Principles of the Method

The procedure developed for this purpose may be defined as a parallel Monte Carlo method aimed at estimating the probability density functions of reliability indices related to frequency of interruptions, duration of interruptions and non-supplied energy. It can be applied to distribution systems (either overhead or underground) with or without distributed generation.

Input data include system parameters (i.e., network topology, component parameters, including setting of protection devices) and yearly variation of loads and generations. A failure is caused by a fault, a term to which several random variables are related: location, time of occurrence, duration, type. Random variables to be generated during the application of the Monte Carlo method implemented for this work are those related to failures rates, fault characterization and reconfiguration times. The use of a power flow simulator will allow the estimation of the non-supplied energy.

The procedure may be summarized as follows:

1. Run the test system during one year using time-driven simulation and a constant time step (e.g., 1 hour). This run, known in this work as base case, will provide basic information (e.g., energy values) that will be used for later calculations. The simulation can be carried out, depending on the system under study, with or without distributed generation

2. Estimate in advance all the random values related to the faults/failures to be simulated (location/component, time of occurrence, duration, type) for one year. 3. Run the test system again but considering now the possibility of equipment

failure. Regardless of the location, type and duration of the fault, a protective device will always operate. Reliability indices are updated once this sequence of events is finished.

4. Repeat the procedure from step 2 as many times as required to obtain accurate information for reliability index calculation.

The stopping criterion used for assessing convergence is the coefficient of variation (CV) [5.18], which helps to determine if enough executions have been performed in

Chapter 5: Reliability Assessment of Distribution Systems

119 order to estimate a variable’s probability density function (PDF). The Monte Carlo method is assumed to have converged when the CV of all calculated indices is below 5%. Reliability indices are calculated as recommended by IEEE Std 1366 [5.19].

5.2.2. Scenarios

Three different scenarios have been considered to calculate reliability indices with and without distributed generation:

1. A first scenario assumes that only protective devices operate; that is, a protection device locks out after a permanent fault, and service is not restored until the faulted component is repaired.

2. The second scenario considers switching operations aimed at isolating the faulted section. This design may restore service only to load nodes upstream the failed component.

3. The third scenario includes also transfer switches between feeders, so system reconfiguration may be used to restore service also to some load nodes downstream the faulted section, depending on the system design.

The specific steps taken by the procedure when a fault occurs depend on the scenario considered.

First Scenario

1. System is simulated until time reaches the moment when a fault occurs, time- step is one hour.

2. The fault is simulated by placing a short-circuit at the element considered for failure.

3. Simulation time-step is reduced to allow the correct response of protection devices.

4. Simulation time-step is increased to continue with the simulation and perform switching actions; new time-step is 5 minutes.

5. Reconnect DG units disconnected during the fault but can be placed back in service immediately.

6. After repair is finished, return operated protection device to its original closed position, restoring service to all loads.

7. Reconnect DG units that were disconnected.

8. Continue simulation until the time for the next fault is reached. Time-step is increased to one hour.

If a one-phase fault occurs on a distribution line protected by a fuse, two cases can be considered:

• Voltage at load point above 0.9 p.u.: the repair will be carried out in a normal manner; loads will not experience any interruptions.

• Voltage at load point below 0.9 p.u.: the remaining phases will be opened, isolating all elements downstream from the operated fuse.

If a two-phase fault occurs on a distribution line protected by a fuse, the remaining phase must be opened to disconnect all voltage sources from the failed zone.

Analysis of Power Distribution Systems Using a Multicore Environment

120

Second Scenario

1. System is simulated until time reaches the moment when a fault occurs, time- step is one hour.

2. The fault is simulated by placing a short-circuit at the element considered for failure.

3. Simulation time-step is reduced to allow the correct response of protection devices.

4. Simulation time-step is increased to continue with the simulation and perform switching actions; new time-step is 5 minutes.

5. Reconnect DG units disconnected during the fault but can be placed back in service immediately.

6. Failed element is disconnected.

7. If possible, the operated protection device is returned to its original closed position. This action will return service to all loads upstream from the failed element.

8. Reconnect DG units that were disconnected and are located upstream from the failed element.

9. After repair is finished, perform switching actions in order to return the system to its original stated.

10. Reconnect DG units that could not be previously put back in operation.

11. Continue simulation until the time for the next fault is reached. Time-step is increased to one hour.

If a one-phase fault occurs on a distribution line protected by a fuse, two cases can be considered:

• Voltage at load point above 0.9 p.u.: the repair will be carried out without disconnecting the failed element; loads will not experience any interruptions. • Voltage at load point below 0.9 p.u.: the procedure will follow the steps

presented above.

Third Scenario

1. System is simulated until time reaches the moment when a fault occurs, time- step is one hour.

2. The fault is simulated by placing a short-circuit at the element considered for failure.

3. Simulation time-step is reduced to allow the correct response of protection devices.

4. Simulation time-step is increased to continue with the simulation and perform switching actions; new time-step is 5 minutes.

5. Reconnect DG units disconnected during the fault but can be placed back in service immediately.

6. Failed element is disconnected.

7. If possible, the operated protection device is returned to its original closed position. This action will return service to all loads upstream from the failed element.

8. Reconnect DG units that were disconnected and are located upstream from the failed element.

Chapter 5: Reliability Assessment of Distribution Systems

121 9. Check for availability of back-up feeder.

10. If there is an available back-up feeder, load is transferred.

11. If load is transferred, reconnect DG units that were disconnected and are located downstream from the failed element.

12. After repair is finished, perform switching actions in order to return the system to its original stated.

13. Reconnect DG units that could not be previously put back in operation.

If a one-phase fault occurs on a distribution line protected by a fuse, two cases can be considered.

• Voltage at load point above 0.9 p.u.: the repair will be carried out without disconnecting the failed element and no load transfer will be performed; loads will not experience any interruptions.

• Voltage at load point below 0.9 p.u.: the procedure will follow the steps presented above.

All switching actions aimed at returning the system to its original state are performed simultaneously by the procedure; therefore all possible effects on loads and generators are neglected.

Figure 5.1 shows the simplified sequence of events and time step size for every interval when switching devices can be used to isolate the faulted device and when additional switches are available to reconfigure the system and transfer some load between feeders.

a) Sequence of events to isolate the faulted section

b) Sequence of events to reconfigure the system Figure 5.1. Sequence of events after a failure.

Fault occurs _{Protective device}

operation finishes t

t

0

t

1 Failed component is isolated

t

2

Original system configuration

t

4

Δt=0.001s Δt = 5 min Δt = 5 min

Fault occurs

Protective device operation finishes

t

Failed component

is isolated

Original system configuration

t

0

t

1

t

2

t

3

t

4 Δt = 5 min Δt = 5 min Δt = 5 min Δt=0.001s Connection to back -up feeder finishes

Analysis of Power Distribution Systems Using a Multicore Environment

122

The operation of the protective device finishes at t1 and may include some reclosing

operations. If the protective device that isolates the fault/failure is a fuse, then its replacement must be made after the fault location is isolated. The service is restored to loads upstream from the point of failure at t2. If possible, the connection of the back-up

feeder will be performed at t3. Finally, the system will return to its original state after the faulted element has been repaired/replaced at t4.

In document Analysis of power distribution systems using a multicore environment (Page 142-146)