? and the estimated value V will not agree. Agreement would be reached to a good degree of accuracy after several iterations and would be the correct value for V- with the estimated voltages but without regard to power at the other buses. This value would nor be the solution for Vy for the specific power-flow conditions, however, because the voltages on which this calculation for ¥ depends are the estimated values V& and V{ at the other buses, and the actual voltages ar¢ not yet known. As the corrected voltage is found at each bus, it is used to calculate the corrected voltage at the next bus. Therefore, substituting V7: in Eq. (9.16), we obtain for the first calculated value at bus @) = (Va + YaQlf + YVsery (a ‘The process is repeated at bus (4) and ut euch bus consecutively throughout the network (except at the slack bus) to complete the first iteration in which calculated values are found for each state variable. Then, the entire process is carried out again and again unti! the amount of correction in voltage at every bus is less than some predetermined precision index. This process of solving the power-flow equations is known as the Gauss-Seidel iterative method. Convergence upon an erroncous solution is usually avoided if the initial values are of reasonabie magnitude and do not differ too widely in phase. It is common practice to set the initial estimates of the unknown voltages at all load buses equal to 1.0/0" per unit. Such initialization is called a flar start because of the uniform voltage profile assumed Fora systemof \ buses the general equation for the calculated voltage at any bus @ where P and Q are scheduled is y ¥,WeP! (9.19) i) jor The superscript (k)} denotes the number of the iteration in which the voltage is92 THEGAUSSSEIDEL METHOD 337 currently being calculated and (& - 1) indicates the number of the preceding iteration, Thus, we see that the values for the voltages on the right-hand side of this equation are the most recertly calculated values for the corresponding buses (or the estimated voltage if k is ] and no iteration has yet been made at that particular bus). Since Eq. (9.19) applies only at load buses where real and reactive power are specified, an additional step is necessary at voltage-controlled buses where voltage magnitude is to remain constant, Before investigating this additional step. ict us look at an exarip‘e of the calculations at a load bus, Example 9.2, Figure 92 shows he one-line dingram of a simple power system. Generators are connected at buses (1) and (4) while loads are indicated at all four buses. Base values for the transmission system are MII MVA, 230 KV. The fine data of Table 9.2 give per unit series impedances and Ine-charging suseeptances for the nominal-w equivalent of the four Lines identified by the buses at which they terminate. The bus data in Table 9.3 list values for P, O, and V at each bus, The Q values of load are calculated from the coresponéing P values assuming a power hector of O85. The net scheduled values, Pen a%d Q, soyy ATE Negative at the load buses @) and G). Generated Q,; 18 not specified where voltage O Tarch | | Oo Elm @ Pine _Naple FIGURE 9.2 nad One-line diagram for Example 9.3 showing the bus names and 7 numbers, TABLE 9.2 e data for Exa ' Shunt ¥ Line Total bus to R x G B charging ys? bus perunit — perunit — perunit per unit Mvart perunil Le COINS — Ousvae 3.8156 19st 1025 0.05125 1-3 003720 5.369861 =25,847899 775 0.03875 0.037. 5.169561 =25.847509 195 0.03875 0.01272 0.0636 3023705 = 15.1) 8528 1275, 0.06375 4 Pase JOOMVA, 230 kV. EALZAOKV,338 CHAPTERS POWER-FLOWSOLUTIONS TABLE9.3 Bus data for Example 9.2 Generation Load Bus P,MW Q,MvarP,MW @, Mvart_-perunit Remarks - - 50 3099 L00Z0" Slack bu: 2 0 0 170 105.35 10070 Load bus (inductive) 3 0 0 200 14 L00Z0? Load bus (inductive) 4 318 - 80 49.58 4102/0" Voltage cont tThe Q valuesof load are calculated from the eorzesponding P values assuming a power fictor of 85. magnitude is constant, In the voltage column the values for the load buses are flat-start estimates. The slack bus voltage magnitude |V,| and angle 6,, and also magnitude [V1 at bus 4), are to be kept constant at the values listed. A. power-flow study is to be made by | the Gauss-Seidei method. Assuming that the iterative calculations start at bus (2), find the value of V, for the first iteration Solution. In order to approach the accuracy of a digital computer, we perform the computations indicated below to six decimal places. From the line data given in Table 9.2 we construct the system Y,,,, of Table 9.4. For example, associated with bus @) of Fie. 9.2 are the nonzero off-diagonal elements a, and Yq, which are equal to the negative of the respective line admittances (3.8159 ~ j19.078144); Fan = (5.169561 — j25.847809) Since Yq is the sum of all the admittances connected to dus (2), including the TABLE 9.4 Bus admittance matrix for Example 9.24 Bus a no. ® ® 7 8.985190 = 3815629 =5.169561 6 vD ~444.835953, + 519.078144 $425.847509 > ~ 3.819629 8.985190 6 = 5.169561 @ 1 19.78144 =)44.835953 4+)25.847809 ~ 5.169561 , 8.193267 3.023705, +325.847809 ° —j40.863838 4 j15.118528, > ~5.169561 ~3.023705 8.193267 ® ° +)25.847809 +715.118528 —)40,863838 4 Per-unit values rounded to six decimal places92 THEGAUSSSFIDELMETHOD 339 shunt susceptances for line charging of lines (2)- @) and @)- G), we have Yay = (= ¥o1) + (0.05125 + (= ¥54) = f0.03875 = 8.985190 ~ j44.835953 Substitution in Eq. (9.17) yields the per-vnit voltage 217+ /10535 0; that is, until |A.x,| and |4.x| are both less than e. The concepts underlying the Newton-Raphson method are now exemplified numerically. Example 9.4. Using the Newton-Raphson method, solve for x, and x, of the nonlinear equations Bi X 1) X20) = Ay(y, 2,4) — by = 4uxz sin x, + 0.6 = 4x3 — dur, cos x, +0.3= 0° eal ys Fart) = hati ta) ~93 THENEWTONRAPHSONMETHOD 345, Treat the parametet w as a fixed mumber equal to 1,0, and choose the initial conditions 2 = Orad and x® = 1,0. The precision index is 107°. Solution. Partial differentiation with respect to the x's yields 8B, 98s — 410, 008 Susin xy j= 5 4urysin xy 8x5 — Au cOsx Ilere the parameter w has a fixed value equal to 1.0, but in some studies it could be treated as @ specifiable or control variable. First iteration. Setting « = 1.0 and using the initial estimates of x, and x, we calculate the mismatches + dcas(0) = -03 which we use in Eq. (9.34) to yield the mismatch equations 4c08(0) 4sin(0) 52) [-06 | ~4cos(0)}/ax®] | 03 Inverting this s'mple 2 x 2 matrix, we determine the initial corrections far] [4 0] 'P-o6) [-o10) dx o4 0.3} | 0.075 which provide first iteration values of xy and 2x2 as follows: xP xf) = Sx = 0.04 (0.150) = —0.150rad rasa M + axM = 1.0 + (0.075) = 0.95 corrections exceed the specified tolerance, and so we continue Second iteration, The new mismatches are ‘The go —0.6 — 4(0.925)sin( - 0.15) —0.087079 ae | -03- + 4(0.925)cos( - 0.15) ~ 0.064047346 cHaPrER9 POWERFLOW SOLUTIONS and updating the jacobian, we compute the new cotrections Ax 3.658453 —().597753 =0.017079 0.016335 7 0.064047) ~ | 0.021214 Ax? — 0.552921 3.444916 These corrections also exceed the precision index, and so we move On to the next iteration with the new corrected walucs x) = -0.150 + (0.016335) = -0.166335 rad x@ = 0.925 + (—0.021214) 0.903786 Continuing on to the third iteration, we find that the corrections Ax and Ax are each smaller in magnitude than the stipulated tolerance of 107%. Accordingly, we calculate the solution xO = -0.166876rad; —_x$? = 0.903057 The resultant mismatches aze insignificant as may be easily checked. In this example we have actually solved our first power-fiow problem by the Newton-Raphson method, This is because the two nonlinear equations of the example are the power-flow equations for the simple system shown in Fig. 9.3, Bay Sa) = Pa( 41) 220K) ~ (Pee ~ Par) = 4\V,) [VzIsin 5, + 0.6 = 0 (9.36) Bx 44) % 254) = O2( 24, 42,4) ~ (Opn > Oar) = 4|V,)? — 41) Vyleos 8, + 0.3 = 0 (9.37) oO @ | Vile I all {Qe fos To load FIGURE 9.3 The system with power-flow equations corresponding to those of Eamaple 9.4EWTON-RAPHSON POWER. FLOWSOLUTION 347 where x, represents the angle 8, and x2 represents the voltage mag at bus @). The control w denotes the voltage magnitude || of the slack bus, and by changing its value from the specified value of 1.0 per unit, we may control the solution to the problem. In this textbook we do not irestigate this control characteristic but, rather, concentrate on the application of the Newton-Raphson procedure in power-flow stucies. nitude IV5! 9.4 THE NEWTON-RAPHSON POWER-FLOW SOLUTION To apply the Newton-Raphson method to the solution of the powerflow equations, we express bus vo tages and line admittances in polar form. When 7 is set equal to ¢ in Eqs. (9.6) and (9.7) and the corresponding terms are separated from the summations, we obtain N P= IG + LAW EY, |cost,, + 6, ~ 6,) (9.38) Nw Q,= AWB, = LIME sin Wye +8, = 8) (9.39) nal These equations can be readily differentiated with respect to voltage a magnitudes. The terms involving G, and B,, come from the definition of ¥,, in Eq (9.1) and the fact that the angle (6, — 4,) is zero when 2 = Let us postpone consideration of voltage-controlled buses for now and regard all buses (except the slack bus) as load buses with known demands Py: and Q,, The slack bus has specified values for 6, anc [M4], and each of the other buses in the network has the two state variables 6, and || to be calculated in the power-flow solution. The known values of Py, and Ou correspond to the negative of the & constants shown in Eqs. (9.27)and (9.28), as onstrated in Example 9.4. At cach of the nonslack buses estimated values of 6, and |M| correspond to the estimates x and x6? in the preceding section. Correspondence to the Ag mismatches of Eq. (9.34) follows from Eqs. (9.8) and (9.9) by writing the power mismatches for the typical load bus @, AP, = Pose Piccate (9.40) AQ; = Qian > @ (4) For simplicity sake, we now write mismatch equations for a four-bus system, and it wil! become obvious how to extend those equations to systems with more than four buses . ’348 cHapreRs rOweR-sLOW SOLUTIONS For real power P, we have AP, a a6 ei b6 i ar I == + — Ab, + A8,+ — AI, 05,7? a3 86, IYI? ap, +a ALV,| + = (942) The last three terns can be multiplied and divided by their respective voltage magnitudes without altering their values, and so we obtain pw egg a eas 4 Oh wy 2 Po =" Bd, + —— AS, + — AS, + IV,}——— 5 06,7 7 a5, 8 By a VA yj oe SIL yy) 2h ale + hie + Miso 9AS “ava Wal lay Tal (8) There are advantages to this form of the equation as we shal] soon see. A similar mismatch equation can be written for reactive power Q,, 9Q; 9Q; 3Q; 6Q, AlMsi AQ, = a8, + a8, + =! Aa, + HIE 85, 085 05, aval {¥z) @Q, AlYs! aQ, AIM} #1 ee Ut jy SO (9.44) av) VI avi Yi Each nonslack bus of the system has two equations like those for 4 P, an Collecting all the mismatch equations into vector-matrix form yields oP, oP, j a oo » AP, 36, a6 * Ps AQ, AiV,) AQs mar | | a Jacobian Corrections Mtismafches (95)9.4 THENEWTON.RAPHSON POWER FLOW SOLUTION 349. We cannot include mismatches for the slack bus since AP, and AQ, are undefined when P and Q, are not scheduled. We alo omit all terms involving Ad, and A!Y,| from the equations because those corrections are both Zero at the slack bus. ‘The partitioned form of Eq. (945) emphasizes the four different types of partial derivatives which enter into the jacobian J. The elements of Jy. and Jay have voltage-magnitude multipliers because a simpler and more symmetrical jacobian results. In choosing this format, we have usec the identity aly ep. ) Al = a XM (9.46) and the corrections become AlVj| / |). as shown rather than AVI The solution of Eg. (9.45) is found by iteration as follows: « Estimate values 6 and || for the state variables + Use the estimates to calculate PO. and O.. from Eqs. (9.38) and (9.39), mismatches 4P and AQ from Eqs. (9.40) and (9.40), and the partial cerivative elements of the jacobien J. + Solve Eg, (9.45) for the initial corrections 46 and VIP [VI * Add the solved corections to the initial estimates to obtain 30 = 5 4 Aa! (9.47) w|i + ype (IO = r+ aly, (9.48) ae + Use the new values 6!) and (¥!"" as starting values for iteration 2 and continue In more general terms, the update formulas for the st variables are ng values of the state BD = Bf) > ABit? (9.49) IYI = IO 4 ayy® = [YIM] +350 CHAPTER 9 POWER.HLOW SOLUTIONS For the four-bus system submatrix J,, has the form [aP, AP, aP, 0, 4, 9 oP, oP, aP, 9, 4, 5, oP, OP, oP, I= (251) Expressions for the elements of thts equation are casily found by difierentiating the appropriate number of terms in Eg. (9.38), When the variatle 7 is the particular value j, only one of the cosine terms in the summation of Eg, 0.38) contains 6,, and by partial differentiating that single term with respect 10 6,, we obtain the typical off-diagonal element of Jy, On the other hand, every term in the summation of Eq. (9.38) contains 6,, and so the typical diagonal element of Jj, is By comparing this expression and that for Q, in Ee. (9. J, we obtain ap, a —WwPR 3b; VB Ina quite similar manner, we can derive formulas for the Ja as follows: nenis of submatrix — VV, ¥,,leos(@,, + 8; — Yi ) (9.55) : woo) LIKY,Y,leos(@,, +6, - 6.) = — L aoe net nt Fn (9.56)94 THENEWTON RAPHSONPOWER.FLOWSOLUTION 351 _ Comparing this equation for 0Q,/é8, with Eq. (9.38) for P., we can show that (957) The elements of submatrix Jjz are easily found by first finding the expression for the derivative dP, /a|,| and then multiplying by 1) to obtain aP, Vi ging ~ WEY, eos(@, ~ 8-8) (9.58) Comparison with Fig, (9.55) yields =--= (9.59) This is a most useful result, for it seduces the computation involved in forming the jacobian since the off-diagonal elements of J, are now simply the negatives of the corresponding elements in J... This fact would not have become apparent if we had not multiplied dP, /al/] by the magnitude |W] in Eq. (9.43). In an analogous manner, the diagonal elements of Jy are found to be ap, Nw = VAG + YL WaYinleos(8jn + 8, = 8))} (9.60) Wil a av ell and comparing this result with Eqs. (9.56) and (9,57), we arrive at the formula ea OV, P+ WG, (9.61) Finally, the off-diagonal and di are determined to be nal elements of submatrix J,, of the jacobian 90, . - “lav = = IV YY, Isin(8,; + 3; (9.62) a oP, TAI QL -— — 21V,B, = O, - VIB, (9.63) “av 05;i4 H 352 CHAPTER 9 POWER-FLOW SOLUTIONS Let us now bri tions’ Offiagonal elements, #5 together the results developed above in the following defini- . (9.64) (9.65) Diagonal elements, AQ, yi = -M,-- 211,78 (9.66) uy, Wi ap 2 Ma IED N,& 3, =N,t 2VPGy (9.67) iE a5, “i : 4 Intemelationships among the elements in the four submatrices of the jacobian are more clearly scen if we uSe the definitions to rewrite Eq. (9.45) in the following form: Mz Mx My | No + 2IKa?Gn =Nay Nau My May Maa Ny Nay + DWAPG ay N45 Ma Ma Mas | “Ne Na Nye + 21¥, Na Nay Nex | Mr ~21Y)IBoa M. Mh Ny Nyy Nyg My = My — 214,178 My Nao Nay Nye May My = My, ~ 214,7°B, as, | far. M53 AP, by AP, 7 7 x _ —| = |— (9.68) AlYI/ Val 40, Alls V3) AQ; alvisiyi | [8s94 THE NIWTON RAPHSON roweR-FLowsOLUTION 353 So far we have regarded all nonslack buses as load buses. We now consider voltage-controlled buses also Voltage-controlled buses. In the polar form of the power-How equations vollage-controlled buses are easily taken into account. For example, if bus @ of the four-bus system is voltage controlled, then |V/| has a specified constant value and the voltage correction AlV,)/ IV) must always be zero. Conse- quently, the sixth column of the jacobian of Eq. (968) is alvays multiplied by zero, and so it may be removed, Furthermore, since Q, is not specified, the mismatch AQ, cannot be defined, and so we must omit the sixth row of Eq. (9.68) corresponding to Q,. Of course, Q, can be calculated after the power-flow solution becomes available. In the general case if there are voltage-contralled buses besides the « bus, a row and column for each such bus is omitted from the polar form of stem jacobian, which then has (2M — A, — 2) rows and (2N — N= 2) the columns consistent with Table 9.1 Example 9.8. The smal] power system of Example 9.2 has the fine data and bus Gata given in Tables 9.2and9.3,A studyo! the system is to be made by the Newlon-Raphson nietiod using the polar form af the equations for P and Q. Determine the number of rows and columns in the Jacobian, Calculate the initial mismatch AP and the initial values of the jacobian elements of the (second row, third column); of the (second row, second column): and of the (fifth row, fifth column). Use the specified values and initial voltage estimates shown in Table 9.3 & bus has no rows or columns in the jacobian, a 6 x 6 matrix would be necessary if P and Q were specified for the remaining three In fact, however, the voltage magnitude is specified (held constant) at bus }, and thus the jacobien will be a 5X 5 matrix. In order to calculate Ps.cac based on the estimated and the specified voltages of Tab c 9.3, we need the polar form af the alf- AQ; | (9.87) Alvi SWAY IB, -IYMiBas — MaYal Baa wal AQ, To show how the voltages are removed from the entries in the coefficient matrix of Eq. (9.87), let us multiply the first row by the correction vector and then divide the resultant equation by [V4] to obtain AQ -By AIV| — (9.88) Ws] = BaelY) = ves Wal The coefficients in this equation are constants equal to the negative of the susceptances in the row of ¥j,,, Corresponding to bus (2). Each row of Eq. (9.87) can be similarly treated by representing the reactive mismatch at bus () by the quantity AQ,/ IV. All the entries in the coefficient matrix of Eq, (9.87) become constants given by the known susceptances of Yj. We can also modify Eq. (9.86) by multiplying the first row by the vector of angle corrections andae 9.9 THEDECOUPLEDPOWERFLOWMETHOD 371 rearranging the result to obtain AP, IWalBoy 832 ~ IV5IBos M8 ~ iIB. 88, = Fp (9.89) _ ‘Vhe coefficients in this equation can be made the same as those in Eq, (9.88) by setting |V5|, IVs, and |V,| equal to 1.0 per unit in the left-hand-side expression. Note that the quantity AP,/|%| represents the real-power mismatch in Eq, (989). Treating all the rows of Eq, (9.86) in a similar manner leads to two decoupled systems of equations for the four-bus network [By -B» -By -Bs 90) : a ; | AQ, By —Byy —Bp: ][S1Vai Al - -By —-B yifo} and By ~By —B:s || AlVs Tal (9.91) AO, |S By ~ Bis | al) Wil B Metrix B is generally symmetrical ancl sparse with nonzero elements, which are constant, real numbers exactly equal to the negarive of the susceptances of Yyus- Consequently, mattix B is casily formed and its tiangular factors, once com- puted at the beginning of the solution, do not have to be recomputed, which leads to very fast iterations. At voltage-controlied buses Q is not specified and AV is 2ero; the rows and columns corresponding to such buses are omitted from Fe. (9.91), One typical solution strategy is to: 1, Calculate the initial mismatches 4P/ INI, 2. Solve Eq, (9.90) for & 3. Update the angles 6 and use them to calculate mismatches AQ/ |V| 4. Sclve Eq. (9.91) for &/V’| and update the magnitudes |V|, and 5 . Return to Eq. (9.90) to repeat the iteration until all mismatches are within specified tolerances.372 CHAPTERS POWER-FLOW SOLUTIONS Using this decoupled version of the Newton-Raphson procedure, faster power- flow solutions may be found within a specified degree ot precision Example 9.9. Using the decoupled form of the Newton-Raphson method, determine the first-iteration solution to the power-flow problem of Example 9.5, Solution. Tne B matrix can be read directly from Table 9.4 and the mismatches corresponding to the initial voltage estimates are alwady calculated in Example 9.5 so that Eq, (9.90) becomes 44835953 QO +28 789 | 36, = 1.59661] 0 40.8638 ~ 15.118528 |} as, | = | - 1.93953} ~25.847809 15118528 40.803838 |] a5, 2.21280 | Solving this equation gives the angle corrections in radians 36) = -0.02057; Ad, = =0.03781; 48, = 0.02609 Adding these resuits to the fat-start estimates of Table 93 gnes the upcates values of 5, 55, and 54, which we then use along with the elements of Yq, to calculate the reactive mismatches AQ, 1 Tym pry (zen 7 Qavcate PAN vai! sn Qa) 1 | Oran [=I al'Bag ~ WeMVelsin(@ig + 8) ~ Wl ~ lFawVa¥isin(B2s + 85 ~ 82)) | ~ 1.0535 + 1.02( -44.835953) + 19.455965 \ = Trop) sin(101.30093 x 97/280 10 + 0.02097) + 26 359695 | 1.02sin(101.30993 x 27/180 + 0.02609 + 0.02057) | = -0.80370 per unit AQ, J 1 - ih 7 rq 2ren Os 1 [Qssen~ [-IViI?Bs9 — [¥aMVaisin( ys +8, ~ 6) Wl =1¥yYyYlsin(Os, + 84 83)] | 1 [1.28 94 + 1,02( -40.863838) + 26.359695 "im \ sin(101.30993 x 7/180 + 0 + 0.03781) + 15.417934 1,02sin(101.30993 x 7/180 + 0.02609 + 0.03781) ~ 1.27684 per unit97 ‘THE DECOUPLED PowER-FLOWMETHOD 373 A reactive mismatch calailation is not required for bus (@), which is voltage controlled. Accordingly, in this example Eq, (9.91) becomes 44835953 0 AlV, 0 40.863838 |] AIM sl 0.80370 * | 1.27684 which yields the solutions [V1 = voltage magnitude at buses (2) and (3) are IV. 0.96875, which completes the first iteration. Updated mismatel ation of Eq, (9.90) are calculated using the new voltage values. Repeating the procedure over a number of iterations vields the same solution as tabulated in Fig. 9.4. 5. The new Often in industry-based progra (90) and YD. The mouil ms certain modifications are made in Eqs ns to B in Eg. (9.9) are generally as follows: + Omit the angle-shifting effects of phase shifters from B by setting f= 1.0/0° When rows ane columns for voltage-controlled buses are also omitted a8 previously indicated, the resulting matrix is called B The coef Eq. (9.90) is generally modified as follows: * Omit from B those elements that mainly affect megavar flows such as shunt cupacitors and reactors, and set taps ¢ of off-nominal transformers equal to 1 Also, ignore series resistances in the equivalent-7 circuits of the transmission lines in forming Yj, from which B in Faq. (9.90) is obtained, The resulting matrix is called BY, When B in Eg, (9.90) is replaced by B’, the model becomes a lossless network. If, in addition, all dus voltages arc assumed to remain constant at nominal valves of 1.0 per unit, the so-called de power flow model results. Under these additional assumptions, Eq, (9.91) is not necessary (since A|l/,) = 0 at cach bus ()) and Liq. (9.90) for the de power flow becomes. -By -Br, ~By,|[ 48, AP, By, ~ By —Bay || 48, |= | APs (9.92) —By —Byy ~ Bar |] 48s AP, where it is understood that the elements of B’ are calculated assuming all lines re lossless, De power-flow analysis can be used where approximate solutions are acceptable, as in the contingency studies discussed in Chap. 14.374 CHAPTERS POWER-FLOW SOLUTIONS, 9.8 SUMMARY This chapter explains the power-flow problem, which is the determination of the voltage magnitude and angle at each bus of the power network under specified conditions of operation. The Gauss-Seidel and Newton-Raphson iterative pro- cedures for solving the power-flow problem are described and numerically exemplified In addition to discussing how power-fow studies are undertaken, some methods of real and reactive power-flow contiol are presented, The results of paralleling two transformers when the voltage-magnitude ratios are different or when one transformer provides a phase shift are examined. Equations are developed for the nodal admittances of these transformers. and equivalent circuits which allow analysis of reactive-power contro! are introduced The computer-based power-fiow program can be used to study the appli- cation of capacitors at a load bus simply by incorporating the shunt admittance of the capacitor into the system Y,,,,. Voitage control a: a generato: bus can also be investigated by specifying the values of the PV’ bus voltages Fast, approximate methods for solving the power-flaw problem are intio- duced by means of the de power-flow model, which depends on the linkages between real power P and voltage angle 6, and reactve power Q) and voltage magnitude. ‘Table 9.5 summarizes the equations for each of the methods of power-flow analysis. TABLE 9.5 . Summary of power-flow equations and solution methods Power-flow equations | Pp= Y. I% dl 60s10,, + Q-- rail AP, = Prsty — Prcate S0;= Gauss-Seidel method | To oltain V at a bus fwithknown P and Q: 1 Pian ~IQisch ‘St i ee -'S ye FS yy ie) = Tooblain Q ata regulated bus / Use ofacceleration factor or at bus ¢ in iteration & HER = (aE) + » Lywes Lyre VERO HEY ~ VEDTABLE 95 (Contined) 98 summary 375 Newton-Raphson method a 5p Py 33 HILLY leosltyy +3, © 6) State variable update formulas: apy gh 4 35H oPy - Wl |] a8, ar aval Ja 5 ; Mby AP Y, lv! 20 V2 Jn = BIMyt “ yl Qn ab, OP, ae baie ctl oo 0, & ag, QA gg oy, 7 Dag, RO PG a 21V,PG Regt tag TMG. 2. 2, Wen ay - 210, Wi ys all Decovalzdd power Row solution technique Bro — By Bs By Biys Boy Bay Bue aly = 1H) : 5 = IK (\- Ta 36; paps iyi ABy Wal aby | [a ee/ Hal alval S0z/ Wal - all AOS IWsd al | [S0u/ Mel 7 B,, are the imaginary parts of the corresponding Y,,, elements. Voltage-cantrolled buses are not represented in &|V| equations. ~376 CHAPTER POWERFLOW SOLUTIONS: PROBLEMS 94. In Example 9.3 suppose that the generator’s maximum reactive power generation atbus is limited to 125 Mvar, Recompute the first-iteration value of the voltage at bus @ using the Gauss-Seidel method 92, For the system of Fig. 9.2, complete the second iteration of the Gauss-Seidel! procedure using the first-iteration bus voltages obtained in Examples 9.2 and 9.3 Assume an acceleration factor of 1 6. 9.3. A synchronous condenser, whose reactive power capability is assumed to be undmited, is installed at Joad bus @) of the system of Exumple 9.2 to hold ibe busvoltage magnitude al 0.99 per unit, Using the GaussScidel method, find the voltages at buses @) and 3) for the first iteration 9.4. Take Fig. 9.12 as the equivalentar representation of the (ransmission line betw bus G) and bus (4) of the system of Fig. 9.2. Using the power-low solution in Pig. 94, determine ane indicate on Fig. 9.12 the vies of (a) P and Q leavin buse G) and G) on line G)- G), (6) charging megavars of the equivalent 7 of line GG) and(c) P and Q atbothend of the series part of 12 equivalent wen ® ——-—, =} T FIGURE 9.12 Diagram for Prob. 94 9.8. From the line flow information of the power-fiow solution given in Fig. 94 determine /?R loss in cach of the four transmission tines, and verify that the sum ‘of these line losses is cqual to the total system loss of 4.81 MW 9.6. Suppose that a shunt capacitor bank rated 18 Mvar 's connected between bus (3) and the reference node in t m of Example 9.5. Medify the Y,,. given in Table 9.4 to account for this capacitor, and estimate the actual megavar reactive power injected into the system from the capacitor 9.7. For the system of Example 9.5 augmented with a synchronous condenser as described in Prob. 9.3, find the jacobian calculated at the initialestimates. Hin: It would be simpler 10 modify the jacabian matzix shown in See. 94 following Example 9.5 than to start calculations from the begicning, 9.8. Suppose that in Fig. 9.7 the tap is on the side of node @) so thatthe transforma- tion ratio is ¢:1. Find elementt of Y,.,. similar to those in Eg. (9.74), and draw the equivalentar representation similar to Fig. 9.8. 9.9. In the four-bu system of Example 9.5 suppose that a magnitude-regulating transformer with 0.2 per-unit reactance is inserted between the load and the bus at bus @), as shown in Fig, 9.10, The variable tap is on the load side of the transformer, If the voltage magnitude at the new load bus () is prespecified, and9.10. 91. 9.12 9.13, 914. proptems 377 therefore is not 2 state variable, the variable tap ¢ of the transformer’should be regarded as a state variable. The Newton-Raphson method is to be applied to the solution of the power-flow equation (a) Write mismatch equations for thi problem in a symbolic form similar to Eq. (9.45) (b) Write equations of the jacobian elements of the column corresponding to variable £ (that is, partial derivatives with respect 10 £), and evaluate them using the initial voltage estimate shown in Table 9.3 and asuming that the voltage magnitude at bus (5) is specified to be 0.97 The initial estimate of 55 is 0. (e) Write equations of P and Q mismatches at bus (S) and evaluate them for the first iteration. Assume the initial estimate of variab’e 1 is 1.0 If the tap setting of the transformer of Prob. 9.9 is prespecified instead of the voltage magnitude at bus (3), then M% should be resarded as a state variable. Suppose that the tap setting ¢ is specified to be If (a) Th this case write mismatch equations in at symbolic form similar to Eq, (9.49). (L) Write equations ef the jacobian elements which are partial derivatives with espect to Il, and evaluate them using the initial estimates. The initial estimate of Ve is 10/0" (c) Write equat ons for the ? and Q mismatches at bus ©), and evaluate them for the first iteration Redo Example 9.8 fo: 1.0/ —3°, and compare the two results as to changes in real and reactive power flows, The generator at bus (4) of the system of Example 9.5 is to be represented by a generator connected to bus (@) through a generator step-up transformer, as shawn in Fig. 9.13. The reactance of this transformer is 0.02 per unit; the tap is on the n-voltage, side of the transformer with the off-nominal turns ratio of 1.05. Evaluate the jacobian elements of the rows corresponding to buses (4) and (5) Ay an ® FIGURE 9.13 For the system of Prob. 9.12, find matrices Band B” for use in the decoupled power-fiow method A five-bus power system is shown in Fig. 9.14. The tine, bus, transformer, and capacitor data are given in Tables 9.6, 9.7, 98, and 9.9, respectively. Use the Gauss-Seidel method to find the bus voltages for the first iteration.378 CHAPTERS POWERFLOW SOLUTIONS FIGURE 914 ® System for Pros, 9.14 thous T A) 9.18. The line anc bus daca are given in Tables 9.6 though 9.9 TABLE 9.6 Line data for the system of Figure 9.14 Persunit Per-unit ate eres Charging R Mvar ®@ 0.0108 0.0649 25 6.6 QD @ 0.0235 Guat 2 40 @©® cons aor7l 50 70 @®Q cam o0sse 49 80 @o® 0.08 6.0529 40 ~18 60 TABLE 9.7 . Bus data for the system of Figure 9.14 Generation Load Y Bus POMW! @Ghad PIMW) QiMvan) pu. Resnark D LOiZ0" Shack bus @ 60) 35 10/0" ® 10 42 10Z0° @® 80 50 10/08 © 190 65 36 10/0 PV bs TABLE 9.8 Transformer data for the system of Figure 9.14 Transformer bustobus Per-unitreactance Tap setting aos 0975prosiems 379 TARLE 9.9 Capacitor data for the system of Figure 9.14 Rutingin Bus Mvar Q 18 @ 5 9.15, To spply the Newtnn-Raphson method to the power-flow solution of the system of 9.16. ols. 919, 9.201 Fig. 9.14, determine (a) Yj, of the system, () the m smatch equation at bus evaluated at the initial vollage estimates of Table 9.7 for the first iteration, and (c) write seh equations in a form similar to Eq. (5.85) For the yen of Pig 914 find matrices BY and BP for use in the decoupled powerflow riethod, Abo. determing the firstateration # and QO mismatch equa- tions at bus C3), and find the voltage magnitude at bas G) at the end of the first Neration. pose that in Fig. 9.14 the transiormer belween beges (2) and () is a phase shifter where ¢ is now a complex variable and is 1.0/ — 2°. (a) Find ¥,,y, of this system, () When compared with the power-flow souron of Prob. 9.15, will the real power flow in the line fram bus (5) to bus (3) inerease ar decrease? What about the reactive power flow? Explain why qualitatwely ‘To apply she decoupted powerfiow method to the sysiem of Prob. 9.17, find matrices B’ and B Redo Exemple 9.10 when an 18-Mvar shunt capacitor bank 1s added to bus G) In applying the Newton-Raphson method, if the amount of reactive power recuired to maintain the specified voltage at a PV’ bus exceeds the maximum limit of its reactive power generation capability, the reactive power at that bus is set to that limit and the type of bus becomes a load bus. Suppose the maximum reactive power generation at us G) is limited to 150 Mvar ia the system of Example 9.5 Using the frst-iteration result given in Sec. 9.4 following Example 9.5, determine pether ar not the type of by (2) should be converted to & load bus at the start of tue scound iteration. If so, calculate the reactive power mismatch at bus (4) that show d be used in the sceond-iteration mismateh equabon.CHAPTER 10 | SYMMEIRICAL FAULTS A fault in a circuit is any failure which interferes with the normal flow of current. Most faults on transmission lines of 115 kV and higher are caused by lightning, which results in the flashover of insulators. The high voltage beween a conductor and the grounded supporting tower causes ionization, which pro- vides a path to ground for the charge induced by the lightningstroke. Once the ionized path to ground is established, the resultant low impedance to ground allows the flow of power current from the conductor to ground and through the ground to the grounded neutral of a transformer or generator, thus completing the circuit. Line-to-line faults not involving ground are Jess common. Opening circuit breakers to isolate the faulted portion of the line from the rest of the system interrupts the flow of current in the ionized path and allows deionization to take place. After an interval of about 20 cycles to allow deionization, breakers can usually be reclosed without reestablishing the arc. Experience in the operation of transmission lines has shown that ultra-high-speed reclosing breakers successfully reclose after most faults. Of those cases where reclosure is not successful, many are caused by permanent faults where reclosure would be impossible regardiess of the interval between opening and reclosing. Permanent faults are caused by lines being on the ground, by insulator strings breaking because of ice loads, by permanent damage to towers, and by surge-arrester failures. Experience has shown that between 70 and 80% of transmission-line faults are single line-t0-ground faults, which arise from the fiashover of only one line to the tower and ground. Roughly 5% of all faults involve all three phases. These are the so-called symmetrical three-phase faults which are considered in this chapter. Other types of transmission-line faults are ine-to-line faults, which do not involve ground, and double line-to-ground faults. All the above faults 280,10) TRANSIENTS IN KESERUSCIRCUITS 381 except the three-phase type cause an imbalance bewween the phases, ard so they are called unsymmerrical faulis, These are considered in Chap. 12 The currents which flow in different parts of a power system immediately ter the occurrence of a fault differ from those flowing a few cydes later just before circuit breakers are called upon to open the line on both sides of the fault. Andall of these currents differ widely from the currents which would flow under steady-state conditions if the fault were not isolated from the rest of the system by the operation of circuit breakers. Two of the factors on which the proper selection of circuit breakers depends are the current flowing immediately after the fault occurs and the current which the breaker must interrupt. In faul analysis values of these currents are calculated for the different types of faults at various locations in the system. The data obtained from fault calculations also serve to determine the settings of relays which control the circuit breakers. 10.1 TRANSIENTS IN RL SERIES CIRCUITS The selection of a circuit breaker for a power system depends not only on the eurent the breaker :s to cary under normal operating conditions, but also on the maximum current it may have to carry momentayly and the current it may have to pverupr at the voltage of the line in which it is placed In order to approach the problem of calculating the initial current when a system is short-circuited, consider what happens when an ac voltage is applied to a circuit containing constant values of resistance and inductance. Let the applied voltage be M,, sinlor + a), where ¢ is zero at the time of applying the vowtage. Then, @ determines the magnitude of the voltage when the circuit is closed. If the instantaneous voltage is zero and increasing in a positive direction when it is applied by closing a switch, @ is zero. If the voltage is at its positive maximum instantancous value, @ is 7/2. The differential equation is di Vi. sin(@r tc) = REEL (10.1) dt The solution of this equation is v i a [sin(wor + a — 8) ~ 7 * /* singe — 8) (10.2) where IZ| = YR? + (wf) and @ = tan (wL/R). Phe first term of Eq. (10.2) varies s:musoidally with time. The second term is nonperiodie and decays exponentially with a time constant L/R. This nonperiodic term is called the de component of the current. We recognize the sinusoidal term as the steacy-state value of the current in aa RE circuit for the given applied voltage. If the value of the steady-state term is not zero when t= 0, the de component appears in the solution in order to satisfy the physical382 CHAPTER 10 SYMMETRICAL FAULTS FIGURE 10.1 . Currest a5. fuection of time in an RL circuit for: (@) @ == (O) @—0 = ~ 2/2, where tan Gal/RY The woltage is Ving Sloot * ee) appligd ate = 0. condition of zero current at the instant of clasing the switch. Note that the de term does not exist if the circuit is closed at a point on the voltage wave such thata — @=0ora— m. Figure J0.1(@) shows the variation of current with time according to Eq. (10.2) when a — 6 = 0. If the switch is closed at a point on the voltage wave such that @ — @ = +77/2, the de component has its maximum initial value, which is equal to the maximum value of the sinusoida! component. Figure 10.1(5) shows current versus time when @—@= —7/2 The de component may have any value from 0 to V,.,./ IZ], depending on the instantaneous value of the voltage when the circuit is closed and on the power factor of the cirauit. At the instant of applying the voltage the de and steady-state components always have the same magnitude but are opposite in sign in order to express the zero value of current then existing. In Chap. 3 we discussed the principles of Operation of a synchronous generator consisting of a rotating magnetic field which generates a voltage in an armature winding having resistance and reactance. The current flowing when a generator is short-circuited is similar to that flowing when an alternating voltage is suddenly applied to a resistance and an inductance in series. There are important differences, however, because the currents in the damper windings and the armature affect the rotating field, as discussed in Secs. 3.8 and 3.9, If the de component of currentis eliminated from the short-circuit current of each armature phase, the resulting plot of each phase current versus time is that shown in Fig. 3.19. Comparison of Figs. 3.19 and 10.i(@) shows the difference between applying a voltage to the ordinary RL circuit and applying a short circuit to a synchronous machine. There is no de component in either of these figures, yet the current envelopes are quite different. In a synchronous machine the flux across the air gapis not the same at the instant the short circuit occurs as it is a few cycles later. The change of flux is determined by the combined action of the field, the armature, and the damper windings or iron parts of the round rotor. After a fault occurs, the subtransient, transient, and steady-state Periods are characterized by the subtransient reactance X!, the transient182. INTERNAL VOLTAGES OF LOADED MACHINES UNDER FAULT CONDITIONS 383, reactance X',, and the steady-state reactance X ,, respectively. These reactances have increasing values (that is, ¥% <.X', <.X,) and the corresponding compo- nents of the short-circuit current have decreasing magnitudes U/l > 17/1 > UD. With the de component removed, the initial symmetrical rms current is the rms value of the ac component of the fault current immediately after the fault occurs, In analytical work the interna] voltage to the machine and the subtran- sient, transient, and steady-state currents may be expressed as phasors. The voltage induced in the armature windings just after the fault occurs differs from that which exists after steady state is reached. We account for the differences in induced votage by using the different reactances (A%, X/, and X,) in series with the internal voltage to calculate currents for subtransient, transient, and steady-state conditions. If a generator is unloaded when the fault occurs, the machine is represented hy the no-load voltage to neutral in series with the proper reactinge, as shown in Pig, 3.20. The resistance ts taken into account if greater accuracy ts desired. I there is impedance external to the generator yetween its terminats and the short circuit, the external impedance must be included in the circuit We shall examine the transients for machines carrying a load in the next section Although machine reactances are not true constants of the machine and depend on the degree of saturation of the magnetic circuit, their values usually lie within certain limits and can be predicted for various types of machin Table A2 in the Appendix gives typical values of machine reactances that are needed in maxing fault calculations and instability studies, In general, subtran- sient reactances of generators and motors are used to determine the initial current flowing on the occurrence of a short circuit. For determining the intermpting capacity of circuit breakers, except those which open instanta- neously, subtransient reactance is used for generators and transient reactance is used for synchronavs motors. In stability studies, where the problem is to determine whether @ fault will cause a machine to Jose synchronism with the rest of the system if the fault is removed after a certain time interval, transient reactances apply 10.2. INTERNAL VOLTAGES OF LOADED MACHINES UNDER FAULT CONDITIONS Let us consider a generator (rat's leaded when a fault occurs, Figure 10.2(a) is the equiveleat citcuit Of & generator that has a balanced three-phase load. Interal voltages and reactances ol the generar are now identified by the sudscript g since some of the circuits to be considered also have motors. External impedance is shown between the gencrator terminals anid the point P where the faultoccurs. The current flowing before the fault occurs at point P is 7, the voltage at the fault ts ¥,, and the terminal voltage of the generator is V,. The steady-state equivaicnt circuit of the synchronous generator is its no-load voltage £, in series with its synchronous reactance %,,. If a three-phase fault3840 CHAPTER 10. SYMMETRICAL FAULTS, Lowe Fees Keg |" a (a) FIGURE 10.2 Equivalent circuit for a generator supplying a balanced three-phitse Joa Application of a phase fiavit at P is simulated by closing switch S, (a) useal steadlystite generator equivalent circ: joad: (b) circuit for calculation of 1° occurs at point P. we see thata short circuit from P to neutral in the equivalent circuit does not satisfy the conditions for calculating subtransient current, tor the reactance of the generator must be X¥,, if we are calculating subtransient current J" or X), ifwe are calculating transient current /'. The circuit shown in Fig. 10.X%6) gives us the desired result. Here a voltage £7 in series with X{, supplies the steady-state current /, when switch S is open and supplies the current to the short circuit through X%, and Z,, when switch S is closed, If we can determine Ef, the current through X7,, will be J”, With switch S open, we sec that Epa, + iXbgh = Ve + (Zou + IXdg) he (03) and this equation defines E%, which is called the subrransienr intemal voltag Similarly, when calculating transient current J’, which mustbe supplied through the transient reactance %,,, the driving voltage is the transient internal voltage E,, where 4? Ea V4 igh, = Vy (Zou t+ Xie)” (10.8) Thus, the value of the load current J, determines the vaiues of the voltages Ey and £%, which are both equal to the no-load voltage E, only whea J, is zero SO that E, is then equal to V, At this point it is important to note that the particular value of Fy in series with X, represents the generator immediately before and immediately after the fault occurs only if the pretault current in the generator has the corresponding value of /,. On the other hand, E, in series with the syn- chronous reactance V4, is the equivalent circuit of the machine under steady- state conditions for any value of the load current. The magnitude of &, is determined by the field current of the machine, and so for a different value of102 INTERNAL VOLTAGES OF LOADED MACHINES UNDER FAULT ConpiTIONs 385 I, in the circuit of Fig. 10.2(a) |£,| would remain the same but a new value of E% would be required Synchronous motors have reactances of the same type as generators. When a motor is short-circuited, it no longer receives electric energy from the power line, but its field remains energized and the inertia of its rotor and connected load keeps it rotating for a short period of time. The internal voltage of a synchronous motor causes it to contribute current to the system, for it is then acting like a generator. By comparison with the corresponding formulas for a generator the subtransient internal voltage E", and transient internal voltage £/, for a synchronous motor are given by = IN ly (10.5) = IX ily (106) where V, is now the terminal voltage of the motor. Fault currents in systems contain ng generators and motors under load may be solved in either one af two ways: (1) by calculating the subtransient (or transient) internal voltages of the machines or (2) by using Thévenin’s theorem. A simple example will illustrate the two approaches : Suppose that a synchronous generator is connected to a synchronous or by a jine of external impedance Z.,,. The motor is drawing load current I, from the generator when a symmetrical three-phase fault occurs at the motor terminals. Figure 10.3 shows the equivalent circuits and current flows of the sysicm immediately before and immediately after the fauit occurs, By replacing the synchronous reactances of the machines by their suitransient reactances as shows in Fig. 10.3(a), we can calculate the subtransient internal voltages of the i Bon a a * G01 Ty IXGe > (0.20) (0.20) RIX 5X4 ES (0.20) XG v, 7 + + ea «5 O “TL_Neutral | Z i. Neutral Before the fault (b) After the fault FIGURE 103 Equivalent cirewits and current Rows before and alter a fault at the terminals of a synchronous mo‘or connected 10 & synchronous generator by line impedance Z.,,. Numeric values are for Excmple 101386 CHAPTER 10 SYMMETRICAL FAULTS and I, in the equations EL = Vi + (Zon + iXde) I (10.7) Em = Vy ~ iXinde. (10.8) When the fault is on the system, as shown in Fig. 10.3(b), the subtransient currents J] out of the generator and". out of the motor are found from the relations I, (40.9) (10.10) These two currents add together to give the total symmetrical fault current /7 shown in Fig. 10.3(6). That is, + le Xm hy ny Westt ls 10,11) zi (10.11) where I"; and I" are the respective contributions of the generator and motor to the fault current /7. Note that the fault current does not include the prefault (load) current. The altemative approach using Thévenin’s theorem is based on the observation that Eq. (10.11) requires a knowledge of ony ¥j, the prefault voltage of the fault point, and the parameters of the network with the subtran- sient reactances representing the machines, Therefore, 17 and the additional currents produced throughout the network by the fault can be found simply by applying voltage V; to the fault point P in the dead subtransient network of the system, as shown in Fig. 10.4(a). If we redraw that newyork as shown in Fig. 10.4(b), it becomes clear that the symmetrical values of the subtransient fault currents can be found from the Thévenin equivalent circuit of the subtransient network at the fault point. The Thévenin equivalent circuit is.a single generator and a single impedance terminating at the point of application of the fault. The equivalent generator has an internal voltage equal to V,, the voltage at the fault point before the fault occurs. The impedance is that measured at the point of application of the fault looking back into the circuit with all the generated voltages short-circuited. Subtransient reactances are used since the initial sym102, INTERNAL VOLTAGES OF LOADED MACHINESUNDERFAULTCONDITIONS 387 | FIGURE 10.4 Chrewis illustrating the wetionsl current flows due #0 the thre-phase fsl: at P: (a)applying ¥; t0 dead netwark to simelite the fault, (b) Thévenin equivelent looking into the cizcuit at point P. | (@) ) 104(b) the Thevenin impedance Zy, is LS CT (10.12) OU Lew ti Xig + Xam) Upon the occurrence of a three-phase short circuit at P. sinvs nS, the subtransient current in the fault is lated by closing y, [Zou HM + Xa) Zn Ron Zee + IX ie (10.13) Thus, three-phase symmetrical faults on systems containing generators and motors under load may be analyzed by either the use of subtransient internal voltages or Thévenin's theorem, as illustrated in the following examples. Example 1.1. Asynchionous generator and motor are rated 30,000 kVA, 13.2 kV, and both have subtransient revetances of 20%. The line connecting them has a reactanee of 104 on the base of the machine ratings. The motor is drawing 20,000 KW at 0 ter leading and a seminal voltage of 12.8 kV when a symmetical three-phase fault eceury at the motor tarminuls, Pind the spbtransient currents ip Ue genersitor. the motor, andl the faut by using tae internal vollages of tie macinnes, Solution. The » Choosing a t asthe referenc efaull equivalent circuit of the system corresponds to Fig. 10.3(a). of 30,000 kVA, 13.2 kV and using the voltage M) at the fault point pheso:, we abtain 12.8 “as oof ° per unit388 CHAPTER 10 SYMMETRICAL FAULTS Base current 1128/36." 1312 For the generator 0.970 + j0.1(0.69 + 0.52) 0.918 + j0.069 4 j0.2(0.69 + 0 O81 + j0.207 j03 Fu “ Ve. °/ per unit =09N + 169 + j9.52) = 1.074 = j0.1381, 74 138 LN DSS = 069 = 1312(-0.0 In the fault Tp [+ Ih = 069 = 7271 ~ 0.69 = -j8.08 x 1312 Figure 10.3(6) shows the patiis of / 918 + ~ j5.37) = -905 — jiu. ~-/10,600 A. 36.7 A 86/ 36.9 per unit S6(0.8 + j0.6) = 0.69 + jO.S2peranit 369 per unit U.814 + J0.207 per unit = 0.69 = 72.71 per unit 0.69 = j2.71) = 905 — j3550 A 70 ~ j0.138 + 0.104 per unit - unit J9.37 = ~ JBL. and 110.2 INVERNAL VOLTAGES OF LOADED MACHINES UNDER FAULT CONDITIONS 389 Example 10.2. Solve Exampte 10.1 by the use of Thévenin’s theorem Solution. The Thévenin equivalent circuit corresponds to Fig. 10.4 In the fault —/8.08 per unit This foult current is divicied between the parallel circuits of the machines inversely ances, By simple current division we obtain the fault currents ~ J02 From generator = + j323perunit From mor. 4.85 per unit Neglecting load current gives Fault current from generator = 3.23 x 1312 = 4240 - 485x131 Fault current from moti 6360 A, Curent in fault = 8.08 x 1312 = 10,600 A The current in the fault is the same whether or not joad current is considered, but the cu in the lines differ, When load current /, is included, we find from Example 19.4 that J3.23 + 0.09 + j0.52 = 0.69 perunil Moe hy 74.85 ~ 0.69 = j0.52= =0.69 + 5.37 per unit Note that f,, isin the same direction as J but opposite to £. The per-unit values found for Jj, If, and #!, are the same 8 in Fxample 10.7, and so the ampere values will aso be the same Fault current from generato: CO0A ult current from motor = 72008 7 of the magnitudes of the gencrator and motor currents does not equal the foull current because the currents from the generator and motor are not in phase when lod current is included.390 CHAPTER 10 SYMMETRICAL FAULTS Usually, load current is omitted in determining the current in each line upon occurrence of @ fault. In the Thévenin method neglect of load current means that the prefault current in each line is not added to the component of current flowing toward the fault in the line. The method of Example 10.1 neglects load current if the subtransient internal voltages of all machines are assumed equal to the voltage V; at the fault before the fault occurs, for such is the case if no current flows anywhere in the network prior to the fault Resistances, charging capacitances, and off-nominal tap-changing of transform- ers ar. also usually omitted in fault studies since they are not likely to influence the fevel of fault current significantly. Calculation of the fault currents is thereby simplified since the network model becomes an interconnection of inductive reactances and all currents throughout the faulted systent are then in phase, as demonstrated in Example 10.2 10.3. FAULT CALCULATIONS USING 2. Our discussion of fault calculations has been confined to simple civcuits, but now we extend our study to general networks. We proceed to the general equations by starting with a specific network with which we are already familiar tus In the circuit of Fig. 7.4 if the reactances in series with the generated voltages are changed from synchronous to subtransient values, and if the generated voltages become subtransient internal voltages, we have the network shown in Fig. 10.5. This network can be regarded as the “perphase equivalent of a balanced three-phase system. If we choose to study a fault at bus @), for example, we can follow the notation of Sec. 10.2 and designate Vas the actual vdtage at bus (2) before the fault occurs A 0.10 j0.10 Vs 4 FIGURE 10.5 J a Reactance diagram obtained from Fig. 74 by substituting sut ent 4020 70.20 value! synchronous reactances + f and synchronous internal voltages Et Et | of the machines. Reactanee values és - are marked in per unit103 FAULT CALCULATIONS USING Zru, 391 [8800 it of Figure 105 with a phase fault on bus @ Simulitedt by Fy and =¥y in COI A three-phase fault at bus is simulated by the network of Fig. 10.6, where the source voltages ¥ and -V, in series constitute a short-circuit branch, Source voltage I, acting alone in this branch would match the prefault voltage already at bus (2), and therefore would not cause current to flow in the ranch, With V, and ~V, in seties, the branch becomes a short circuit and the branch current is {7 as shown, It is evident, therefore, that J is caused by the addition of the -V% source. The current /f distributes itself throughout the sysiem from the reference node before flowing owt of bus (2) through the —V, source. In doing so, it produces whatever bus voltage changes that occur in the system due tothe fault, If E”, £2, and V; are short-circuited, then =Vy is left to act alone and =f" into bus (2) is then the only current entering the network from external sources, With a8 the only source, the network has the nodal impedance equations in the Z),,, mettix form ® ® @ AY] ay] Ay Zyxy & v av4,|=]-%] = Ly ta Z|) 777 (10.14) ay AVY, Ly Ly Zul] 0 al, My Zy Za 2 0 in the voltages at the buses due ta the cutrent =1" injected into 5us @) by the fault The Zs, dwiding algorithm, or some other means such as Yj. ttiangular ization and inversion, Can be used to evaluate the bus impedance matrix for the392 CHAPTER IO SYMMETRICAL FAULTS network of Fig, 10.6. The numerical values of the elements of the matrix will be different from those in Example 7.6 because subtransient reactances are now being used for the synchronous machines. The changes in the bus voitages due If are given by AV, AV, [ -2y¢7 ay)" | av, Noa | | eal om yy AY -Za1 The second row of this equation shows that (10.16) le senting the Thévenin ). Substituting the expression for £7" into We recognize Z,, as the diagonal element of Zu, TEP! impedance of the network at bus Eq. (10.15) gives Zn AV, “Sy, Zn]. , ayy hy, aie (0.27) 32 2 -S 1, aK zal 2a2 1 -Sy, [Any [OZ When the generator vollage —V, is short-circuited in the network of Fig. 10.6 and the sources E%, £/', and Vy are reinserted into the network, the currents and voltages everywhere in the network will be the same as those existing before the fault. By the principle of superposition these prefault voltages add to the changes given by Eq, (10.17) to yield the total voltages existing after the fault occurs ‘The faulted network is usually, but not always, assumed to be without load before the fault occurs. In the absence of loads, as remarked previously, no prefault currents flow and there are no voltage difierences across the branch impedances; all bus voltages throughout the network are then the same as V;, the prefault voltage at the fault point. The assumption of no prefault current simplifies our work considerably, and by applying the principle of superposition,s 13 PAULTCALCULATIONS USING Ze we obtain the bus voltages MY VY, AV, . + = (10.18) V; vy, AV, Vy, | yy AY, Z L L Job L 22 ‘Thus, the voltages at aif buses of the network can be calculated using the prefault voltage V7, of the fault bus and the elements in the column of Zaye corresponding to the fault bus. The calculated values of the bus voltages will yie.d the subtransient currents in the branches of the network if the system Z. has heen formed with subtiansient values for the machine ances, In more general terms, when the three-phase fault o ‘on bus (k) of a le newwork, we nave la V, Mat (10.19) Zee and neglecting prefault load currents, we can then write for the voltage at any bus ) during the fault wmv vy, — Vy (10.20) where Z, and Z,, are elements in column & of the system Z,,,,- If the prefault voltage ofbus () is not the same as the prefault voltage of fault bus @), then we simply replace V7, on tre felt in Eq, (10.20) by the actus. prefault voltage of bus GZ). Knowing the bus voltages during the fault, we can calculate the subtransient current [ from bus @) to bus (2) in the ine of impedance Z, connecting those two buses, "% This equation shows J’, as the fraction of the fault current 7 appear line flow from bus @) to bus GZ) in the faulted network If bus () is directly connected to the faulted bus (&) by a line of series impedance Z,, then the current contributed from bus @) to the current in the fault at bus () is simply V,/Z,,, where V, is given by Eq. (10.20).3940 cHapTER 10 SYMMETRICAL FAULTS The discussion of this section shows that only column k of Z,,s, which we now denote by Z*, is required to evaluate the impact on the system of a symmetrical three-phase fault at bus ©). If necessary, the elements of 2%? can be generated from the triangular factors of Y,,,,, aS demonstrated in Sec. 8.5 Example 10.3. A three-phase fault accur at bus @) of the network of Fig, 10.5 Determine the initial symmetrical rmgcurzent (hati the subtran ient current) in the fault; the voltage at buses CQ), @), and (4) during the fault; the current flow in the line from bus ® to bus (); and the current contributions to the fault from, lines G)-@), O-@ }. Take the pretiult vole Yat bus (2) equal to 1.0/4? per unit and negleet all preluult curcuts Solution. Applying the 2. building algorithm to Fig. 10.5,we finel that oO ® ® : GDI i0.2436 70.1938 jOs44 —j0.1256 QI j0.1938 702205 jO.1494 70.1506 Bus = QJ s0.1s45 f0.1494 jO1954 70.1056 @® [0.1456 0.1506 j0.1046 j0.1954 | Since load currents are neglected, the prefault voltage at eacn bus 1s 1.0/0° per unit, the same as Vj at bus @). When the fault occurs, et 1 isc, i ~ 7 Faas 7 MASTS per unit and from Eq, (10.18) the voltages duving the fault arc VY, a D158 : ~ F0295 018% V; 0 0 y JO.1d94 | ~ ous per unit - 3490 vs j0.2295 | - 40-1506 - 0.343 a yous | |° 8 The current flow in line @)- (1) is Vs - 0.3490 — 0.1556 = ~j0.7736 per unit104 FAULT CALCULATIONS USING2,, EQUIVALENTCIRCUITS 395 Fault currents contributed to bus (2) by the adjacent unfaulted buses are 0.1556, G01 u From bus () =)1.2448 per unit Vv; 0.3490 40.25 3960 per unit 03438 70.20 From bu 7190 per unit Fecept for round-olf cn ors, the sum of these current contributions equals 17. 10.4 FAULT CALCULATIONS USING Z. EQUIVALENT.CIRCUITS We cannot devise a physically realizable network which directly incorporates all the individual elements of the bus impedance matrix. However, Fig. 84 shows that we can use the matrix elements to construct the Thévenin equivalent circuit between any pair of buses in the network that may be of interest. The Thévenin equivalent circuit is very helpful for illustrating the symmetrical fault equations, which are developed in Sec. 10.3 _ In the Thévenin equivalent circuit of Fig. 10.%a) bus ©) is assumed to be the fault bus and bus G@) s unfaulted. The impedances shown correspond directly to the elements af the network Zyy, and all the prefault bus voltages are the same as Vj of the fault bus if load currents are neglected. The two points Urfaulted bus Urfaulted bus ne a = Sy, -2, 2 FIGURE 10.7 Phévenin equivalent between duses (1) and @) of system with ro piefault load currents: (a) before the (aul: (8 openk (b) ducing the fall (S closed) ‘396 CHAPTER 10 SYMMETRICAL FAULTS marked x have the same potential, and so they can be joined together to yield the equivalent circuit of Fig. 10.7() with a single voltage source V, as shown. If theswitch S is open between bus (©) anc the reference node, there is no short circuit and no current flows in any branch of the network. When S is closed to represent the fault on bus ), current flows in the circuit toward bus &). This current is [7 = ¥%//2,,, which agrees with Eq. (10.19), and it induces a voltage drop (24/24); in the diteetion from the relerence node toward bus @). The voltage at bus @) to the reference changes therefore by the émount —(Zip/Zy4)V, so that the voltage at bus G) during the fault is Vp- Zig] Zaps Which is consistent with Eq. (10.20). ‘Thus, by substituting appropriate numerical values for the impedances in the simple equivalent circuit of Fig. 10.7(6), we can calculate the dus voltages of the system before and after the fault occurs. With switch § open in the circxit, the voltages at bus ) and the representative bus (7) are equal to ¥,, The same uniform voltage profile occurs in Fig. 10.6 if there are no prefault currents so that E% and Ef equal ¥,, If S is closed in Fig. 10.7(b), the circuit refiects the voltage of representative bus (J) with respect to reference while the fault ison bus Therefore, if a three-phase short-circuit fault occurs at bus (&) of a large-scale network, we can calculate the current in the fault and the voltage at any of the unfaulted buses simply by inserting the proper impedance vaiues into elementary circuits like those in Fig. 10.7. The following example illustrates the procedure. Example 10.4. A five-bus network has gencratomt at buses (i) and (3) rated 270 and 225 MVA, respectively. The generator subtransient reactanc plus the reactances of the transformers connecting them to the buses are cack 0.30 per unit on the generator rating as base. The tums ratios of the transformer are such that the voltage base in each generator circuit is equal to the voliage rating of the generator. Line impedances in per unit on a 100-MVA system base are shown in Fig. 10.8. All resistances are neglected, Using the bus impedance nuatrix for the network which includes the generalor and transformer reactances, find the FIGURE 10.8 Impedance diagram for Example 10.4. Generator reactanees include subtcansient values plus Teactances of set-up transformess, All values in pes unit on 100-MVA base104 FAULTCALCULATIONSUSINGZy, FOUVALENTEIRCUITS 397 subtransient current in a three-phase fault at bus (4) and the current coming to the faulted bus over each line. Prefault current is to be neglected and all voltages asumec to be 1.0 per unit before the fault occurs. Solution. Converted to the 1((-MVA base, the combined generator and transformer reactances are ~ 100 . Generator at bus D): ¥ = 0.30 x <= = 0.11 Iperunit x=oa0% y} 0.0 x 55 Generator at bus = 0.1333 per unit 31S These values, along with the line impedances, are marked in per unit in Fig, 10.8 from which the bus impedance matrix can be determined by the Zp, building algoritim to yield @ @ ad oO 30.0793 70.0558 70.0382 j0.0511 j0.0608 J0.05S8 50.0664 0.0630 0.0605 j00382 j0.0875 —j0.0720 j0.0603 4) | sO.051L j0.0630 70.0720 70.2321 0.1002 G [0.0608 70.0605 jo.0503 70.1002 j0.1301 Since we are to calculate the currents from buses (3) and () into the fault at bus Unfaulted bus Unfaulted bus 4 + = 10/0" Zag 2 foe 2y = Ze Za Ze >s s ae 0.0720 © j0.1601 Ze =2s,| | 30.1002 j0.1319 17 = ~ 4.308 1; = =j4.308 fa) (6) FIGURE. 10.9 . Use of Thévenin equivalent circuits to calculate voltages at (a) bus (9) and (b) bus (@) due to fault o :398 CHAPTER 10. SYMMETRICAL FAULTS. we need to know V3 and Vs during the fault, Visualizing equivalent circuits like those of Fig. 10.9 helps in finding the desited currents and voltages. ‘The subtransient currentin the three-phase fault at bus (4) can be calculated from Fig. 10. a), Simply closing switch $ gives = —/4.308 per unit From Fig. 10.%a) the voltage at bus @) during the fault is Vy = Vy ~ If 255 = 1.0 = (+4.308)( 10.0720) = 0.68 Sper unit From Fig. 10.9(b) the voltage at bus (5) curing the fault is (WZ, = 1.0 4.308 )( j0.1002) = 0.5683 per unit Currents into the fau't at bus (4) over the line impedances Z, are -j2.053 per unit per unit Hence, total fault current at bus (4) — j4.308 per unit Other equivalent circuits based on the given bus impedance matrix can be developed for three-phase faults on any of the other buses or transmission fines of the system. A specific application will demonstrate how this is accomplished ‘Three-phase faults occur more often on transmission lines than on substa- | tion buses because of the greater expesure of the lines to storms an¢ accidental disturbances. To analyze a fine fault, the point of fault on the line can be assigned a new bus number and Z,,, for the normal configuration of the network can then be modified to accommodate the new bus. Sometimes the circuit breakers at the two ends of the line do not open simultaneously when a line fault is being cleared. Ifonly one circuit breaker has opened and the fault is not fully cleared, short-circuit current persists. The so-called fine-end fault represents the partiaular situation where the three-phase fault occurs very close to one of the teminating buses of the line, on the line side of the first breaker (near the fault) to open. The line breaker near the fault is called the nearend breaker and that at the enc away from the fault is called the remore-end breaker. ‘The single-line diagram of Fig. 10.10 shows a fourbus network with a line-end fault at point P on the line connecting buses (1) and (2). The line has series impedance Z,. The near-end breaker at bus (2) is open and the remote-end breaker is closed, leaving the fault still on at point , which we now104 FAULT CALCULATIONS USING Znys EQUIVALENTCIRCUTS 399. FIGURE 19.10 Line-end! fix Fig. 108 2) of system of poin: Pon line of series impedance Z, between boses CD) and In order to study this fault condition, we need to modify the existing bus impedance matrix Z,.;. for the normal configuration of the system to reflect the near-end breaker operation. This is accomplished in twa steps: 1, Establish the new bus © by adding a line of series impedance Z,, between bus () and bus @ 2. Remove the line between bus (2) and bus @) by adding line impedance ~Z,, between those two buses in the manner explained in Sec. 8.4 The first step follows the procedure given for Case 2 in Table 8.4 and yields, in terms of the elements Z,; of Zoips the first five rows and columns of the symmetric matrix @ @ @ zy 2722 2 2a-2n zm Mn 2 7 a Za 7 Za 2s Zur 2 | 2n-2e Bi ~ Za) [Zn Zu) | Zur Zo (10.22) where Zyy12 = Zi; + Zz — 2Zip when Zorg is symmetric. The second step ean be accomplished by forming row @) and column @) es shown and then Kron400 CHAPTER 10 SYMMETRICAL FAULTS reducing the matrix Z to obtain the new 5 XS matrix Zi new including bus ®), as explained for Case 4 of Table 8.1. However, since Zeg,neu iS the only clement required to calculate the current in the fault at bus ©) (point P of Fig. 10.10), we can save work by observing from Eq. (10.22) that the Kron reduction form gives (10.23) Again, we note that Z,) = Zp) and Zy,2 = Zn + Zag — 2Zya. By neglecting prefault currents and assigning prefault voltage 1’, = 1.0/0" per unit to the fault point P, we find the line-end fault current #7 out of be 1.0 10 = SC) [ee Zarvew Zi + (Zu = ZaY HZ 26) f Thus, the only elements of Zocg entering into the calculation of If are Zy, Zyy = Zp, and Zyp. It 8 worthwhile observing that the same equation for the line-end fault current can be found directly by inspection of Fig. 10.11(a), which shows the Thévenin equivalent circuit between buses (1) and (2) of the prefault network The impedances Z, and —Z,, are connected as shown in accordance with steps 1and2 above. Circuit analysis then shows in a straightforward manner that the impedance looking back into the circuit from the terminals of the open switch Sis (21) = 2n)(Zr = Zekonew = Zp + pe en tanee OY Ty Tig # Bay = Fas = (10.28) Since Zz = Za and Zy12 = Za + Zap ~ 2Zy2y Eq. (10.25) can be reduced to give (2n > Za) (Zn > 25) — Zun> Za Z, Z¢-— + +Z akanew ™ Zp Zia hh 2 (Zn- =Zy,+Z,- (10.26) " Zia Thus, simply by closing switch § as shown in Fig. 10.11(b) and using elementary circuit analysis, we can calculate the line-end fault current /> in agreement with Eq, (10.24). Of course, the circuit approach using the Thévenin equivalent must yield the same results as the matrix manipulations of Eq. (10.22)—for the sameN04 FAULT CALCULATIONS USING2,, FOUIVALENTCIRCUTTS 401 FIGURE 10.11 Simulating the ling-ead fault Of Fig. 10.10 by Thérenin equivalent cizevit: (a) with line open before the foull; (b) during the fault (S closed) external connections are being made to the la stem model as to its Thévenin equivalent, Other uses of the equivalent circuits based on the bus impedance matrix are possible Example 10.5. In the five-bus system of Fig. 10.8 a line-end, short-circuit fault occurs on iine @)~ @), on the line side of the breater at bus (2). Neglecting prefaull currents and assuming rated system voltage at the fault point, ealeulate the subzansient current into the fault when only the neer-end breaker at bus 2) opens Solution, Figure 10.8 shows that the impedance of fine D)- @) is Z,, = j0.168 per unit and the required elements of Z;,,, are given in Exemple 10.4. The Thévenie equivalent circuit looking into the intact system between buses -G) and @)402 cHaPTER10 SYMMETRICAL FAULTS corresponds to Fig. 10.114) The numerical values of the impedances shown in parallei are calculated as follows: 2; = j0.0793 ~ j0.0558 = j0.0235 Zq ~ Zp = j01338 — j0.0558 ~ j0.168 = ~ 70.09 The new Thévenin impedance seen looking back into the faulied system between the fault point P and the reference is therefore given by Ec, (10.25) as (40,0235)( =j9.09) Zee sou = 0.168 ets oow “JOBS “C535 TOD) + 10.0558 = 30.2556 per anit ‘Thus, the subtinsient current inte the linc-end fill is —j3.912 perunit 10.5 THE SELECTION OF CIRCUIT BREAKERS ‘The electric utility company furnishes data to a customer who must determine the fault current in order to specify circuit breakers properly for an industrial plant or industrial power distribution system connected to the utility system ata certain point. Instead of providing the Thévenin impedance of the system at the point of connection, usually the power company informs the customer of the short-circuit megavoltamperes which can be expected at nominal! voltage; that is, Short-cireuit MVA = v3 x (nominal kV) X Mycl «1074 (10.27) where [fgcl in amperes is the rms magnitude of the short-circuit current in a three-phase fault at the connection point. Base megavoltamperes are related to base kilovolts and base amperes [yasel by Base MVA = ¥3 X (base kV) X Vinsel X 107%) (10.28) If base kilovolts equal nominal! kilovolts, then dividing Eq. (1027) by Eq. (10.28) converts the former to per unit, and we obtain Short-circuit MVA in per unit = [gel in per unit (10.29) Atnominal voltage the Thévenin equivalent cizcuit looking back into the system from the point of connection is an emf of 1.0/0” per unit in series with the105. THE SELECTIONOF CIRCUIT BREAKERS 403 per-unit impedance Z,,. Therefore, under short-circuit conditions, 10 10 Ty Pet unit [Zul = ee "eel short-circuit MV perunit (10.30) Often resistance anc shunt capacitance are neglected, in which case Zy, = Xp. Thus, by specifying short-circuit megavoltamperes at the customer's bus, the electric utility is effectively describing the short-circuit current at nominal voltage and the reciprocal of the Thévenin impedance of the system at the point of connection Much study has been given to circui:-breaker ratings and applications, and our discussion here gives some intraduetion to the subject. The presentation is not intended as a study of breaker applicauons but, rather, 10 indicate the portance of understanding fault calculations, For additional guidance in specifying breakers the reader should consult the ANSI publications listed in the footnotes which accompuny this section. From the cwrene viewpoint two factors to be considered in selecting circuit breaker are: + The maximum instantaneous current which the breaker mu and * The totai current when the breaker contacts part to interrupt the circuits carry (withstand) Up to this point we have devoted most of our attention to the subtransient current called the initial symmetrical current, which does not include the de component. Inclusion of the de component results in a rms vale of current immediately after the fault, which is higher than the subtransient current. For o'l circuit breakers above 5 kV the subtransient current multiplied by 1.6 is considered to be the rms value of the current whose disruptive forces the breaker must withstand during the first half cycle after the fault occurs. This cursent is called the momentary current, and for many years circuit breakers were rated in terms of their momentary current as well as other criteria.’ The inentpting rams of a circuit breaker was specified in kilovoltam- peres or megavoltemperes. The interrupting kilovoltamperes equal v3. x (the kilovolts of the bus to which the breaker is connected) x (the current which the breaker must be cxpable of interrupting when its conticts part), This interrupt ing current is, of course, lower than the momentary current and depends on the speed of the breaker, such as 8, 5, 3, or 2 cycles, which is a measure of the time from the occurrence of the fault to the extinction of the arc, Breakers of dificrent speeds are classified by their rated interrupting times. The rated 'See G. N. Lester, “High Voltage Circuit Breaker Standards in the USA: Pait, Present, and Future,” IEEE Transactions on Power Apparatus and Systems, vol. 93, 1974, pp. 590-600,404) CHAPTER 10 SYMMETRICAL FAULTS Extinctionof . arcon Initiation of primary contacts short circuit —— Energization oftrip circuit Parting of primary arcing contacts a Time] Interrupting time Tripping] Opening] Arcing delay |_ time | _ time FIGURE 10.12 J+] Definition of interrupting time given in Contact ANSI/IEEE Standard C57.010-1979, Applica parting time tion Gusle for AC High Voltage Circuit Breakers sd on a Symmetrical Bass interrupting time of a circuit breaker is the period between the instant of energizing the trip circuit and the arc extinction on an opening operation, Fig 10.12. Preceding this period is the tripping delay time, which is usually assumed tobe 2 cycle for relays to pick up. ‘The current which a breaker must interrupt is usually asymmetrical since it still contains some of the decaying de component. A schedule of preferred ratings for ac high-voltage oil circuit breakers specifies the interrupting current ratings of breakers in terms of the component of the asymmetrical current which is symmetrical about the zero axis. This current is properly called the required symmetrical interrupting capability or simply the rated symmetrical short-circuit current. Often the adjective symmetrical is omitted. Selection of circuit breakers may also be made on the basis of total current (de component included)? We shall limit our discussion to a brief treatment of the symmetrical basis of breaker selection See Preferred Ratings and Related Required Capabilities for AC High-Voltage Circuit Breakers Rated ona Syrimetrical Current Basts, ANS] C37.06-1987, and Guide for Calculation of Fault Currents for Application of AC High-Voltage Circuit Breakers Rated on a Total Current Basis, ANSI C315-1979, ‘American National StandaaJs Institute, New York.WS THE SELECTION or circurTBREAKERS 405 Breakers are identified by nominal-voltage class, such as 69 kV. Among other factors specified are rated continuous current, rated maximum voltage, voltage range factor K, and rated short-circuit current at rated maximum kilovolts. The rated maximum voltage of a circuit breaker is the highest rms voltage for which the circuit breaker is designed. The rated voltage range factor K is the ratio (rated maximum voltage = the lower limit of the range of operating voltage). K determines the range of voltage over which the product (rated short-circuit current X operating voltage) is constant. In the application of circuit breakers it is important not to exceed the short-circuit capabilities of the breakers. A breaker is required to have a maximum synunetrical interrupting capability equal to K x rated short-circuit current. Between the rated maximum voltage and 1/K times the rated maximum voltage the symmetrical interrupting capability is defined as the product {rated short-circuit current x (rated maxi- mum voltage /operating voltage)]. Example 10.6. A 69-kV circuit breaker having a voltage range factor K of 1.21 and a continuous current rating of 1200 A has a rated short-circuit current of 19,000 A. at the maximum rated voltage of 72. kV, Determine the maximum symmetica! mterrupling capability of the breaker and explain its significance at lower operating voltages. Solution, The maximum. symmetrical interrupting capability is gwen by K & rated short-circuit current = 1.21 = 19,000 = 22,990 A This value of symmetrical interrupting current must not be exceeded, From the definition of K we have rated maximum voltage Lower limit of operating voltage = K = Hence, in the operating voltage range 72.5-60 kV, the symmetrical interrupting curreat may exceed the rated short-circuit current of 19,000 A, but it is limited to 22,990 A. For example, at 66 kV the interrupting current can be RS 19,000 = 20,871 A. Breakers of the 115 class and higher have a K of 1.0. A simplified procedure for calculating the symmetrical short-circuit cur-_ rent, called the E/X method disregards all resistance, all static load, and all ‘See Application Guide for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis, ANS] C37.010-1979, American Notional Standards Institute, New York, This publication is also TEEE Std. 320-1979 .406 cuarreR10 SYMMETRICAL FAULTS prefault current. Subtransient reactance is used for generators in the E/X method, and for synchronous motors the recommended reactance is the X of the motor times 15, which is the approximate value of the transient reactance X¥4 of the motor. Induction motors below 50 hp are neglected, and various multiplying factors are applied to the X% of lamer induction motors depending on their size. If no motors are present, Symmetrical short-circuit current equals subtransient current. The impedance by which the voltage V, at the fault is divided to find short-circuit current must be examined when the E/X methed is used. In specifying a breaker for bus ®, this impedance is Z,, of the bus impedance matrix with the proper machine reactances since the short-circuit current is expressed by Eq. (10.19). If the radio of X/R of this impedance is 15 or less, a breaker of the correct voltage and kilovoltamperes may be used if its interrupt- ing current rating is equal to or exceeds the calculated cument. Ifthe X/R ratio is unknown, the calculated current should be no more than 80% of the allowed value for the breaker at the existing bus voltage. The ANSI application guide specifies ’@ corrected method to account for ac and de time constants for the decay of the current amplitude if the Y/R ratio exceeds 15. The corrected method also considers breaker speed. Example 10.7. A 25,000KVA 138-kV generator with Xj = 15% is connected through a transformer toa bus which supplie four identical motors, as shown in Fig. 10.13. The subtransient reactance Xf of each motor is 20%on a base of 5000 kVA, 6.9 kV. The three-phase rating of the transformer is 25,000 «VA, 138/6.9 KV, with a leakage reactance of 10%. The bus voltage at the motors is 6.9kV when a three-phase fault occur at point . For the fault specified, determine (a) the subtransient current in the fault, (b) the subtransient current in breaxer A, and (c) the symmetrical short-circuit interrupting current (as defined for circuil-breaker applications) in the fault and in breaker A Solution. (a) For a base of 25,000 kVA, 13.8 kV in the generator circuit the base for the motort is 25,000 kVA, 6.9kV. The subtransient reactance of each motor is 9.20 25:000 ari) 0 per unit Figure 10.14 is the diagram with subtransient values of reactance marked, For a Motors Gen, CO) FIGURE 10.13 a ap One-line diagrain for Example 10710S THESELECTIONOF cIRcUITBREAKERS 407 FIGURE 10.14 Reactance diagram for Example 10.7 fault at P. V,= 10/0" perunit — 2,, = 70.125 per unit /_ 10/0 . I = Sige = IR nor nic ‘The base currentin the 69 kV circuit is and so pls (5) Through breaker A comes the contabution from the generator and three of the four motors. The generator contributes a current of 0.25 050 ~j8.0x <= = —j4.0per unit Each motor contributes 25% of the remaining fault current, or —j 1.0 per-unit amperes cach. breaker A 14,6304, =j4.0 + 3(—J1.0) = -j7.0perunit or (©) To compute the current to he intemupted by breaker subtransient reactanee of j1.0 by the transient reactance of jl cireuits of Fig. 10.14. Then, replace the in the motor The generator contributes a current of 10 _ 0375 —— x js 4.0 per unit408) ciarTER 10 SYMMETRICAL FAULTS Each motor contributes a current of 1. 10 © 025 4 * 075 * 00s ~ —j0.67 per unit The symmetrical short-circuit current to be interrupted is (4.0 4 3 x 0.67) x 2090 = 12,560 Suppose that all the breakers connected to the bus are rated on the basis of the current into a fault on the bus. In that case the short-circuit current interrupting rating of the breakers connected to the 6.9kV bus must he at least 444x067 =6.67perunit or 6.67 x 2090 = 13,940 A AI44-KV circuit breaker has a rated maximum voltage of 15.5 kV and a K of 2.67, At 155 kV its rated short-circuit interrupting current is 8900 A. This breaker is rated for a symmetrical short-circuit interrupting current of 2.67 x 8900 = 23,760 A, at a voltage of 15.5/2.67 = 5.8 kV. This current is the maximum that canbe interrupted even though the breaker may be in a circu't of lower vollage The short-circuit interrupting current rating at 6.9 kV is, 15.5 = 0 = 20,000 65 x 890 20,000 A The required capability of 13,940 A is well below 80% of 20,000 A, and the breaker is suitable with respect to short-circuit current The short-ciruit current could have heen found by using the bus impedance matrix. For this purpose two buses (J) and (2) are identified in Fig 10.14. Bus (1) is on the low-voltage side of the transformer and bus (2) is on the high-woliage side. For motor reactance of 1.5 per unit Yy= -j10+ = —/12.67 jsf Ya —J10 = j6,.67 = -j16.67 The node admittance matrix and its inverse are ® ® ® () @f i267 j10.00 _ @[j0.150 j0.090 | Yous = @| jie00 —/16.67 ws Q)i0.090 j0.114105 THE SELECTIONOFCIRcUITBREAKERS 409 0.15 j0.09 FIGURE 10.15 Bus impedance equivelent cizeuit for the 2), of Fas. 19.14 Figure 10.15 is the network corresponding to Zyqy and Vy S, with S; open represents a fault on bus G). The symmetrical short-circuit interrupting current in a three-phase fault at bus () is 1.0 per unit. Closing 10 1s —j6.67 per unit JO. which agrees with our previous calculations. The bus impedance matrix also gives us the voltage at bus (2) with the fault on bus (1). V, = 10> IgeZp = 1.0 — (~j6.67)( j0.09) = 04 and since the admittance between buses (1) and @ the fault from the transformer is 1s ~/10, the current into (0.4 - 0.0)(- f10) = ~j4.0 per unit which also agrees with our previous result We also know immediately the short-circuit current in a three-phase fault at bus @), which, by referring to Fig. 10.15 with S, open and S; closed, is 1.0 en Joana = ~/8-77perunit This simple example illustrates the value of the bus impedance matrix where the410° cHaFTER 10 SYMMETRICAL FAULTS effects of a fault at a number of buses are to be studied. Matrix inversion is not necessary, for Zs can be generated directly by computer using the Los building algorithm of Sec. 8.4 or the triangular factors of Y,.,. aS explained in Sec. 8.5. Example 10.8. The gencrators at bu es @) and @) of the network of Fig. 10.16(a) have ynchronous reactanees X= Xy =J1.70 per unit (as marked) and subtransient reactances X%, 25 per unit. If a three-phase shorcircuit fault occurs at bus (3) when there is no load (all bus voltages equal 1.0/0° per unit), find the initial symmetrical (subtransient) cu ent (a) in the fault (Yan line =), and (c) the voltage at bus 2). Use triangular factors of Y,, in the calculations Solution. For the given fault covelitions the network has the subtransient reactance diagram shown in Fig, 10,1604), and the corresponding Yj, has the ular factors 510 Yous = Ao -s79 bp pS - L U Since the faultis at bus (3), Eqs. (10.19) through (10.21) show that the calculations involve column 3, Z@,, of the subtransient Z,,, , which we now generate as follows: —710 » \fey) fo Jl 79 0 JS j3S - 3.94937 1 ® ® cq i05338 10.25 Q + (@) @) FIGURE 10.16 The reactance diagram for Example 108 with generators represe: behind X45 (b) equivalent current source in paralel with X% ed by: (a) series voltage source196 suMMaRy 411 Solving, we obtain 1 x, = ———— = 0.25320 pes 5 = Sypagay 7102530 per unit and so the element of 26) are given by 1-01-05 |f2p 0 1 0.44304 |] Z., | = 0 1 25 0.25320 Weind tsa Za = 301.2532 perunit Zn = #11218 per unit Zy = J0.13782 per unit (i) Acoording 10 Eq. (£0.19), the subtransiont current in the fault is, sp = PRM peru: Im 2 i (6) From Eq. (10.21) we can write for the current in line @) a Vi (25725) re (Zao Zi \eaeos 1 f §0.13782 ~ j0.25320 7 L{ ae 2.27844 per unit } During the fault the voltage at hus Q) is giver: by Eq. (10.20) as follows: ) =05. 5695 per unit 10.6 SUMMARY ‘The current flowing immediately after a fault occurs in a power network is determined by the impedances of the network components and the synchronous machines. The initial symmetrical rms fault current can be determined by representing cach machine by its subdtransient reactance in series with its subtransient internal voltage. Subtransient currents are larger than the transient412° CHAPTER 10 SYMMETRICAL FAULTS. and steady-state currents. Circuit breakers have ratings determined by the maximum instantaneous current which the breaker must withstand and later interrupt. Interrupting currents depend on the speed of breaker operation. Proper selection and application of circuit breakers should follow the recom- mendations of ANSI standards, some of which are referenced in this chapter. Simplifying assumptions usually made in industry-based fault studies are: * All shunt connections from system buses to the reference node (neutral) can be neglected in the equivalent circuits representing transmission lines and transformers. + Load impedances are much larger than those of network comporents, and so they can be neglected in system modeling * All buses of the system have rated/nominal voltage of 0/0" per unt so that no prefault currents flow in the network. + Synchronous machines can be represented by voltage of 1.0/0° per unit behind subtransient or transient reactance, depending or the speed of the circuit breakers and whether the momentary or interrupting fault current is being calculated (ANSI standards should be consulted), + The voltage-source-plusseries-impedance equivalent circuit of each syn- chronous machine can be transformed to an equivalent current-source-pl shunt-impedance model. Then, the shunt impedances of the machine models. represent the only shunt connections to the reference node. The bus impedance matrix is most often used for fault current calcula- tions. The clements of Z,,, can be made available explicitly using the Zeus building algorithm or they can be generated from the triangular factors of Ys. Equivalent circuits based on the elements of Z,, can simplify fault-current calculations as demonstrated in this chapter for the line-end fault PROBLEMS 10.1. A 60-Hz alternating voltage having a rms value of 100 V is applied to a series 2. circuit by closing a switch. The resistance is 15 © and the inductance is 0.12 H. (a) Find the value of the de component of current upon closing the switch if the instantaneou value of the voltage is 50 V at that time: (b) What is the instantaneous vaiue of the voltage which will produce the maximum de component of current upon closing the switeh? (c) What is the instantaneous value of the voltage which will result in the absence of any de component of current uion clesing the switch? +s (d) If the switch is closed when the instantaneous voltage is zero, find the instartancou current 0.5, 1.5, and 5.5 cycles later. 10.2. A generater connectad through a S-cyce circuit breaker toa transformer is rated 100 MVA, 18 KV, with reactances of X{1= 19%, X%, = 26%, and Xg = 130%. It10.3. 10.4. 10.6. 10.7 Prontems 413 is operatingat no load and rated voltage when a three-phase short circuit occurs’ between the breaker and the transformer. Find (a) the sustained short-circuit curzent in the brecker, (b) the initial symmetrical rms current in the breaker, and (c) the maximum possible de component of the short-circuit current in the breaker, ‘The three-phase transformer connected to the generator described in Prob. 10.2is rated 100 MVA, 240Y/184 kV, X = 10%. Tf a three-phase short circuit occurs on the high-voltage side of the transformer at rated voltage and no load, find (a) the initial symmetrical rms cursent in the transformer windings on the high-volt- age side and () the initial symmetrical rms current in the Line on the low-voltage side A 60-Hz generator is rated 500 MVA, 20 KV, with X!, = 0.20 per unit. It supplies ly resistive load of 400 MW at 20 kV. The load is connected directly across, minals of the generator. If all three phases of the load are short-circuited simultaneously. find the initial symmetrical rms current in the generator in per unit on a base of S00 MVA, A generator is connceted through # transformer to a ynehronous motor, Reduced to the same base, the per-unit subtransient reactanees of the generator and motor are 0.15 and 0,35, respectively, and the leakage reactance of the transformer is 0.10 per unis. A three-pha € fault occurs at the terminals of the motor when the terminal voltage of the generator is 0.9 per unit and the output current of the generator is 10 per unit at 0.8 power fictor leading. Find the subtransient current in per unit in the fault, in the generator, and in the motor. Use the terminal voltage of the generator as the reference phasor and obtain the solution (a) by computing the voltages behind subtransient reactance in the generator and motor and (6) by using Thévenin’t theorem, ‘Two synchronous motor having subtansient reactances of 0.80 and0.25 per unit, respectively, on a base of 480 V, 2000 kVA are connected to a bus. This motor is connected by a linc having @ reactance of 0.023 2 10 a bus of a power system. At the power systerr: bus the short-circuit megavoltamperes of the power system are 9.6 MVA for a nominal voltage of 480 V. When the voltage at the motor bus is 440 V, neglect load current and find the initial symmetrical rms current in a three-phase fault at the motor bus. The bus imp nce matrix of a fous-bus network with value in per unit is [i015 j0.08 j0.04 j007 O08 jO1S j0.06 j009 0.04 70.06 j0.13 j0.0S 70.07 j0.09 70,05 jO12 Generator connected to buses (1) and (2) hase their subtransient reactances included in Z,,.. If prefault current is neglected, find the subtransient cument in per unit in tye fault for a thsee-phase fault on bus (4). Assume the voltage at the fault is 10/0 per unit before the fault occurs, Find also the per-unit current from generator 2, whose subtransient reactance is 0.2 per unit4140 CHAPTER 10 SYMMETRICAL FAULTS 10.8. 10.9. 10.10. 10.11. 10.12. 10.13, For the network shown in Fig. 10.17, find the subtransient current in per unit from generator 1 and in line (1)~ (2) and the voltages at buses (1) and (3) for a three-phase fault on bus (2). Asume that no curren} is flowing prior to the fault and that the prefault voltage at bus (2) is 1.0/0° per unit. Use the bus impedance matrixin the calculations. FIGURE 10.17 - Network for Probs 1]8and 109, + For the network shown in Fig. 10,17, determine Yy, and its triangular factors. Use the triangular factors to generate the elements of Zp, needed to salve Pro’. 08. Ifa three-phase fault occurs at bus (1) of the network of Fig. 10.5 when there is no load (al! bus voltages equal 1.0/0° per unit), find the subtransient current in the fault; the voltages at buses ), G), and G); and the current from the generator connected tobus (4). Use equivalent circuits based on Z,,, of Example 10.3 and similar to those of Fig. 10.7 to illustrate your ealculations The network of Fig, 10.8 has the bus impedance matrix given in Example 104. 1a short-circuit crip pe wn @ of the network when there is no load (all bus voltages equal 1.0/0° per unit), find the subtrensient current in the fault, the voltages at buses (2) and G), and the current from the generator connected to bus (1), Use equivalent circuits based on Zp, and similar to those of Fig 10.7 to illustrate your calculations. Zs for the network of Fig. 108 is_given in Example 10.4. If a line-end short-circuit fault occurs on line @)= (S) of the network on the line side of the breaker at bus @) calculate the subtransient current in the fault when only the near-end breaker at bus (3) has opened. Use the equivalent circuit approach of Fig. 10.11 Figure 9.2 shows the one-line diagram of a single power network which has the line data given in Table 9.2. Each generator connected to buses (1) and (4) hasa subtransient reactance of 0.25 per unit. Making the usual fault study assumptions, summarized in Sec. 10.6, determine for the network (a) Yigg (b) Zr (6) the subtransient current in per unit in a three-phase fault on bus @) and (d) the contributions to the fault current from line (1)— 3) and from line @) -G)rrosLeMs 415 10.14. A 625-kV generator with X% = 0.20 per unit is connected to a bus through a circuit breaker, as shown in Fig. 10.18. Connected through circuit breakers to the same bus are three synchronous motor rated 250 hp, 2.4 kV, 1.0 power factor, 90% efficiency, with X") = 0.20 per unit. The motort are operating at full load, unity power factor, and rated voltage, with the load equally divided among the machines (a) Draw the impedance diagram with the impedances marked in per unit on a hase of 625 kVA, 2.4 KV. () Find the symmetrical short-circuit current in amperes, which must be inter= rupted by breaker A and B for a three-phase fault at point P. Simplify the calculations by neglecting the prefault current (c) Repeat pact (b) for a three-phase fault at point Q (4) Repeat part (b) for a three-phase fault at point R a [O-O SFO) FIGURE 1018 ve One-line diagram for Prob. 10.14 10.18. A circuit breaker having a nominal rating of 34.5 kV and @ continuous current rating of 1500 A has a voltage range factor K of 1.65. Rated maximum voltage is 38kV and the rated short-circuit current at that voltage is 22 kA. Find (a) the voltage below which rated short-circuit current does not increase as operating voltage decreases and the value of that current and () rated short-circuit current at 34.5 kV.CHAPTER 11 | SYMMETRICAL COMPONENTS AND SEQUENCE NETWORKS One of the most powerful tools for dealing with unbalanced pdyphase circuits is the method of symmesical components introduced by C. L. Fortescue! Fortescue’s work proves that an unbalanced system of 7 related phasors can be resolved into n systems of balanced phasors called the symwmetrical componenes of the original phasors. The m phasors of cach set of components are equal in length, and the angles between adjacent phasors of the set are equal. Although the method is applicabie to any unbalanced polyphase system, we confine our discussion to three-phase systems. In a three-phase system which is normally balanced, unbalanced fault conditions generally cause unbalanced currents and voltages to exist in cach of the phases. If the currents and voltages are related by constant impedances, the system is said to be lmear and the principle of superposition applies. The voltage response of the linear system to the unbalanced currents can be determined by considering the separate responses of the individual elements to the symmetrical components of the currents. The system elements of interest are the machines, transformers, transmission lines, and loads connected to 4 or Y configurations. 1G. L. Fortescue, “Method of Symmetri | Coordinate Applied to the Solution of Polyphase Networks,” Transactions of vol. 37, 1918, pp. 1027-1140 AIGMLL SYNTHESIS OF UNSYMMETRICAL PHASORS FROM THEIR SYMMETRICALCOMPONENTS 417 In this chapter we study symmetrical components and show that the __ response of each system element depends, in general, on its connections and the componeat of the current being considered, Equivalent circuits, called sequence circuits, will be developed to reflect the separate responses of the elements to each current component. There aze three equivalent circuits for each element of the three-phase system, By organizing the individual equivalent circuits into networks according to the interconnections of the elements, we arrive at the concept of three sazuence networks. Solving the sequence networks for the fault conditions gives symmetrical current and voltage components which can be combined together to reflect the effects of the original unbalanced fault currents on the overall system. Analysis by symmetrical components is a powerful tool which makes the calculation of unsymmetrical faults almost as easy as the calculation of three phase faults. Unsymmetrical faults are studied in Chap. 12. 11,1 SYNTHESIS OF UNSYMMETRICAL PHASORS FROM THEIR SYMMETRICAL COMPONENTS According to Fortescue’s theorem, three unbalanced phasors of a three-phase system can be resolved into three balanced systems of phasors. The balanced sets ‘of components are: 1. Positive sequence components consisting of three phasors equal in magnitude, displaced from each other by 120° in phase, and having the same phase sequence as the original phasors, yativesequence components consisting of three phasors equal in magni- tude, displaced from each other by 120° in phase, and having the phase sequence opposite to that of the original phasors, and 3. Zero-sequence components consisting of three phasors equal in magnitude and with zero phase displacement from each other. It is customary when solving a problem by symmetrical components to designate the three phases of the system as a, &, and ¢ in such a manner that the phase sequence of the voltages and currents in the system is abc. Thus, the phase sequence of the positive-sequence components of the unbalanced phasors is abc, and the phase sequence of the negative-sequence components is ach. If the original phasors are voltages, they may be designated V,, V,, and V,. The three sets of symmetrical components are designated by the additional super- script 1 for the positive-sequence components, 2 for the negative-sequence components, and 0 for the zero-sequence components. Superscripts are chosen So aS not to confuse bus numbers with sequence indicators later on in this chapter. The positive-sequence components of V,, Vj, and V, are VS, VO, and VO, respectively. Similarly, the negative-sequence components are Vi, V,2,418 CHAPTER 11 SYMMETRICAL COMPONENTS AND SEQUENCE NETWORKS vn yo . ve LA yo LE ve a “wo ve vee é Posttve-sequence Negative-sequence —Zero-sequence Components components components FIGURE 11.1 ‘Three sets of balanced phasors which are the symmetrical components of three urb: ed phasors and V@, and the zero-sequence components are V, 1), and V2, respec- tively. Figure 11.1 shows three such sets of symmetrical components, Phasors representing currents will be designated by J with superscripts as for voltages Since each of the original unbalanced phasors is the sum of its compo- nerts, the original phasors expressed in terms of their components are SV 2 VO 42 (a) VO + VEO4 VE (11.2) SVM WV + Ve The synthesis of a set of three unbalanced phasors from the three sects of symmetrical components of Fig. 11.1 is shown in Fig. 11.2. ‘The many advantages of analysis of power systems by the method of symmetrical components will become apparent gradually as we apply the method to the study of unsymmetrical faults on otherwise symmetrical systems. Tt is sufficient to say here that the method consists in finding the symmetrical components of current at the fault. Then, the values of current and voltage at various points in the system can be found by means of the bus impedance matrix. The method is simple and leads to accurate predictions of system behavior, 11.2. THE SYMMETRICAL COMPONENTS OF UNSYMMETRICAL PHASORS In Fig. 11.2 we observe the synthesis of three unsymmetrical phasors from three sets of symmetrical phasors. The synthesis is made in accordance with Bas,11.2 THESYMMETRICAL COMPONENTS OF UNSYMMETRICAL PHASORS 419 VO VE ve FIGURE 1.2 Graphical slditon of the components shown in Fig, 121 to obtain three un- balanced phasors, yor (11.2) through (11.3). Now let us examine these same equations to determine how to resolve three unsymmetrical phasors into their symmetrical components. First, we note that the number of unknown quantities can be reduced by expressing each component of ¥, and ¥. as the product of a component of Vz and some function of the operator a 120°, which was introduced in Chap. 1. Reference to Fig. 11.1 verifies the fol ing relations VO = yo VO = yoo VP CVO — VL = aV@ (ia) Y= AVP VM = av Repeating Bq. (1 i)and substituting tgs. (1.4) in Eas. C112) and (11.3) yield Vat cs WAae Ay (Ls) Y= VO + VM + av? (11.6) a ae i a7) or in matrix form fv} faa od “° [yo vy,)=]1 a? @ l]V@)=apvm (1g) v, 1a @&llve Ve ;420 CHAPTER 11 SYMMETRICAL COMPONENTS AND SEQUENCE NETWORKS where, for convenience, we let 1a ot Atle eee] (119) 1 a Then, as may be verified easily, Pao Ape ape eag (11.10) 1 aa and premultiplying both sides of Bq, (118) by A~! yields vo] 11 1 [y 1 ‘a v, VOl=esll a aly) =aoty, (iil) VP 1@ally v. which shows us how to resolve three unsymmetrical phasors into their symmetri- cal components. These relations are SO important that we write the separate equations in the expanded form YP=SY + Wt Vv) (1.2) VO=3V, + av, + art.) (11.13) ; YP=3(V, tah, > av.) (il) If required, the components V,, V6, V,, VO, Vv, and VY. can be found by Eqs. (11.4). Similar results apply to line-to-line voltages simply by replacing V,, V,, and 7, in above equations by V,.. /,, and V,,,, respectively Equation (11.12) shows that no 2ero-sequence components exist if the sum of the unbalanced phasors is zero. Since the sum of the line-td-line voltage phasors in a three-phase system is always Zero, zero-sequence components are never present in the line voltages regardless of the degree of unbalance. The sum of the three line-to-line neutral voltage phasors is not necessarily zero, and voltages to neutral may contain zero-sequence components. The preceding equations could have been written for any sct of related phasors, and we might have written them for currents instead of for voltages. They may be solved either analytically or graphically. Because some of the112. THE SYMMETRICAL COMPONENTS OF UNSYMMETRICALPHASORS 421 preceding equations are so fundamental, they are summarized for currents: I, 19+ 1+ 1 1,=12 +a? 1+ al? (11.15) HID + a 10+ aI 1=3U,+ It 1) IP = 3, + al, + al.) (11.16) 12 =\(1, 4 a7L, + al.) Finally, these results can be extended to phase currents of a A circuit (such as that of Fig. 11.4(@)) by replacing /,, fy. and /, by fs e.and J,,. respectively Example 1.1, Onc conductor of a three-phase Line is open. The the dconnected load through line is 10.4. With the cure rence and assuming that line ¢ 18 open, tind the symmetric the Jine currents. urrent flowing to pt in line @ as components of Solution. Figure 11.3 is a dis ram of the ciraait, The line currents are 0/0A f= 10/180 A, = 0A From Fas. (1116) 1 = $(10/ 0° + 10/ 180° +0) =0 1 =4(10/ 0° + 10/ 180° + 120° + 0) ~/280=5 18/0 A 4(10/0 + 10/ 180° + 240° 4 0) 5 +289 = 5.78/30? A FIGURE 11.3 ‘ Cirewit for Example 1.1422. CHAPTER 11. SYMMETRICALCOMPONENTS AND SEQUENCE NETWORKS From Eqs.(11.4) 19 =0 1 =0 1M=s78/-190 A 1 = 5.78/90? A sa8/ 150° A 1 = 5.78/ -90 & The result 9 =f, = 1 =O holds for any three-wire system In Example 11.1 we note that components {(? and £.° have nonzero values although line ¢ is open and can carry no net current. As is expected, therefore, the sum of the components in line ¢ is zero. Of course, the sum of the components in line @ is 10/0” A, and the sumof the components in line 6 is 10/180" A. 113 SYMMETRICAL Y AND A CIRCUITS In three-phase systems circuit elements are connected between lines a, b, and ¢ in cither Y or A configuration. Relationships between the symmetrical compo- nents of Y and A currents and voltages can be established by referring to Fig. 11.4, which shows symmetrical impedances connected in Y.and 4. Let us agree that the reference phase for 4 quantities is branch a—b. The particular choice of reference phase is arbitrary and does not affect the results. For currents we have (11.17) FIGURE 11.4 7 Symmetrical impedances: (a) A-conrected, (b) Y-connected.IZ SYMMETRICAL YaNDAciRCUITS 423 Adding all three equations together and invoking the definition of zero- sequence current, we obtain 1 = (1, + 1, + 1,)/3 = 0, which means that line currents into a connected circuit have no zero-sequence currents. Substituting components of current in the equation for [, yields 1 4 1P = (1 + 1D + 18) ~ (18410418) DCD 1D) + (LAP) 1.18) Evidently, if a nonzero value of circulating current /? exists in the A circuit, it cannot be determined from the fine currents alone. Noting that / = af? and 12 = 15), we now write Eq. (11.18) as follows: 10 412 = (1 aI + 1 = a1 (11.19) A similar equation for phase 6 is (+1 = (1 —aysfP + — a7)42, and expressing [2 1. 1{P, and J}? in terms of £0, 1, 1°), and 1, we obtain a resultant equation which canbe solved along with Fg, (11.19) to yield the important results 1 = 43/-30 x1) 1% = 5/30 x 1D (11.20) These results amount to equating currents of the same sequence in Eq. (11.19). Complete sets of positive- and negative-sequence components of currents are shown in the phasor diagram of Fig, 11.5(a) In a similar manner, the line-to-line voltages can be written in terms of line-to-neutral voltages of a Y-connected system, v, in Va Bo (1.21) Veg = Kno, = Kau Vin Adding together all three equations shows that 4) = (VV, + Vie + V.,./3 = 0. In words, line-todine collages have no zero-sequence components. Substituting components of the voltages in the equation for ¥,,, yields VE) 4 VIB = (VAD + VAD + VD) = (HD + HD = VR? ‘ab + Val ‘an = (YO ~ VA) + (VAD ~ VIR) + (VE? VB) (11.22) a :4240 CHAPTERIL SYMMETRICAL COMPONENTS ANID SEQUENCE NETWORKS sia) 12) 1 Positive-sequence Negative sequence ‘components ‘components @ ve Positive sequence Negative-sequence components components ) FIGURE 115 Positive- and negative-sequence components of (a) line and delta currents and (b) fine-to-line and ine-to-neutral voltages of a three-phase system, Therefore, a nonzero value of the zero-sequence voltage V{® cannot be deter- mined from the line-to-line voltages alone. Separating positive- and negative- sequence quantities in the manner explained for Eq. (11.19), we obtain the important voltage relations VQ = (1 = aby 1G) = VS L30" x HD ai) VP = Ca VD = 5 £ =30" x VPUa SyMMETRICALYANDaciRcUITS 425 Complete sets of positive- and negative-sequence components of voltages are shown in the phasor diagrams of Fig. 11.5(b). If the voltages to neutral are in per unit referred to the base voltage to neutral and the line voltages are in per unit referred to the base voltage from line to line, the 3. multipliers must be omitted from Eqs. (11.23). If both voltages are referred to the same base, however, the equations are correct as given. Similarly, when line and A currents are expressed in per unit, each on its own base, the v3 in Eqs. (11.20) disappears since the two bases are related to one another in the ratio of ¥3 21. When the currents are expressed on the same base, the equation is correct as written From Fig. 11.4 we note that V, Z. when there are no sources or pen positve- and negative-sequence r mutual coupling inside the A circuit. Wh quantities are both present, we have Van Vu 5 “in is (11.24) Substinuting from Eqs. 11.20) and (1.23), we obtain Y3V.2/ -W 7 = £30" v3 voog sothat jo 737 (11.25) which shows that the A-connected impedances Z, are equivalent to the per- phase or Y-conrected impedances Zy = Z,/3 of Fig. 116(a) insofar as posi- ) noe 1, = 31 (a) () FIGURE 116 (a) Synimetticul A-eannected impedances sind their Y-eanneeted equivalents rekuted by Zy = Z/3s (b) ¥-connected impedances with ceutral connection to ground .426 CHAPTER 11 SYMMETRICAL COMPONENTS AND SEOULNC: NIEFWORES tive- or negative-sequence currents are concerned. Of course, this result could have been anticipated from the usual A-Y transformations of Table 1.2. The relation Zy = Z,/3 is correct when the impedances Zy and Zy are both expressed in ohms or in per unit on the same kilovoltampere and voltage bases. Example 11.2. Three identical Y-connected resistors form a load bank with a three-phase rating of 2300 V and 500 kVA. If the load bank has applied voltage Wal = 1840V 0 1. = 2760V [J = B00V find the line voltages and currents in per unit into the load. Assume that the neutral of the load is not connected to the neutral of the system and select a base of 2300 V, 500 kVA. Solution. The rating of the luad bank coincides with the specified base, and $0 the resistance values are 1.0 per unit. On the same base the given line voltages in per unit are Wl 08 Wyel=12 Wal = 10 Assuming an angle of 180° for V., and using the law of cosines to find the angles of the other line voltages, we find the peranit valucs, 2f =e ¥,. = 1.0/ 180° The symmetrical components of the line voltages are VSP = $(0.8/ 82.89 + 1.2/ 120° - 41.4? + 1.0/ 240° + 180° ) (0.1003 4 j0.7937 + 0.2372 4 j1.1763 + 0.5 + 70.8660) Vin = 08/828 — V,, 0.2792 + 0.9453 = 0.98577 73.6 per unit (line-to-tine voliage base) VP = (0.8/ 82.89 + 1.2/ 240° — 41.47 + 1.0/ 120° + 180° ) 0.1003 + j0.7937 ~ 1.1373 — j0.3828 + 0.5 — 10.8660) = 0.1790 ~ 0.1517 = 0.2346/ 220.3" per unit (line-to-line voltage base) The absence of a neutral conncetion means that 2cr-sequenee currents are not present. Therefore, the phase voltages at the load contain positive- and nepative- sequence comporents only. The phase voltage are found from Eqs. (11.23) with the v3 factor omitted since the line voltages are expressed in terms of the base voltage from line to line and the phase voltages are desired in per unit of the’base114 POWER IN TERMS OF SYMMETRICALCOMPONENTS 427 voltage to neutral. Thus, VE) = 0.9857/ 73.6° — 30° = 0.9857/43.6 per unit (line-to-neutral voltage base) V2 = 0.2346//220.3° + 30° = 0.2346/ 250.3° per unit (line-to-neutral voltage base) Since cach resistor has an impedance of 10/0? per unit 10 5 OTe = 0.9857 /43.6" per unit ve = — = 0.2346/ 250.2 per unit 1 0/ 0 ee The positive direction of current is chosen to be from the supply toward the load. 11.4. POWER IN TERMS OF SYMMETRICAL, COMPONENIS, If the symmetrical components of current and voltage are known, the power expended in a three-phase circuit can be computed directly from the compo- nents, Demonstration of this statement is a good example of the matrix manipulation of symmetrical components. ‘The total complex power flowing into a three-phase circuit through three lines a, b, and ¢ is (1126) ne PAID Kyle + Vt A KL where |, V,, and V, are the voltages to reference at the terminals and Z,, J, and [, are the currents flowing into the circuit in the three lines. A neutral connection may or may not be present. If there is impedance in the neutral connection to ground, then the voltages ¥,, V;,, and V,, must be interpreted as voltages from the line to ground rather thaa to neutral. In matrix notation [uel Ssg=["%e % Ye) | te | (11.27)428 CHAPTER 11 SYMMETRICALCOMPONENTS AND SEQUENCE NETWORKS where the conjugate of a matrix is understood to be composed of elements that are the conjugates of the corresponding elements of the original matrix. Tointroduce the symmetrical components of the voltages and currents, we make use of Eq. (11.8) to obtain 7 2) Ssq = [AVor2]"[Aloia]” (11.28) vor] [2 where Vor = | VS? and Igy = | 1? (11.29) vo [2 ‘The reversal rule of matrix algebra states that the transpose of the product of two matrices is equal to the product of the transposes of the matrices in reverse order. According to this rule, [AVo2)” = Vda" (11.30) andso Sag = Wor A“[AT gio)” = Wi AAG (11.31) Noting that A? = A and that @ and a? are conjugates, we obtain 1 o1]fia i 1 Sop = [VO VO VE} 1 a a 1 a 1) (11.32) 1a @ili a ele 17 or since i le tod fr oo] 2 [i @ a@ li @ a VO yO vol ToT 1 (31.33) 12 So, complex power is Say = Vali + Volt + VoIE = VO1O* BVAIO* + 3VA2* (11.34) which shows how complex power (in voltamperes) can be computed from the symmetrical components of the voltages to reference (in volts) and line currents118 SEQUENCECIRCUITSOF Y AND AIMPEDANCES 429 (in amperes) of an unbalanced three-phase cirauit, It is important to note that the transformation of a-b-c voltages and currents to symmetrical components is power-invariant in the sense discussed in Sec. 8.6, only if each product of sequence voltage (in volts) times the complex conjugate of the corresponding sequence current (in amperes) is multiplied by 3, as shown in Eq. (11.34). When the complex power 53, is expressed in per unit of a three-phase voltampere base, however, the multiplier 3 disappears. Example 113. Using symmetrical components, calculate the power absorbed in the load of Example 11.2 and check the answer, Solution. In per unit of the three-phase 500-kVA base, Eq. (1134) becomes 95 = VOID. YI 4 YTB Substituting the componcnts of voltages and currents from Example 11.2, we obtain Sug = 04 0.9857/ 43.6" x 0.9857/ --a3.6% + 0.2346/ 250.3 x 0.2346/ — 250.3° = (0.9857)? + (0.2346)? = 1.02664 per unit = 51332kW The per-unit value of the resistors in each phase of the Y-connected load bank is 1.0 per unit. In ohms, therefore, From the given line-to-lne voltages we calculate directly 5, = Wee Po Wel? | Weal? p ocd Mel ne Ry Ry Ra oy? + (2300)? 115 SEQUENCE CIRCUITS OF Y AND A IMPEDANCES If impedarce Z, is inserted between the neutral and ground of the Y-connected impedances shown in Fig. 11.6(b), then the sum of the line currents is equal to430 CHAPTER 11 SYMMETRICAL COMPONENTS AND SEQUENCE NETWORKS the current /, in the return path through the neutral. That is, Lalth+tt. . (1.35) Expressing the unbalanced line currents in terms of their symmetrical compo- nents gives Ly = (UO + HP 1D) + (BO HY TP) & (1 1 (1 + 1 £1) + (1 + 19 41) + (1 & BD + 1) a a = 31 ( Since the positive-sequence and negative-sequence currents add separately to Zero at neutral point n, there cannot be any positive-sequence or nega sequence currents in the connections from neutra’ to ground regardless of the value of Z,. Moreover, the zero-sequence currents combining together at 7 become 3/°, which produces the voltage drop 3/°Z, benween neutral and ground. It is important, therefore, to distinguish between voltages to neutral and voltages to ground under unbalanced conditions. Let us designate voltages of phase a with respect to neutral and groundas V,,, and V, respectively. Thus, the voltage of phase @ with respect to ground is given by V, = V,, + V,, where V, = 310Z,,. Referring to Fig. 11.6(b), we can write the voltage drops to ground from each of the lines a, b, and ¢ as Vv, Von My 1 Vib= Van ed 3102, 1 | (1.37) Vey [Yen | [Ys 1] The voltages and currents ia this equir cua be replaced by their symmetrical components as follows: vo 12 fa AL VO] = ZA] LO] + 3192, 1 (1.38) Vo 12 fa Multiplying across by the inverse matrix AW!, we obtain vo Pr] 1 vO] =2,) 1} + 31z,a— fa vo 12 fa :