IMPLEMENTATION OF ADVANCED DISTRIBUTION AUTOMATION IN U.S.A. UTILITIES (Summary) N S Markushevich and A P Berman, C J Jensen, J C Clemmer Utility Consulting International, JEA, OG&E Electric Services, U.S.A. Advanced Distribution Automation implies data preparation, decision making, and control of field devices automatically adjusted to real-time changes of the operating conditions of the power system. Automation of data preparation includes real-time data acquisition, automated data retrieving from various utility databases and other automated systems, and realtime comprehensive modeling of the distribution operations. Automated decision making means real-time adaptive optimization procedures. Automated control of devices in distribution systems means execution of the recommendations of the optimization algorithms by using closed-loop control of switching devices, voltage controllers, and capacitor. Advanced Distribution Automation Systems with realtime closed-loop applications are in different stages of implementation in several U.S.A. utilities. The real-time Distribution Operation Model and Analysis (RTDM), the Fault Location, Isolation, and Service Restoration (FLIR), and the coordinated Voltage and Var Control (VVC) functions are the most effective DA applications. The RTDM application periodically runs in the background and performs the following actions: topology modeling and connectivity checking; load modeling and balancing; capacitor state estimation; facility modeling; modeling of immediate transmission; power flow modeling; voltage related contingency analysis; voltage quality checking; loss analysis; report preparation, including reporting to the operator the summaries of the analysis. The field tests of the accuracy of distribution operation models show satisfactory results. The FLIR function monitors the SCADA statuses of the automated switching devices, the statuses of the fault indicators, and the commanded/uncommanded switching operations of the circuit breakers and automated switching devices along the feeders. It also monitors the pseudo-statuses of other switching devices represented in SCADA and performs the following actions: checks the circuit connectivity and changes the connectivity model in accordance with the real-time situation; checks the loading of the circuits under normal conditions for a given look-ahead time interval; recommends feeder reconfiguration to eliminate overload in case of an existing or a possible near future overload under normal conditions; analyzes the combination of fault indications, switch statuses, and lockout conditions to determine the most likely fault location; recommends operation sequences to the operator after a fault occurs; executes the switching operations via SCADA in closed-loop mode if authorized by the operator. The function selects the optimal reconfiguration based on the following criteria: Minimum unserved kws; no overloads; no voltages beyond the emergency limits; minimum switching operations; preferable use of switching devices which are closer to the faulted feeder; minimum losses. The field test of the Fault Location, Isolation, and Service Restoration application demonstrated satisfactory fault location and optimal service restoration under different load, connectivity, and switch availability conditions. The optimal solutions included simple as well as cascading transfers of loads. The Voltage and Var Control application calculates and executes the optimal voltage controller settings and capacitor statuses while meeting one or a combination of the following objectives: maximize the kwhs supplied within the standard (or better) voltage quality; reduce the peak load while keeping the voltages within either standard limits or emergency limits; minimize the utility production (purchase) cost under conditions of high real-time energy prices; eliminate or reduce feeder segment overloads; eliminate or reduce transmission line overloads; eliminate or reduce under/over voltages in transmission; provide reactive power support in transmission. The field test of the Voltage and Var Control application demonstrated that a significant improvement in the voltage quality and load management could be obtained by using coordinated closed-loop voltage and var control. The integrated field test of advanced DA demonstrated a rapid adjustment of the VVC and RTDM applications to the feeder connectivity changes. Effective utilization of advanced DA requires a significant change of the paradigm in the area of distribution operation control.
IMPLEMENTATION OF ADVANCED DISTRIBUTION AUTOMATION IN U.S.A. UTILITIES N S Markushevich and A P Berman, C J Jensen, J C Clemmer Utility Consulting International, JEA, OG&E Electric Services, U.S.A. 1. INTRODUCTION Advanced Distribution Automation (ADA) implies data preparation, decision making, and control of field devices automatically adjusted to real-time changes of the operating conditions of the power system (1). Automation of data preparation includes real-time data acquisition, automated data retrieving from various utility databases and other automated systems, and realtime comprehensive modeling of the distribution operations. Automated decision making means real-time adaptive optimization procedures. Automated control of devices in distribution systems means execution of the recommendations of the optimization algorithms by using closed-loop control of switching devices, voltage controllers, and capacitors. The stages of implementation of advanced DA are as follows (2): 1. Determine via study the utility readiness for ADA by examining existence of sufficient room for optimization; 2. Determine the feasible upgrade of the distribution system needed for effective ADA; 3. Determine the feasible degree of automation of the particular distribution system; 4. Prove the practicality and acceptability of the recommended automation; 5. Specify the system-wide integrated automated system; 6. Conduct implementation of ADA in phases. The main methodologies of the first three stages of implementation of advanced DA are comprehensive computer-aided simulation of utility-specific distribution operations with and without ADA (3), optimization procedures for selection of automated normally open points, automated sectionalizing switching devices, feeder capacitors and voltage regulators either with local or with central control (4,5), and economic analysis of the expected benefits and cost of ADA (6). The proof of the recommended concept of ADA is provided by conducting a pilot project where the main modeling and optimization algorithms are implemented under typical conditions of the particular utility, and by studying the behavior, the benefits, and the acceptance of the trial functions by utility personnel and by utility customers (7). The specification of the system-wide automation is based on the specific operational and business requirements of the utility, on the results of the feasibility study, on the experience derived from the pilot project, on the availability of the automated systems and utility resources. The advanced DA is specified as an integrated part of the overall automation of the power system (8). The description of the experience with the ADA implementation in the utilities will mainly address the following three applications: the real-time Distribution Operation Model and Analysis (RTDM); the Fault Location, Isolation, and Service Restoration (FLIR); and the Optimal Voltage and Var Control (VVC). Several North American utilities are in different stages of implementation of advanced Distribution Automation. The main characteristics of these applications and the implementation results based on the experience of two US utilities: Jacksonville Electric Authority (JEA) and OG&E Electric Services (OG&E) are briefly described in the following sections. 2. DISTRIBUTION OPERATION MODEL AND ANALYSIS The RTDM application periodically runs in the background and performs the following actions: a. Topology modeling and connectivity checking; b. Load modeling and state estimation; c. Capacitor status estimation; d. Facility modeling; e. Modeling of immediate transmission; f. Power flow modeling; g. Voltage related contingency analysis; h. Voltage quality checking; i. Loss analysis; j. Report preparation, including reporting to the operator the summaries of the analysis, such as: load and voltage violations distances to the limits available change of loads due to voltage change
The real-time Distribution Operation Model is the background sub-function of all advanced DA applications. The accuracy of the model impacts the accuracy of the distribution operation analyses, the correctness of voltage and var optimization, and of the recommendations for post-contingency configurations. The most critical parameters for the objectives of the advanced applications are the voltages in the sites where the voltage limits are checked and the amperes in the most loaded segments of the circuit mains. The accuracy of the modeled voltages depends on the accuracy of the measurement of the substation bus voltage, which is the reference voltage for the distribution operation model, on the correctness of the model of the circuit connectivity and parameters, on the accuracy of the nodal load models, and on the accuracy of the power flow program used as the background of the model. The accuracy of the results produced by the RTMD application was tested by comparing the model results with the output of digital measurement devices placed in the primaries and in secondaries. The manufacturers of these devices guarantee an accuracy of 0.4% for voltage measurements. Typical results of this comparison are presented in Figures 1 (for the primaries) and in Fig 2 (for the secondaries). As seen in the figures, the difference in the model and measurement outputs is very much within the ±0.5% range for the primaries and ±1% for the secondaries. 3. FAULT LOCATION, ISOLATION, AND SERVICE RESTORATION The Fault Location, Isolation, and Service Restoration function monitors the SCADA statuses of the automated switching devices, the statuses of the fault indicators, and the commanded/uncommanded switching operations of the circuit breakers and automated switching devices along the feeders. It also monitors the pseudo-statuses of other switching devices represented in SCADA and performs the following actions: 1. Checks the circuit connectivity and changes the connectivity model in accordance with the real-time situation 2. Checks the loading of the circuits under normal conditions for a given look-ahead time interval (e.g., three hours) 3. Recommends feeder reconfiguration to eliminate overload in case of an existing or a possible near future overload under normal conditions 4. Analyzes the combination of fault indications, switch statuses, and recloser lockout conditions to determine the most likely fault location 5. Recommends operation sequences to the operator after a fault occurs 6. Executes the switching operations via SCADA in closed-loop mode if authorized by the operator. The function selects the optimal reconfiguration based on the following criteria: FREQUENCY (%) DISTRIBUTION OF DIFFERENCES BETWEEN MEASURED AND MODELED VOLTAGES IN PRIMARIES 25% 20% 15% 10% 5% 1. Minimum unserved kws 2. No overloads 3. Minimum voltages below voltage limits 4. Minimum switching operations 5. Preferable use of switching devices which are closer to the faulted feeder 6. Minimum losses. 0% -1.5-1 -0.5 0 0.5 1 1.5 POWER AVE SOUTHPOINT SAN SOUCI -5% CB517 CB540 CB543 CB503 DIFFERENCES (%) Fig.1 Modeling accuracy in primaries AR156 AR154 AR155 AR104 AR101 AR103 FREQUENCY (%) DISTRIBUTION OF THE DIFFERENCES BETWEEN MEASURED AND MODELED VOLTAGES IN SECONDARIES 40 35 30 25 20 15 PHILLIPS CB525 CB527 AR113 AR 106 AR110 AR102 AR107 AR112 AR109 AR115 AR100 AR105 AR114 AR111 CB589 CB572 AR108 CB546 CB547 CB545 BAYMEADOWS 10 GRAVEN SOUTHEAST JAX 5-4 -3-2 -1 0 1 2 3 DIFFERENCES (%) Fig.2 Modeling accuracy in secondaries 0 Fig. 3 FLIR test circuit at JEA The performance of the FLIR function was tested at JEA for the distribution circuits comprising eleven
interconnected 26.4 kv feeders with 25 remotely controlled circuit breakers and automated reclosers, 16 normally closed and 9 normally open (Fig.3), and at OG&E for a circuit comprising three interconnected 12.5 kv feeders with six remotely controlled circuit breakers and feeder sectionalizing switches, four normally closed and two normally open. Several faults were simulated in a section under different load and switch availability conditions. There were eight possible configurations for service restoration at JEA and three at OG&E. The application selected the best solutions in accordance with the criteria mentioned above. In some cases the solution required only two switching operations, and in other cases four switching operations with cascading load transfer. The average time for the closed-loop execution of the recommended switching order was as follows: 1. At JEA, in the two-operation case 90 seconds and in the four-operation case 100 seconds. 2. At OG&E, in the two-operation case 110 seconds and in the four-operation case 170 seconds. These times included time taken by the operator to authorize the switching order. In some instances the service restoration solution was different from the solution expected by the utility personnel and, in the beginning, it raised suspicions that the application did not recommend the correct switching order. The analysis of the situations showed that the application solutions were different mostly because more information was available to the application than to the personnel. In some cases the application took into account that the quality of data exchange with a particular switching device was unsatisfactory, considered the device unavailable, and recommended an alternative solution. In other cases the real-time load flow analysis performed by the application led to a different solution. At JEA, all normally closed sectionalizing switching devices along the feeders as well as the normally open tie devices are protective devices with reclosing and fault interrupting ratings. These devices are remotely monitored and controlled from a central SCADA master computer. In addition to collecting data gathered by the recloser microprocessor (including fault detection), SCADA is able to control the status of the reclosers, change the mode of operations from recloser mode to load-switch mode and vice versa, and change the settings of the relay protection and reclosing cycles. The average number of reclosers (AR) per automated feeder is about three. This number of in-series AR units creates protective zone coordination problems. The configuration of the distribution circuits is frequently changed due to maintenance, load balancing, fault isolation and service restoration. When multiple ARs are installed in many feeders, almost every change of the circuit configuration changes the relative allocation of the protective devices, and the coordination can be lost. A re-coordination procedure following the realtime field reconfiguration is imperative when implementing feeder automation by using multiple protective devices. The procedure should cover all cases of feeder reconfiguration, including maintenance and load balancing schemes created by using both automated and manual switches, paralleling, and restoration schemes following contingencies. 4. VOLTAGE AND VAR CONTROL The Voltage and Var Control application calculates and executes the optimal voltage controller settings and capacitor statuses while meeting one or a combination of the following objectives (1): 1. Maximize the kwhs supplied within the standard (or better) voltage quality 2. Reduce the peak load while keeping the voltages within either standard limits or emergency limits. 3. Minimize the utility production (purchase) cost under conditions of high real-time energy prices 4. Eliminate or reduce feeder segment overloads within either standard or emergency voltage limits 5. Eliminate or reduce transmission line overloads within either standard or emergency voltage limits 6. Eliminate or reduce under/over voltages in transmission within either standard or emergency voltage limits 7. Reactive power support in transmission within either standard or emergency voltage limits. The effectiveness of the Volt/Var Control application depends very much on the actual long-term load-tovoltage sensitivities. Field tests were conducted in several North American utilities at different times of year (9,10). The results of the field test at OG&E did not differ much from the average data from other utilities and were as follows: for real load 0.8% change in kw per percent in voltage change, and for reactive load 4.7% change in kvar per percent change in voltage. The performance of the VVC function in closed-loop mode was evaluated for the nominal voltage objective (provide better than standard voltage maximize the consumption with close to nominal voltage) and for the load reduction objective. The daily ranges of voltage deviations in representative secondary nodes were compared with the daily ranges of voltage deviations with a standard voltage objective. The range was about ± 1% in the first case and about ±2% in the latter case. The load reduction objective was evaluated by comparing measurements of the substation bus load for days with VVC in load reduction mode with the measurements for days with standard voltages (no load reduction). Typical results of these comparisons are presented in Figures 4 and 5. As seen in figure 4, the voltage without DA has been kept close to the upper voltage limit. Implementation of the VVC application in the Nominal Voltage objective resulted in a voltage profile on a lower level, which provided some load reduction. The Peak Load Reduction mode of VVC was
applied for four hours, when the voltage was reduced to the minimum limit. As seen in Fig. 5 the load profile measured during hours with VVC in the load reduction mode is lower than the load profiles measured during hours with VVC in the Nominal Voltage mode. The load reduction effect is in the range of 1% in kw reduction per percent in voltage reduction. V Bus Voltage in 120-Volt scale 124.5 123.5 122.5 121.5 Fig. 4 Substation bus voltage profiles Real load (% of daily average) 128 127.5 127 126.5 126 125.5 125 124 123 122 121 140 130 120 110 100 90 80 70 60 Bus Voltages Profile, 7/14-9/14 0:00 0:55 1:50 2:45 3:40 4:35 5:30 6:25 7:20 8:15 9:10 10:05 11:00 11:55 12:50 13:45 14:40 15:35 16:30 17:25 18:20 19:15 20:10 21:05 22:00 22:55 23:50 Time of the day Average Real Load Profile, 7/14-9/14 0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 Time of the day Fig. 5 Substation load profile 1.9% of Peak Nominal Voltage Objective Load Reduction Objective Voltage without DA Nominal Voltage Objective Load Reduction The VVC application can switch from one mode of operations to another mode either following a prepared daily scenario with different modes of operations at different times or on the demand of the operator, who has the options for changing the modes of the application. The performance of the applications can change automatically depending on the change of the real-time input data, e.g., the cost of generation. When the cost of generation exceeds the customer rates, and if the VVC application is in the real-time pricing mode, then the function will automatically switch to load reduction within the standard voltage limits. The closed-loop execution of the optimal solution derived by VVC is arranged in stages in a way, which minimizes limit violations during the transition from one intermediate stage to another and allows enough flexibility for re-optimization in the case of unsuccessful execution. The centralized voltage and var control system adjusts the settings of the controllers and the states of the capacitors to the changing objectives and operating conditions optimizing the operations within the operational tolerances. In many cases the operating conditions are driven by the application close to the operational limits to obtain maximum benefits. If for some reasons the central control is lost, and the voltage settings and the states of capacitors are frozen, then in a short period of time the operations can be beyond the acceptable range. A concept of backing up the centralized control with local controllers was introduced. According to this concept, if the centralized control is disabled, the local controllers with the technically acceptable settings should automatically be enabled. When the centralized control is restored, the local controllers should automatically switch back to the remotely controlled mode. When all three applications are in service, they perform in a coordinated manner: the VVC runs right after the FLIR finishes the service restoration and adjusts the voltages in accordance with the new configuration. RTDM models and analyzes the new operating condition and informs the personnel about the performance of the newly reconfigured distribution system. 5. NEW PARADIGM IN CONTROL OF DISTRIBUTION OPERATIONS Implementation of the advanced Distribution Automation requires a significant change in the utility operator s control patterns. The traditional rules for controlling the distribution operations are based on the results of off-line studies performed by the planning engineers for a selected set of representative operating conditions and on near to real-time operator s estimates based on available realtime measurements from the distribution system. For instance, the acceptance of the voltages in distribution is evaluated by comparing the substation bus voltages with limits derived for typical conditions. The real-time operating condition can be very different from the typical one. Consider the following examples: The RTDM application models the operating conditions down to the secondaries and calculates the real-time substation voltage limits based on respecting the voltage limits in the secondaries and on actual connectivity and loading of the distribution circuits. The operators should first pay the attention to the model output regarding the voltages in the secondaries. If these voltages are within the limits, then the substation bus voltage may be satisfactory even if it is beyond the traditional range. An overload of a non-monitored by telemetry feeder segment can be determined by the model, while SCADA does not report load limit violations in the
monitored sections. The operator should relay on the model output. Another example is the capacitor control in distribution. Many utility personnel believe that the closer the power factor is to unity the better. However, under some power system conditions and objectives it is more efficient to switch some capacitors off even under peak load conditions. For instance, if there is no overload when the capacitor is turned off, and the loss increase is smaller than the load reduction due to voltage reduction, then the total peak load will be reduced. The FLIR application recommends switching orders based on the current loading, connectivity, switch control availability, and multi-level reconfiguration. The solution can look very different from the ones expected by the operator. The coordination of distribution and transmission capacitors recommended by VVC can be different from the traditionally used by the operators because the advanced application considers many more factors in its optimization procedure. 6. CONCLUSIONS. 1. Advanced Distribution Automation Systems with real-time closed-loop applications are in different stages of implementation in several U.S.A. utilities. 2. The real-time Distribution Operation Model and Analysis, the Fault Location, Isolation, and Service Restoration, and the coordinated Voltage and Var Control functions are the most effective DA applications. 3. Field test results for the accuracy of the distribution operation model are satisfactory. 4. The field test of the Fault Location, Isolation, and Service Restoration application demonstrated satisfactory fault location and optimal service restoration under different load, connectivity, and switch availability conditions. The optimal solutions included simple as well as cascading transfers of loads. 5. The field test of the Voltage and Var Control application demonstrated that a significant improvement in the voltage quality and load management could be obtained by using coordinated closed-loop voltage and var control. 6. The field test of Advance DA demonstrated a rapid adjustment of the VVC and RTDM applications to the feeder connectivity changes. 7. Experience with the advanced DA shows that the requirements for reliability, accuracy and the capacity of the IEDs, along with the requirements placed on communications systems are much more demanding with real-time closed-loop control of distribution operations than with local and supervisory control of operations. 8. Effective utilization of advanced DA requires a significant change of the paradigm in the distribution operation control area. Intensive personnel training is a critical component of successful implementation of advanced DA applications. 7. REFERENCES. 1. Markushevich N, 1993, Voltage and VAR Control in Automated Distribution Systems, 3 rd International Symposium on DA/DSM, 478-485 2. Chan E and Markushevich N, 1996, Justification And Planning Of Distribution Automation, 11CEPSI, 1, 796-803 3. Markushevich N, 1994, "Modeling Distribution Automation", 4 th International Symposium on DA/DSM, pp. 59-67. 4. Jensen C, Markushevich N, Berman A, 1998, Optimizing Feeder Sectionalizing Points for Distribution Automation, DistribuTech 98, 147-150 5. Markushevich N, Clemmer J, Berman A, Royz A, 1998, Analysis of Capacitor Control Under Conditions of Distribution Automation at OG&E, DistribuTech 98, 143-146 6. Markushevich N, Herejk I, Nielsen R, 1994, "Functional Requirements and Cost-Benefit Study For Distribution Automation At B.C. Hydro", IEEE, Vol. 9, No. 2, 772-781 7. Markushevich N, Nielsen R, 1995, Update on DA Pilot Project at B.C. Hydro, 5 th International Symposium on DA/DSM 8. Chan E, Markushevich N, 1994, "Integration of Distribution Automation into Power System Operations 4 th International Symposium on DA/DSM, 277-285 9. Markushevich N, Nielsen R, Dwyer A, Stangl J, 1994, "Load to Voltage Dependency Tests at B.C. Hydro", IEEE, Vol. 10, No. 2, 709-715 10. Markushevich N, Nielsen R, Hall J, Nakamura A, Nuelk R, 1996, "Impact of Automated Voltage/Var Control in Distribution on Power System Operations", 6 th International Symposium on DA/DSM, 100-106