1.4 GENERAL Electric power systems came into service in the 1880s and since that time have grown enormously in size and complexity. Their power generation, transmis sion, and distribution methods and equipment have consistently improved in performance and reliability. The power industry provided the first large-scale application of nuclear energy, was among the first to ‘use analog controls for turbine-generators, and introduced the use of on-line digital control computers. These advances were often mandatory because of the enormous growth in megawatt requirements of the power systems. As_power systems increased in size, so did the number of lines, substa- tions, transformers, switchgear, and so on. Their operation and interactions became more complex; therefore, it became essential to monitor this informa- tion simultaneously for the total system at a focal point, which is now called an energy control center. A fundamental design feature of energy control centers is that they increase system reliability and economic feasibility but system opera- tion is possible without the energy control center. This is essentially a fail-safe type of operation. In practice, all communication links between equipment and the control center could be interrupted and the electric service maintained. For ‘example, most systems maintain “local” as opposed to centralized control of protective switchgear which must function within several cycles of the 60-Hz line upon detection of an overload. The same switchgear may be operated on a slower basis from the energy control center. As another example, a generating plant in the system remains synchronized to the transmission network andFigure 1.1 Interior view of an energy control center. Dispatcher sits at a con- sole containing computer CRT interface, keyboard to enter control commands, and telephone lines. Wall contains a mimic board showing one-line diagrams of the power system with appropriate lights for warning and indications. Figure 1.2 Selecting a numeri- cal value to be changed in the database: The operator points a light pen to a sensitive area of the CRT (poke point); then the com- puter software prompis him to enter the value via the keyboard. ‘The new value is then displayed. (Photograph courtesy of West- inghouse IED, Pittsburgh, PA.)maintains its existing power output level even without signals received from the control center. An energy control center fulfills the function of coordinating the response of the system elements in both normal operations and emergency conditions. The burden of repetitious control in normal situations is delegated to the digi- tal computer, and selective monitoring is performed by human operators. Es- sentially, the digital computer is used to process the incoming stream of data, detect abnormalities, and then alarm the human operator via lights, buzzers, or cathode-ray-tube (screen) presentations. Many “lower-level” or less serious cases of exceeding normal lir -its are routinely handled by the digital computer. For example, if an increase in generation must be allocated to all machines of the system and a generator is already at its high power limit, its share would be distributed among the remaining units. A more serious abnormality detected by the digital computer may cause suspension of normal control functions, In extreme emergencies, such as loss of a major generator or excess power demands by-a neighboring utility on the tie lines, many alarms would be de- tected and the system would enter an emergency state. The operator may be flooded with information and no diagnostics. The present state of the art is that the digital computer programs perform the first attempts at diagnosing the source of several simultaneous alarms, but remedial or corrective actions re- main the responsibility of the human operator, usually called a dispatcher. 1.2 CONTROL CENTER OF A POWER SYSTEM Figure 1.1 is a photograph of the interior of an energy control center. An opera- tor or dispatcher is shown seated at one of several consoles which command the system simultaneously. The interfaces and equipment available to the operator are termed a human-machine interface and often consist of the following: 1. Color cathode-ray-tube (CRT) presentations of transmission networks, substations, data summaries, performance indices, and so on. Selection of presentations is by dedicated function paging buttons, and by light-sensi- tive areas on the CRT which are called poke points. 2. Editing keyboards to change operating conditions or system parameters, by means of typing the alphanumeric characters that are read into the digital computer software codé 3. Special function keyboards to raise/lower transformer tap settings, switch line capacitors, and so on. 4. Light-pen cursor controls to open or close circuit breakers or switches and to activate data presentation or other operations directly on the CRT dis- play. Figure 1.2 shows an operator performing a light-pen action.5. Alarm lights, audible alarms, dedicated telephone communications with system generating stations and transmission substations, and conven- tional telephone links to neighboring utilities. Selected CRT or alarm information may be presented simultaneously on the large wall display (mimic board) of transmission circuits located in front of the consoles (see Figure 1.1). The CRT display monitors and encourages oper- ator actions to the digital computer commands. Displays are characterized by a fixed background and foreground which is updated with measured data. The field, size, and display locations remain fixed until and if the computer source code is edited. New CRT displays can be created and entered into the com- puter system, and the colors, symbols, and blinking or nonblinking formats may be changed from the console through a display compiler. In Figure 1.1, a portion of the transmission network is displayed on an operator's CRT. The bus us voltages, line flows, and locations of breakers, trans- formers, and other equipment are presented on this display. The selection of thé display may be changed by paging forward or backward with dedicated pushbuttons located on the keyboard in front of the operator or by means of the light pen to sénsitive areas of the display called poke points. Using a light pen, an operator may typically do the following: 1. Position the pen to a poke point associated with the device. In the case of a circuit breaker on a power line, the status of the device is represented by a separate symbol, adjacent to the poke point. In the case of a tap-chang- ing transformer, the display field is numeric, adjacent to the poke point. 2. After transmitting the cursor position, the poke point blinks and the cur- sor moves to the home position to give a clear view of the selection. In the case of a two-position device, the pushbutton of the only available option is backlighted. Hence, for a breaker that is closed, the TRIP pushbutton lights. For a tap-changing transformer, both the “raise” and “lower” pushbuttons light (see Figure 1.3). 3. The operator presses the desired pushbuttons, which causes the clearing of the light, the stopping of the poke-point blink, and the logging of the operator request. 4. After successful recognition of the breaker action, for example, the breaker symbol changes either (or both) symbol or color. The symbols and colors (though frequently" the same) are selectable on a per-field basis. In this case, as with all successful operations, a message docu- menting the action is output as a hard copy from the printer. Each type of controllable device is given a preset time limit (usually in terms of computer scan cycles) within which a normal remote operation should be completed. An OPERATION FAILED is shown on the CRT display and logged should the time interval be exceeded.RTA ET ree mS (st ex es az el (a carwrarevaessour | | el Ey f TET} feel [| fos] fae a = = TP | LJ YW JL A HO = Tehelel| (Sisiel= & fer peels [ (= [se] : =) [asl al fe ra] fo ‘|p [re] [22] fee] [ "er ee) EL TH freaee| | rare DO] + [foe sore | | as x [ese el} |e) fes)[=] fm pen} (vevie] jeasen exce JC DATA ENTRY elelee EIDE) |i Figure 13 Operators deticated conrol paneSome of the other control features available to the operator are the major labels on the keyboard of Figure 1.3. These are: I. System commands: may select the mode by which the digital computer controls the entire power system (explained in section 1.4) 2. Units: selects the manual or base loading mode, and so on, for each gen- erator supplying power to the system 3. AGC (automatic generation control) data entry: permits the operator to establish a new steady-state output power for a generator 4. Data/entry readout: values may be entered on the keyboard for subse- quent introduction to the digital computer program 5. Alarms: allows the operator to find the source of the alarm in the system, then correct the source of the alarm 6. Plant/substation select: allows the dispatcher or operator to bring forth presentations on generating plants or substation power distribution points 7. Special functions: allow entry or retrieval of data used to control the power system 8, Readout control: directs output to different CRTs or line printers 9. CPU control: permits either of the two digital computers to be used on- line, and the other applied to nonessential programs 4.3 DIGITAL COMPUTER CONFIGURATION Consistent with principles of high reliability and fail-safe features, electric util- ities have almost universally applied a redundant set of dual digital computers for the functions of remote data acquisition control, energy management, and system security. Both computers have their own core memory and drive an ex- tensive number of input-output devices, such as Printers, teletypes, magnetic tape drives, disks, and so on. Usually, one computer, the on-line unit, is moni toring and controlling the power system. The backup computer may be ext cuting off-line batch programs such as load forecasting or hydro-thermal allocation. The on-line computer periodically updates a disk memory shared between the two computers. Figure 1.4 is a functional block diagram of the dual-computer system. ‘Upon a failover or switch in status command, the stored information of the common disk is inserted in the memory of the on-coming computer. Thus the information used by the on-coming computer has a maximum age of the update cycle (typically 30 sec). Figure 1.5 is a detailed block diagram of a dual-computer configuration and indicates some of the peripheral equipment. All of the peripheral equipment is interfaced with the computer through input-output microprocessors that have been programmed to communicate as‘On-line computer Backup computer Zaurity functions Fastover | Sesto moppere gi os circu Alarms aisk | Logging jovi Seema generation coneot [A] ust negnuion / Stace extimation Unis commitment.” | states Load forecasting i CRT displays, 7 Failover | sc apa computer Operator imrcréace, SCADA computer SCADA co peri Remote data acquisition system Dara PS kets, switches. kW flows. points) | var hows, volte: ampesen, tranetorne? Figure 1.4 ‘Typical digital computer control and monitoring for power sy: tems, SCADA is an abbreviation for “supervisory control and data acquis tion.” well as preprocess the analog information, check for limits, convert to another system of units, and so on. The microprocessors can transfer data in and out of. computer memory without interrupting the central proces: ing | Often, the microprocessors are also redundapt, in that equipment interfaces may be switched to spare units fipon detecting a malfunction. Figure 1.6 shows the physical size of the equipment cabinets. Asa result of these precautions, for all critical hardware functions there is often a guaranteed 99.8% or more availability. Software (computer programs) also allow fOr multilevel hardware failures and initialization of application programs if failures occur. Another feature of the computer system is that critical operating Tunctions are maintained during either preventive or cor- rective maintenance. Besides hardware, new digital code to control the system may be‘compiled and tested in the backup computer, then switched to on-line status. aaa The digital computers are usually employed in a fixed-cycle operating pei Mame otFigure 1.5 _ Detailed dual-computer control block diagram.Figure 1.6 Central computer cabinets. In the foreground are the two compu- tational units, disk drives, and magnetic tape drives. In the background are cab- inets containing equipment and microprocessors to interface the computer to the telemetry channels. (Photograph courtesy of Westinghouse IED, Pittsburgh. PA) mode with priority interrupts wherein the computer periodically performs a list of operations. The.most critical functions have the fastest scan cycle. Typically the following categories are scanned every 2 sec: All status points, such as switchgear position (open or closed), substation loads and voltages, transformer tap positions, and capacitor banks Tie-line flows and interchange schedules Generator loads, voltage, operating, limits, and boiler capacity ‘Telemetry verification to detect failures and errors in the bilateral corn- munication links between the digital computer and the remote equipment _- ‘The turbine-generators are often commanded to new power levels every 4 sec, sharing the load adjustment based on each unit's response capability in MW/min. The absolute power output of each unit is typically adjusted every 5__ min by the computer executing an economic dispatch program to determine the base power settings. Many other system operations, such as the recording of load, forecasting of load, determination of which generators to start up or stop, are considered noncritical, so the computer executes these programs on an hourly basis. ‘Most low-priority programs (those run less frequently) may be executed on demand by the operator for study purposes or to initialize the power system. An operator may also alter the digital computer code in the execution if a pa- rameter changes in the system. For example, the MW/min capability of a gen- erating unit may change if one of its throttle valves is temporarily removed formaintenance, so the unit's share of regulating power must accordingly be de- creased by the code. The computer software compilers and data handlers are designed to be versatile and readily accept operator inputs. AUTOMATIC GENERATION CONTROL FOR A POWER SYSTEM Automatic generation control and economic load dispatch are two principal areas of concern for generation control on large, interconnected power systems. The role.and aim of each is quite different, but both act on the generator through controlling the prime mover and both vary the generation relatively slowly ‘Automatic generation control (AGC) is an on-line computer control that maintains the overall system frequency and the net-tie-line load exchange be- tween the power companies in the interconnection. The common practice is to carry out generation control on a decentralized basis; that is, each individual area [1]' tries to maintain its scheduled interchange of power. Economic load dispatch, described in Chapter 6, is also an on-line com- puter control, whose function is to supply the existing system load demand from all currently operating generators in the most economical manner in terms of minimal fuel cost. A variety of conditions, such as the presence of hydro along with fossil-fuel power stations, a multiarea structure for the inter- connected power system, or even pollution control, can be features of an eco- nomic dispatch operation [2]. Usually, neighboring power companies are interconnected by one or more transmission lines called tie lines, as depicted graphically in Figure 1.7 The electrical “areas” shown in Figure 1.7 are separate power systems under the control of an AGC in a central digital computer. The “boundaries” of an area are the points on the tie lines where a utility’s ownership, maintenance, and loss. accounting ends and those of its neighbors begin. There are very few isolated power systems that are not connected to neighbors by means of tie lines. The power systems employ tie lines for the following reasons 1. Tie lines allow a local or “pool” exchange and sale of power between the power companies on a predetermined schedule. 2. Tie lines allow areas experiencing disturbances to draw on other areas for help. 3. Tie lines provide a long-distance transmission link for the sale and trans- fer of power (¢.g., on an interstate or international basis). "The numbers in brackets in the text indicate references in the list at the end of each chapter.Tie tine Electrical boundary of the area Figure 1.7 Typical interconnected power systems. Interconnections are made so that operating areas can share generation and load, This sharing is normally on a scheduled basis as forced by the AGC. However, during times of disturbance, when an area is unable to meet its own regulating requirements, unscheduled sharing occurs, as dictated by generator governing responses and by the contributory function of the AGC frequency bias of each areas. Such unscheduled interchanges persist until the disturbed area can either itself fully respond to its local requirement, or until normal schedules are reset so that the contribution is taken out of the unscheduled class and put into the scheduled class. 1.4.1 Area Control Error To maintain a net interchange of power with its area neighbors, an AGC uses real power flow measurements of all tie lines emanating from the area and subtracts the scheduled interchange to calculate’an error value. The net power interchange, together with a gain, B (MW/0.1 Hz), called the frequency bias, as a multiplier on the frequency deviation is called the area,control error (ACE) and is given by ~ ACE= ¥ Pk -P,+ 10B(fia— fo) MW k=l »v Here P, is a MW tie flow defined as positive out of the area, P, is the scheduled MW interchange, and f, is the scheduled base frequency. When a system is not interconnected, only the frequency term is used. The interchange power P, is generally scheduled for periods of the day and is changed as “blocks” of MWh are bought or sold to neighboring utilities. A positive ACE or positive net ex- change of power represents a flow out of the area. Figure 1.8 schematically shows the ACE signal used by the central computer to control generation within the electrical boundary encompassing the system’s transmission net- work.wtf \ -e— Transmission meuwork ‘\ +4 LZ] sums — \ [7p generators \ P+ h08ha, rej ze | Scheduled { interchange | | | Figure 1.8 Using ACE as an input to the AGC. (Tie-line power. flow is mea- sured and telemetered to the central control computer. Four tie lines are shown.) The tie lines are generally connected into the transmission network at lo- cations where their specific power How. must be established by adjusting or shifting the power oifiput of generators in order to achieve a desired How value. fotice that the Teal power summation of the ACE calculation loses informa- tion as to the flow on individual tie lines; thus the AGC sensing only ACE does not control the flow on the individual tie lines but is concerned with area net generation. Often, the tic lines transfer power through the arca_from_one neighbor to the next, Called wheeling power. The wheeling power cancels alge- Sraicaily in the ACE. However, this is exactly the process whereby one area purchases or sells blocks of power (MWh) with nonneighbor utilities, For ex- ample, consider the four-area power pool shown in Figure 1.9, which is operating at 60 Hz, so the frequency error is zero. If area A desires to sell a block of power, p, for | hr beginning at 1 a.M., it would introduce p into its own ACE as a scheduled interchange (with sign change), causing its tie lines to ex- port power until its AGC forces equation 1.1 to zero, Simultancously, arca C introduces —p into its ACE (with sign change) so that power flows into area C until its own AGC reduges equation 1.1 to zero. Notice that areas B and Dfo Sacx = 60 He Figure 1.9. Sale of power from area A to area C. usually participate in this interchange because of the interarea tie lines, so they must be informed of the transaction to determine possible tie-line overloads. Generally, the characte: pattern of area B and D network flows are calcu- lated or known a priori and stored as tie distribution factors for the case when neighbors A and Care exchanging power, so critical situations are avoided. It is common practice in the United States for neighboring areas to use “after the fact” accounting on a monthly basis for the use of their facilities for wheeling power, with the charge based on a percentage cost of the power exchanged. The economic sale or purchase of power and the use of the distribution factors in these transactions are discussed in Chapter 6. The minimum requirements of the AGC on controlling the-interchange of power and frequency have been established by NERC (North American Electric Reliability Council), which is comprised of representatives of the major operating power pools. This committee, which superseded earlier orga- nizations, specifies the following criteria as minimum performance expected from AGC as of 1983. A. Normal system conditions 1. ACE must equal zero at least one time in all 10-min periods. 2. The average deviation of ACE from zero for all 10-min periods must be within allowable limits based on a percentage of system generation. B. Disturbances conditions 1. ACE must return to zero within 10 min, 2. Corrective action (by the AGC) must be forthcoming within 1 min of a disturbance.Maximum rate ~~ Peak daily load ————— of load change Peaking Toad Averaze aly load 4 Intermediate Toad Figure 1.10 Daily load cycle for a power system. (Maximum rate of load change is indicated.) Minimum, daily load Minimum dally toad 12PM, 3AM, 6AM, 9AM 2AM 3PM. 6PM, 9PM 12 P.M. Time of day The allowable limit, Ly of the average deviation on power systems (averaged over 10 minutes) £4=0.0254L+50 MW where AL = AP/At, MW/hr, is the greatest hourly change in the net system (native) load of the control area on the day of maximum peak load. The value of AL is determined annually and istaken from the daily load cycle as shown in Figure 1.10. The maximum rate time varys from system to system, as does peak time. A power system is defined to be in a disturbance condition if the ACE _ signal exceeds 3Z4. This could occur due to the tripping of a generator or a sudden change in system load. The requirements on the AGC are directed to- ward returning the system to normal operation. A typical ACE signal of an AGC that meets minimum performance is shown in Figure 1.11, where the av- erage value is less than 100% of the limit allowed, and the peak excursions are acceptable. t Overgeneration Undergeneration { ime ing -100 Percent permitied ACE Figure 1.11 Example of ACE close to permissible limit.1.5 OPERATION WITHOUT CENTRAL COMPUTERS OR AGC It must be emphasized that power systems are entirely capable of operating ut a central Com central computer and/or AGC, This is a result of turbine-generator speed controls built into generating stations and natural load regulation. These characteristics force generators within an area to share‘Toad and cause connected power areas to share load. To understand the basic concepts involved in the operation of an inter- connected power system, the hypothetical model shown in Figure 1.12 will be analyzed [3,4]. First assume that breaker 7 is open and there is no between areas A and D. Area D represents an operating area of connection in which a Sudden load or generation change occurs. Area A is con- sidered as a single operating area representing the remainder of the intercorinection. This implies a basic assumption in regard to the interconnec- tion; the many operating entities that comprise area_A share a composite ac- tion. Furthermore, it is assumed | di 166 i their Bensrating-capacity size and operating characeristics. Both of these fun- damental assumptions are based on operating experience. Let the area A over- all generation-frequency characteristic be represented by curve GG in Figure 1.13, which is a composite response curve from all the generators in area A. Figure 1.13 essentially summarizes how the rotating shaft speed of steam tur-_ bines, gas turbines, diesel engines, or other prime movers Vales with th 2 elec Figure 1.12 Representation of a sim- ple interconnected system. 60.5 60.0 Frequency (H2) 59.8 Gy base Excess Deficient generation generation Generation (MW) Figure 1.13 Area 4 gencration-frequency characteristic curve GG.trical load on the generator. The shaft speed—and consequently the electrical fine frequency—changes with load reflected onto the The generation-frequency characteristic curvé has a negative slope, or droop, with frequency because each turbine-generator control is a “type 0” control. Typically, the prime mover has approximately a 5% speed droop from no load to full load. For the purposes of the present study, the assumption is made that over the restricted frequency and loading range under consideration, the generation characteristic has a constant incremental slope. Each turbine- generator control may have Ad ies and deadband regions, but consid- ering a number of such generating situa a, & Composite would be as in Figure 1.13. The area connected load is defined by curve LL, as shown in Figure 1.14, which may be intuitively envisioned as the increase in load when the rotating machinery in the area is forced to increase speed. Although difficult to verify by experimental data, these models are gener- ally accepted. Therefore, the basic equations describing generation and load are G4 = Gy + 108; (face fo) = MW (2) Ly = Lg + 10B2 ( fact — fo) MW (1.3) where G, = total generation on system A, MW base generation on system, MW at 60 Hz otal load on system 4, MW + Lo = base load on system, MW at 60 Hz, yystem frequency, Hz ase frequency (60 Hz), often a scheduled quantity 8, = cotangent of generation-frequency characteristic in MW/0.1 Hz; this quantity is negative in sign and is called the natural generation governing, 8, <0 By = cotangent of load-frequency characteristic, MW/0.1 Hz, 6 > 0 Tp base load Load (MW) Figure 1.14 Effective area A load-frequency characteristic curve LL.Sec. 1.5 Operation Without Central Computer or AGC 17 ator CC @ composite generator plus load “ - aw) Figure 1.15 Isolated operation in area A in response to a load change. For a steady-state frequency, total generation must equal total effective load, and prevailing frequency edefined by the point of intersection Zo of the GG and LL curves, shown at 60 Hz in Figure Lis — ———Thé generation characteristic and the load characteristic can be added algebraically to obtain the combined area characteristic, as shown by curve CC in Figure 1.15. The composite generation load-frequency characteristic is given by Ga — La = Go + 1081 Saci — So) — Lo — 1082 (Fact ~ fo) (4) Now assume that there is a load increase in area 4 of magnitude to move the load-frequency characteristic to position L’L’. The new system frequency will now be defined by the intersection of GG generation line and the new load me L‘L’. This point of intersection labeled as J, is shown in Figure 1.15 at 59.9 i. If it is desired to return system frequency to 60.0 Hz, this is possible by of offsetting the automatic generation control (AGC) in area 4, shifting geacration curve GG to the new position G’G’, which intersects the prevailing L-E’ load line at 60.0 Hz, point /;. The resulting combined characteristic is now shown by line C’C’. Both G’G’ and C’C’ curves intersect the Hz coordinate. However, if there is no isochronous (constant frequency) sensing in AGC ta pesform corrective action, only individual generators and foad regulate, and ome. actio crators and foad regulate, and << Equation 1.4 may be written in terms of increments asAg = Gy — Go + Ly — Ly = 108, (face ~ fo) — 10B2 ( face — fo) = 10B 1X4 (face — fo) a5) =10B,X,Af MW where B, is the natural regulation characteristic of area A expressed in percent of generation per 0.1 Hz, and X, is the generating capacity of area A in meg- awatts. Thus, should the load increase + A, MW or generatiorrdecréase in area A, the resulting frequency deviation would be = Hz 6) which is a negative quantity because By frequency) decreases due to added Toa According to equation 1.1, define a real power tie-line flow, AT, as a pos- itive quantity out of the area. The combined effect on frequency for a foad in- crease (of generation decrease) and positive tie flow on area A is then negative. The system speed (ac line Ag+ AT, “-Sonx, | He (16a) where A, + AT; is the net megawatt change. To understand the effect of tie-Tine flow when systems A and D are inter- connected, assut nat the breaker T' of Figure 1.12 is closed with generation and load equal at 60 Hz in both aréas. There is initially no tie-line power inter- change between the two areas..Now a disturbance occurs in area D which causes the common system frequency to drop to 59.9 Hz. Assuming that a sup- plemental control or AGC does not act, the area generation no longer matches the effective load within arca A. The difference between the two is defined by the difference between intercepts J, and J), respectively, of the GG and LL curve with the 59.9-Hz line as shown in Figure 1.16. The difference between generation and éfféctivé Toad, or the net excess power in the area, flows out of area A over the tie lines as a cgntribution to the disturbance in area D. The con- tributory effect is comprised of two co nts: — AL, which represents a de- crease in load power in area A,"and AG, which represents the increase of. generation in area A with decrease in frequency. The tie-line flow between A and Disthen ~~ AT, =AG,— dL, MW (7) where AT; is the netchange in tic-line power flow from initial conditions, which is a positive value directed from A toward D. If area A has an AGC that applies frequency bias, B, the effect of the bias is to increase the MW/Hz. response of the generation. The G’G’ and GG” lines60.1 Increasing bias, Frequency (He) 8 & 539.9} ar; — —— ar; ———-— Figure 1.16 Governing and regulating characteristics of area 4 in intercon- nected operation. of Figure 1.16 represent increasing amounts of bias applied t by the AGC. Cor- responding tie flows AT;, and AT are larger. Bias 1s discussed in Section 1.7. For area D, AT; is the tic-line power flow directed from A to D. Power flow into the area appears as a logd.decrease, so the frequency change due to a disturbance Ap and tie-line flow from 4 to D is ~ Ap- AT, = Av— AT 1 4f = “op, Xp (8) Let A,p = Ap be the magnitude of the disturbance that occurs in area D and A, = 0. Since the frequency is common to both systems, AT, _ Aan = ATx “T0B,X, 10BpXp Hz (sy Solving equation 1.9 for the tic line flow, we obtain: - (OB .X4 Aad - BaXadap w AT. = 10B,X, + 10B>Xp M (1.10) The net power change in area D is: 10BpXp)Aan Bap — ATL = (OB pXv)Ax MW 1) 10B,X, + 10BpXp From this derivation it is seen that the interconnected system comprisedof areas A and D both participate in sharing disturbances as weighted by their generating capacity: a Aap = 10B,X, Af + 10BpXp Af = (10B,X4 + 10BpXp) Af MW Thus by interconnecting the systerhs, the frequency fluctuations due to distur- ‘bances are reduced and system performance is improved by means of tie-line flows. Equation 1.12 is rewritten in transfer function form to show this: —_ of 1 Bap 10B,X, + 10BpXp (12) Hz/MW (12a) For more than one tie line between areas 4 and D, equation 1.10 de- scribes the net flow between the areas. The power flow on each of the multiple tie lines connecting areas is dependent on transmission-line impedances and the location Bf gensrationand Toad soit is ca Seneratiear and Mead to i is calgulatea By The meteor ‘Chap- ter 5. If WO oF more areas are interconnected, Only the net flows into or out of each area are estimated by méans of generalizing equations 1.10 and 1.11. Example 1* ‘Two areas are interconnected (see Figure E1.1). The generating capacity of area A is 36,000 MW and its regulating characteristic is 1.5% of capacity per 0.1 Hz. Area D has a generating capacity of 4000 MW and its regulating characteristic is 1% of capacity per 0.1 Hz. Find each area’s share of a +400-MW disturbance (load increase) occurring in area D and the resulting tie-line flow. Solution Using equation 1.12a, the change in frequency is A — Aap 10B,X, + 10BpXp 0 10(0.015 (36,000) — 10(0.01)(4000) = —0.068964 Hz af The tie-line flow is given by equation 1.10: = U0B4X4)Aap _ _ 400)400 _ AT: = T08,X, + 108 Xp ‘5800 37.4MWw Through the interconnection, the larger system considerably aids the other system. The smaller system needs only to absorb a 27.6-MW load. The frequency regulation is much better; for if area D were alone, its frequency would decrease 1 Hz = A4p/(10BpXp) due to the increased load. a oe ~ by wee te for ar — Disturbance, Aa Figure E111.6 PARALLEL OPERATION OF GENERATORS The tie-line flows and frequency droop described for interconnected power areas are composite characteristics based on parallel operation of generators. That areas must have speed or frequency droop as opposed to isochronous (constant-speed) operation is obvious, for if each aréa could maintain its speed ¢ synchronizing torques, then a load common to both areas, by supérposition, would have the terminal voltage Vioaa = Vy sin wit + VY; sin wz (1.13) where subscripts 1 and 2 refer to the areas and ¢ is time in seconds. Combining the terms of equation 1.13 results in line frequencies that are the sum and dif- ference of f, and f;, which is objectionable. Although it is possi! erence frequency for both areas, in principle both areas as well as generating units must be capable of independent operation should communications links be interrupted. A generator speed versus load characteristic is a function of the type of governor used on the prime mover (type 0 for a speed-droop system, type 1 for an isochronous system, etc.) as well as the capacity of the generator. Consider an extreme case where generator 1, of limited capacity, is paralleled to an infi- nite bus of constant frequency, as shown in Figure 1.17. (An infinite bus can absorb or supply unlimited power at constant voltage and constant frequency.) In this figure, the gencrator-droop characteristic is such that it is loaded to 50% Rowse. # af Generator Tartine vat 1.02 0.98 Speed or frequency iperunt) oninn S| SDSS o oF To Load, P (per unity Figere 1.17 Parallel operation of a “droop” generator with an infinite bus.Constant system _~ frequency (infinite bus) Increasing uunit load Speed or frequency (per unit) motor gx | => Seema 0.5 1.0 Load, P (per unit) Figure 1.18. Adjusting prime-mover torque to load a generator. of its capacity when paralleled to the bus. The regulation of the unit with an implicit algebraic sign is defined as Afipu) _ __AfCH2)/60(Hz) __ AP(@p.u.) AP(MW)/P.ac(MW) where Prac is the megawatt rating of the generator and p.u. represents “per unit.” The regulation is assumed to be constant for the range of interest here. The governor shown in Figure 1.17 has a steady-state regulation of 4%. If it is desired to increase the load on the generator, the prime-mover torque is in- creased, which results in a shift of the speed-droop curve as shown in Figure 1.18, unit speed regulation = R = (1.14) By means of adjusting the prime-mover torque the power output of the generator is set to the desired level, including motor operation. The shifts in generator output are performed by means of momentary shaft speed changes with respect to the infinite bus at constant frequency. Thus Figure 1.18 is equivalent to changing the shaft reference angie 9, of the synchronous machine shown in Figure. 1.19. For a simplified, cylinarical rotor machine the real power flow is given by 1? gin (, — 82) 1.15) Infinite bus vile, Figure 1.19 Synchronous generator Prime-mover operating > irequency g Eh- f £ By =. = 3% r 3 E Peal E s operating é z teauency 2 Py Prete Pap Prete Unie H output (MW) Unit 2 output (MW), Figure 121 Two different generators operated in parallel. Since the frequency is common to both units, they will share according to the ratio AP, _ Rp Pirate SF LA? BP. Ry Parwe a If the initial load is P,o + Pjo, a change in load is satisfied by AL =P, + AP, = 2 Pune 4 Af Powe gw (1.18) Ri Ro Hence the equivalent regulation of the paralleled system is Royse Zz a 1/MW (1.19) Z y Paate Ri Re or in terms of per unit, with the system capacity as base, this is Ryysrom = Reysiem-(Priate + Preate) pu (19a) such that the performance of the system is a capacity-weighted combination o the generators. A change in load AL is satisfied by the two units in shifting the operating points from 4,, A to B,, B, on Figure 1.21, with thé magnitude of the increments given by equation 1.17. * Notice that if 10B, = 1/R, the parallel operation of two generators ex- pressed by equation 1.19 is exactly analogous to equation 1.12a, which de- scribes interconnected power systems. S Example 1.2 ‘Two synchronous generators are initially supplying a common load (see Figure E1.2) at I p.u. frequency (60 Hz). The rating of unit | is 337 MW and-has 0.03-p.u. droop built into its governor. Unit 2 is rated at 420 MW and has 0.05-p.u. droop. Find each unit'sFigure E12 Load share of a 0.10-p.u. (75.7 MW or 10% of the total generation) increase in the load de- mand. Also find the new line frequency. Solution The equivalent system regulation is a weighted average of both units. Ronen ~ A Bo (Prrse + Porate) R 7 - 337, a (337 + 420) 0.03 * 0.05 = 0.0386 pu ‘The new system frequency due to a load increase is ASP.U.) = Royster AL(PU.) = -0.0386 x 0.10 = ~0.00386 p.u which is a decrease of 0.231 Hz. Both generating units have the same frequency. Their share of the load is determined by their regulation: Ap, = MBB) _ “000886 9.129 pu. +433 MW 0.00386 = uM. +324 os = 0.0771 p. MW 1.7 AREA LUMPED DYNAMIC MODEL ‘The power system model described thus far is one of macroscopic behavior in the sense that no effort has been made to indicate instantaneous power flow within the system due to’a tie-line disturbance, magnitudes of the internal line flows, the time history of generator phase angles, and so on. In fact, the simpli- cation from a transmission-line, generator study to the macro model may be impossible to justify analytically. One can consider the macro model as having evolved from both power system operating experience and parallel operation of strongly coupled generators within the power system. These independent generators could not share a common load unless the turbine-generator con- tollers have droop characteristics, or else a noninterruptible, differential powerA Lg R Governor | Tasbine Ref t senerstor Area-response Governor Turbine | APs + Ret? mye >| generator sual : Ay, load I JT, tie lines t r 3 1 1 i t { ! ' L ! Ret. M . Other generating units Figure 1.22 Equivalent model for generation within a power are~. signal is available to every turbine-generator of the system as well as the inter- connection neighbors. ‘The power system macro model may be described by means of a block diagram [5] as shown in Figure 1.22, where the following definitions are used: H, = effective inertia of the rotating machinery loads in the power system area as normalized by the generating capacity; the numerical vaiue of 2H4/10Bz is 2 to 8 sec for ail systems [6] 8 = load frequency characteristic, MW/0.1 Hz Pyrate = Tated power output of generator i, MW AP, = electrical power increment for generator i, MW 1/R, = droop characteristic of generator i, Hz/MW (The algebraic sign is taken into account at the summing junction.) ‘S = Laplace transform variable, 1/sec In this model, which represents an isolated power system without tie lines, the steady-state value of the frequency deviation, Af, for a load change AL_ 4.,/S is found using the final value theorem of Laplace transforms as(1.20) (= where the droop characteristics of the M generating units combine according to equation 1.19: ) ~ 0p, — 1083 ss. (21) Ru With this definition, the static response of the dynamic system of Figure 1.22 is consistent with equation 1.5 for an area. The dynamic system is assumed to be a linearized, small-perturbation model valid for +t 5% variations about the base generation. The area dynamic response is approximated by a single time constant r= 2H,,/10f>. That is, in the absence of turbine-generatof action, if generation or load were suddenly to change ffom base value to an increment A,, the fre- quency deviation would exponentially reach its final value: . Bay ry Ae ye tomtaity) of Tops! ova tog, | e ‘A) (1.22) where ¢ is measured in seconds. Numerical values for 8, and H, which are measured by such a test are difficult to predict for a power system area. The linear smali-perturbation transfer functions for the governors and turbine- generators are available from design data and unit experimental tests. As a re- sult, the governor-turbine-generator models are well known for fossil-fired units, hydro stations, combined cycle, and nuclear, so that they may be mod- eled with any desired degree of accuracy ranging from a single time constant to multipath transfer functions, including deadbands and nonlinearities. The slower-response clements, such as boilers, dominate the governor-turbine- generator transfer functions. For example, a drum-type steam unit with a “tur- bine-leading” control arrangement has its slowest time constant on the order of seconds, while a “boiler-leading” unit has pertinent time constants on the order of minutes. Integrated boiler-turbine control units and hydro stations range between these extremes for small disturbances. ‘Next, consider ths isolated power area where the central computer has an AGC sensing only frequency in equation 1.1. This is called “flat frequency” control of the power system. Assume that the central computer command rate to each generator / is approximately a continuous signal ‘G{s), as shown in Fig- ure 1.23. In order for the AGC of the central computer to allocate the fre- quency error as a change in generation among units, the equivalent transfer function and gain in cascade with urbine-generator appear in the overall response for a load change. A direct reduction Of the block diagram ey Mason’s rule [6] for block diagrams yields the transfer function: Mason's rule [6Ms) AL(s) 4 QH,S + 1082)+ ZS [Gks)/R] + rox] ist — = (1.23) = oxeno| 1 The general behavior of this transfer function is dependent on the com- plexity of the G, and Gj, but for steady-state values assume the following: G(s) | ~o = | = per unit value for unit 7 (1.24) Gis) | .-o = K; = 0, gain allocation to unit i (1.25) where the relative generating capacity of different units is included in R, and K, With these assumptions G/ is a type 0, or proportional AGC, so that the Final value theorem of Laplace transforms yieids the following for step changes in the load, AL(s) = A4/S: ym AS] = (2) = RR 1.26) 50 SS 108. + S U/R)+ 10BS K, 1 - ‘Governor turbine generator 1 Gus) Area response Governor S turbine Ref 2 GS) Lt ar TH, 108, 108 | Frequency Figure 1.23. Isolated power area, AGC regulating frequency.Gone Frequency Isochronous operation only ‘Load applied AGC applied ‘Time (unspecified units) Figure 1.24 Representation of load and AGC effects. In equation 1.26 it is evident that an AGC sensing the frequency error addi- tively contributes to the natural regulation of the area since the frequency change is less for a given disturbance. Th® contribution of the AGC is often called a “supplemental control” [4], and its effect on the system dynamic re- sponse eee oe ‘and G(s). For the purposes of illustration only, a sudden change in the load affects the area as shown. schematically in Figure 1.24. The time scale is not defined in this figure, and the interval be- tween sensing the error arid corrective action by the AGC is expanded, but the effect of the bias, B, is evident. The entire AGC dynamic response, including ACE filtering and time lags associated with digital computer system scan rates, are included in Gis) and its digital equivalent. If the AGC introduces an integrator into the loop, such that Gis) = asso then the final value theorem of Laplace transforms yields i Afs) 7 lis A} = hi oO $0 {ara} 2, { M Mf 1022+ ED 1/R; + 10B E K, i=] Ss which would make the power system isochronous, or force the system back to base frequency following a disturbance, as shown in Figure 1.24. An isolated system may have isochronous operation, but it is unacceptable for intercon- nected operation because the droop characteristic is essential for load sharing. In the Eastern United States power interconnection, the 24 hr. accumu- lated time error due to underfrequency operation (or overfrequency operation) is corrected at midnight E.S.T. by increasing (decreasing) the desired frequency fy in equation 1.1 for a period of time. The Western United States interconnec- ‘tion continuously introduces a time correction term into equation 1.1. How-ever, in both interconnections, the steady state frequency remains very close to 60 Hz. In subsequent chapters it is assumed the generator droop characteristics have forced the system into 60 Hz steady-state power-flow condition. In the steady state, the individual line power flows and network state may be deter- mined from voltages, loads, and line parameters. Initially, an isolated power area, or one in which the tie lines have been replaced by an equivalent load, is considered. This simplication limits the number of lines and elements to a rea- sonable size to begin the analysis of power systems. PROBLEMS 1.1. A power system has a peak-load demand of 1000 MW and is supplying this peak with the on-line units specified below. The area load is assumed to have a 1% change per 0.1 Hz frequency deviation. Determine the composite generation—load frequency dependence for the area. Normalize the result to a 1000-MW base. There is no automatic generation control (AGC). ‘Type of Unit Nominal Rating (MW) Regulation (%) Coal-fired 337 5 Coal-fired 420 6 Oil burning 100 4 Oil burning 100 4 Oil burning, 100 4 Natural gas (peaking) 50 1 Natural gas (peaking) 50 Il 1.2. Areas A, B, and C (shown in Figure P1.2) comprising a power pool initially have zero tie-line flow before a 1050-MW load decrease occurs in area B. Each area has AGC-assisted frequency regulation, which results in 1.5 times its combined load and generator natural characteristic. The frequency is 60.0 Hz. before the load loss. Data for the areas are as follows: Area A base generation = 10,000 MW natural regulation = 1% of base per 0.1 Hz © ©) Figure PL2Chap. 1 Problems, 31 Area B: base generation = 7000 MW natural regulation — 2% per 0.1 Hz Area C: base generation = 11.500 MW natural regulation = 4% per 0.1 Hz (a) What is the system frequency after the load loss? (b) What net tie-line flow (MW) does each pool member have after the distur- bance? 1.3. A two-generator, three-bus power system equivalent is shown in Figure P13.1. Generators 2, = 1.0/8) Terminal 2 by = 1.0 f63 ‘Terminal 3 1 = +90 iw Area inertia and Trequency, dependence Figure P1.3.1 variable ‘Network resistive losses are neglible and all terminals are operated at 1.0-p.u. volt- age. Incremental power changes are desired, so the approximation sin 8 ~ 8 may be used for power flow (equation 1.15). The algebraic signs at the summing junction of the area inertia and load characteristic correspond to power-flow terminology where a load is a negative real injection to the network. (a) Calculate the real power delivered to the load in terms of the variables 6, and 5, when the system reaches steady state. ‘The generators are as follows: Unit 1: 337-MW rating, 0.05-p.u. frequency droop Unit 2: 420-MW rating, 0.06-p.u. frequency droop A simplified block diagram for the power demand-frequency response for the generators is shown in Figure P1.3.2. where