Electrical Distribution Systems: Practical Aspects of Fault Processing (part 1)

Home | Articles | Forum | Glossary | Books


Overview

The fault processing approach to coordinate the fault processing approaches of local, distributed, and centralized intelligence to make the best of them is put forward.

The approach to plan the amount of various FTUs in DAS to meet the requirement of service reliability is described. The test methodologies to verify the properties of fault location, isolation, and service restoration of a fault processing system based on local, distributed, or centralized intelligence are proposed.

Keywords

fault processing procedure, comparison of various fault processing methodologies, coordination of various fault processing approaches, planning the amount of terminal units, feeder terminal unit (FTU), fault indicator, relay protection, hybrid mode of FTU and fault indicator, master injection testing methodology, secondary synchronous injection testing methodology, master and secondary synchronous injection testing methodology, direct short-circuit test

1. Introduction

Each of the fault processing approaches, no matter whether it is based on local, distributed, or centralized intelligence, has its advantages and disadvantages. In practice, we should scientifically plan the fault processing system to coordinate fault processing approaches with each other to make the best of them, which will be discussed in Section 2.

The simpler a system, the more reliable it is. From another point of view, the fewer FTUs, the more economical the system is. Section 3 covers FTU amount planning to meet the requirement of service reliability.

Section 4 describes the test methodologies to verify the property of fault location, isolation and service restoration of a fault processing system based on local, distributed, or centralized intelligence, some of which lead to service interruption, the others do not.

2. Coordination of Fault Processing Approaches

Combining the fault processing approaches of local, distributed, and centralized intelligence may improve the fault processing property.

2.1 Fault Processing Performance of Various Methodologies

The whole fault processing procedure consists of six steps as shown in FIG. 1, in which clear control, fault location, fault isolation, service restoration, and return control can be carried out automatically within a short period. But repairs must be done by man-power, which may take a rather long time.

2.1.1 Clear Control

Clear control must be carried out by local intelligence. A circuit breaker, such as the main source node, is tripped by the relay protection device to cut out the fault current.

The clear control is also the start condition of centralized intelligence based fault location, isolation, and service restoration.

FIG. 1 Six steps of the fault processing procedure:

Clear control -> Fault location ->Fault isolation ->Service restoration ->Repairs ->Return control

2.1.2 Fault Location, Isolation, and Service Restoration

Fault Location, Isolation, and Service Restoration may be carried out by local, distributed, or centralized intelligence.

• Relay protection.

For rural areas, both the three-section type of over-current protection and over-cur rent protections with time delay coordination are feasible. But for urban areas, only over-current protection with time delay coordination of no more than three stages is feasible. The control of fault location, isolation, and service restoration can be able to be completed in no more than 1 sec. The fault region may not be located, however, and isolated to the smallest scope, and the restorable regions may not be fully restored.

Besides, in some cases, coordination of over-current protection is not effective, for example, the sectionalizing switches on a trunk with very short lengths are difficult to coordinate.

• Automatic reclosing control.

Automatic reclosing control is suitable for the feeders with overhead lines. Although automatic reclosing control cannot locate and isolate the fault, it may restore the service immediately in case of temporary faults. If the automatic reclosing control is only implanted on output circuit breakers in the substation, once a fault causes multistage tripping, only part of the feeder may be restored even if the reclosing is successful. But if the automatic reclosing controls are implanted on all of the action nodes with over-current protection, the loading of the feeder may all be restored even if a temporary fault causes multistage tripping.

• Backup automatic switching control.

Although backup automatic switching control cannot locate and isolate the fault, once a fault occurs on the normal power supplying route, it may switch the load to the backup power supplying route within several seconds.

• FA based on recloser and voltage-delay type sectionalizers.

This is suitable for the feeders with overhead lines in the open loop operating mode.

The trip of a main source node, that is, the recloser, is needed to clear the fault cur rent that caused the outage of the whole feeder and all sectionalizers on the feeder to trip due to loss of voltage. After a while, the main source node recloses for the first time, the sectionalizers close in sequence after being energized. For a temporary fault, the feeder will be restored within 1-2 min. For a permanent fault, the first reclosing is doomed to fail but the fault region is discovered. Then, the upstream of the feeder is restored by the reclosing of the second time, the downstream of the feeder is restored by the automatic closing of the loop switch, which causes the corresponding sectionalizers to close in sequence after being energized.

The advantages of FA based on recloser and voltage-delay type sectionalizers lies in its simplicity, reliability, and no need for communication. It can locate and isolate the fault in the area surrounded with the sectionalizers. Both upstream and down stream of the fault may be restored. But it has the shortcomings of reclosing twice and the undesired close looping may be caused once the PT of the sectionalizer implanted on the loop switch is broken. Thus, sometimes the automatic function of loop switch is locked. Besides, the restoration schemes are unchangeable even in the cases of with the risk of causing over-current and therefore the outage area may be enlarged occasionally, especially in the area with loads increasing so fast.

• Reclosing with fast over-current protection mode.

This is suitable for the feeders with overhead lines in an open loop operating mode.

The principle and performance of reclosing with fast over-current protection mode are both similar to FA based on recloser and voltage-delay type sectionalizers, but reclosing is only needed once; that is, the outage of the whole feeder only occurs once. Besides, all the sectionalizing switches should be circuit breakers while the sectionalizing switches of FA based on recloser and voltage-delay type sectionalizers are load switches.

• Fast healing approach based on neighbor communication.

This is suitable for both overhead lines and cables in both the open loop operating mode and closed loop operating mode. The fault area can be located and isolated immediately and the healthy areas have almost no influence of the fault. But high speed communication and reliability are needed. Also, the sectionalizing switches should be circuit breakers.

• DAS based on centralized intelligence.

The minimum area of fault location is the monitoring region surrounded by monitoring nodes, such as fault indicators. The minimum area of fault isolation is the action region surrounded by action nodes, which are usually the switches with FTUs.

Generally, the fault location area of DAS based on centralized intelligence can be much smaller than that of other fault processing approaches. The fault isolation of DAS based on centralized intelligence can be satisfied even in cases where other fault processing approaches are not effective, such as a fault occurs on a trunk cable with numerous sections of short length divided by sectional switches such that none of the approaches of over-current protection, FA based on recloser and voltage-delay type sectionalizers, or reclosing with fast over-current protection mode, are feasible.

Since global information can be collected, DAS based on centralized intelligence may optimize the service restoration schemes and generate correction schemes when some remote controls fail. Also, the DAS may deal with any faults, including cases of modeled fault processing, large area breakdown due to bus voltage losses, simultaneous multiple faults, the fault on a distribution grid with large-scale DGs, and so on. The remote control of a lot of switches may be arranged in methodical steps to guarantee safety.

Centralized intelligence based DAS fault processing needs a rather long time period, typically, several minutes. Also, communication systems are needed to cover the DAS master stations and the FTUs, which usually requires a large sum of money.

2.1.2.1 Return Control

The return control to normal operation mode after repair must be carried out by the centralized intelligence based DAS. Since the global information can be collected, the DAS may optimize the return control schemes and arrange the operation steps to guarantee safety. When some remote controls fail, it can generate correction schemes as well.

Of course, in case of a temporary fault, the fault processing approaches of automatic reclosing control, FA based on recloser and voltage-delay type sectionalizers, reclosing with the fast over-current protection mode and the fast healing approach based on neighbor communication may also return the distribution grid to its normal operation mode. But in the case of permanent faults, they cannot do so.

A comparison of various fault processing approaches is shown in Table 1.

Table 1 Comparisons of various fault processing approaches

2.1.3 Basic Principles of Coordination of Automatic Fault Processing Approaches

According to Table 1, we may draw the following conclusions about automatic fault processing:

• The unique measure to cut out the fault current immediately -- although the fast healing approach based on neighbor communication has the function of clear control, it is also based on relay protection, which is an essential methodology for automatic fault processing in distribution grids, but the fault location and isolation are somewhat rough.

• The unique measure to return to the normal operation mode - although FA based on recloser and voltage-delay type sectionalizers, reclosing with fast over-current protection mode, and the fast healing approach based on neighbor communication retain part of the function of return control, they are not effective in case of permanent faults, DAS based on centralized intelligence is an essential methodology for automatic fault processing of in grids.

• The methodologies based on local intelligence are faster and more reliable. Those based on distribution intelligence are simple and do not need of communication, except for the fast healing neighbor communication approach, which is fast and has little influence on healthy regions. Centralized intelligence based DAS can process complicated fault cases with more accurate fault location and optimized control steps and is very adaptable.

The basic principles of coordinating various fault processing methodologies are as follows:

• Once a fault occurs, local intelligence methodologies work first to clear the fault cur rent immediately and roughly. As for the permanent faults, the fault is roughly located and isolated immediately by relay protections. As for the temporary faults, the automatic reclosing control restore all of the service immediately. In case the loads are with more than one power supplying route, the backup switching devices are helpful to make the restoration in seconds.

• A short time later, all of the fault information has been collected by the master of DAS, based on which the fault may be accurately located and the optimized control steps may be carried out to improve the service restoration result.

• The distributed intelligence without communication based methodologies can be used as the complement to DAS based on centralized intelligence to be installed in areas where the cost of communication is too high, or be installed on the branches.

• The fast healing approach based on neighbor communication is only suitable for installation on feeders with a high reliability requirement.








FIG. 2 An example of coordination of DAS and relay protection for fault processing

With the coordination of various fault processing methodologies, not only does the performance of fault processing remarkably improve, the robustness is also greatly increased. For example, in case of the override trip or multistage tripping due to coordination failure of over-current protection, DAS based on centralized intelligence may locate the fault, isolate the fault region, and restore the service to correct the override trip or multistage tripping. On the other hand, in case where centralized intelligence based DAS fails due to a communication barrier, the fault can be roughly located and isolated by relay protection, and sometimes part of or even the whole healthy region may be restored, especially when the fault is on a lateral or branch.

An example of coordination of DAS and relay protections for fault processing is shown in FIG. 2. The distribution grid is based on cables where all switches are action nodes with FTUs. The squares indicate the circuit breakers, the circles indicate the load switches, the solid ones are in the closed states, while the hollow ones are open.

The circuit breakers in the substations, that is, S1 and S2 , are installed with the time delay instantaneous current protection and over-current protection with set value of time delay of 0.3 and 0.5 sec., respectively. The circuit breakers indicated by big squares are installed with instantaneous current protection without delay and over- current protection with set-values of time delay of 0.2 sec. The circuit breakers indicated by small squares are installed with instantaneous current protection without delay and over-current protection with set-values of time-delay of 0 sec. Thus, coordination of over-current protection can be as much as three stages while the coordination of instantaneous cur rent protection can be two stages. Also, since the set-values of time-delay of the instantaneous current protection C2 and C1 are both 0 s while their set-values of current are almost the same, override trip or multistage tripping of C2 and C1 is inevitable.

In the case where a two-phase short circuit fault occurs in region with C1 as its enter point shown in FIG. 2(b), since the short-circuit current is smaller than the set-value of the instantaneous current protection of S1 , C1 and C2 , only the over-current protections of S1 , C2 , and C1 are started. The set-value of time-delay of C1 is the shortest, it is tripped to clear the fault current and isolate the fault region while there is almost no influence on the other regions of the feeder, which is shown in FIG. 2(b).

In the case where a three-phase short circuit fault occurs in the region with C1 as its enter point shown in FIG. 2(c), since the short-circuit current is very large, the instantaneous current protections and over-current protections of S1 , C2 , and C1 are all started. As the set-values of time-delay of the instantaneous current protections of C2 and C1 are both 0 sec., C2 and C1 are all tripped almost simultaneously causing a multistage tripping, the fault current is cleared and the fault is isolated but not within the smallest area. There is almost no influence on the other regions of the feeder, which is shown in FIG. 2(c). After a while, the exact fault region is located by the centralized intelligence based DAS. Then C2 is closed by the remote control of DAS to restore the corresponding healthy regions. Therefore, the fault processing result is improved, which is shown in FIG. 2(d).

In the case where a three-phase short circuit fault occurs in the region Re(L1 , L2 ) shown in FIG. 2(e), since the short-circuit current is very large, the instantaneous current protections and over-current protections of S1 and C2 both start. Due to the set-values of time delay of C2 being shorter than S1 , only C2 is tripped to clear the fault current but the fault is not isolated within the smallest area, which is shown in FIG. 2(e). After a while, the exact fault region is located by centralized intelligence based DAS. Then L1 is opened and C2 is closed by the DAS remote control to restore the corresponding healthy regions.

Therefore, the fault processing result is much improved, which is shown in FIG. 2(f).

In the case where a three-phase short circuit fault occurs in the region Re(L3 , L4 ) shown in FIG. 2(g), S1 is tripped by instantaneous current protection to clear the fault current but causes the outage of the whole feeder, shown in FIG. 2(g). After a while, the exact fault region is located by the DAS. Then L3 and L4 are opened and S1 is closed by DAS remote control to restore the healthy regions of the feeder. Therefore, fault processing result is remarkably improved, which is shown in FIG. 2(h).

For these cases, if there was only centralized intelligence based DAS, any fault would cause the tripping of the circuit breaker in the substation and the whole feeder would endure outage during the fault processing period, which would typically last several minutes. With only relay protections, override trip or multistage tripping is inevitable. Sometimes, the fault will not be isolated within the smallest area and some of the healthy regions will not be restored. In certain cases, the whole feeder could even endure outage. Also, the return control cannot be automatically carried out without the DAS based on centralized intelligence.

To be worse, once something is wrong with the DAS or relay protection devices, the fault processing result would unsatisfactory.

It can be seen from the example that the performance of fault processing may be improved by coordination of DAS and relay protections.

3. Planning of Terminal Units

In this section, the planning of terminal units in centralized intelligence based DASs to meet the requirement of service reliability is discussed.

3.1 Elements Affecting the Reliability of Service

Reliability is the ability of the power delivery system to make sufficient voltage of satisfactory quality continuously available, to meet consumers' needs.

Equipment outages cause customer service interruptions. Power interruptions have three aspects: frequency, duration, and extent. The reasons for outage include three types: planned outage, outage due to power rationing, and outage due to faults.


Table 2 Fault ratios per kilometer of overhead lines and cables in China: Overhead lines (times/100 km/year) Cables (times/100 km/year)

Outage due to power rationing should be reduced by expanding planning and engineering of the electric power grid, usually carried out satisfactorily. Thus, the ratio of outage due to power rationing in most areas is small, for example, less than 5% in Asia.

The ratios of planned outage are quite different between utilities. These are determined by the level of condition based maintenance and non-blackout working technology, the skills and arrangement of repair, extent of traffic jam, distance, surroundings of the feeder, and so on. The ratios of planned outage of utilities vary from 30-70% in China.

With the development of distribution grids and the maintenance and non-blackout working technology, outage due to faults increases to the most important reason for outage.

The fault processing technologies described in this book are of great help to reduce the area and time of the outage due to faults.

To simplify the problem, the per-unit fault ratio is usually indicated by the times of fault on unit length of the feeder, which can be obtained by statistics. For instance, the fault ratios per kilometer of overhead lines and cables in China are shown in Table 2, which are obtained by statistical analysis of reliability data from 2002 to 2009.

The other important index is the averaged outage duration time of one fault, which consists of the time to find out the exact position of the fault and the time of repair. For instance, the averaged outage duration times due to faults in China are shown in Table 3, which are obtained by statistical analysis of reliability data from 2005 to 2009.

3.2 Cost-Benefit Analysis of Action Node Planning

For a feeder with n customers with the total load of P, it is connected to other feeders and the N - 1 criterion is satisfied. We suppose that sectionalizing switches are sufficient, the fault ratio of the feeder is F, the averaged outage duration time of one fault is T, the cost of a FTU is C, and the number of FTUs to be installed is k, which divide the feeder into k + 1 action regions.


Table 3 The averaged outage duration time of one fault in China: Averaged outage duration time of one fault (hour/times)

3.2.1 The Case of the Amount of Customers Evenly Distributed

Assuming that the amount of customers is evenly distributed, the number of customers in each action region is n/(k + 1).

Since the N - 1 criterion is satisfied, the benefit of installing k FTUs is to maintain the service of the action regions except for the fault action region before repairs being completed, which can be formulated as Equation (eq. 1).

(eq. 1)

It can be seen from (eq. 1) that the increase in benefit becomes weaker and weaker with the increase in the amount of FTUs installed.

The cost of installing k FTUs is approximately

(eq. 2, not shown)

The cost-benefit ratio of installing k FTUs is

(eq. 3, not shown)

It can be seen from (eq. 3) that BC1 (k) is a monotonically decreasing function, that is, the more FTUs installed, the lower the cost-benefit ratio is.

3.2.2 The Case of Evenly Distributed Loads

Assuming that the loads are evenly distributed, the load in each action region is P/(k + 1).

The benefit of installing k FTUs is

(eq. 4, not shown)

Where, ? is the benefit of unit load in 1 h.

It can be seen from (eq. 4) that the increasing of the benefit becomes weaker and weaker with the increasing amount of FTUs installed.

The cost-benefit ratio of installing k FTUs is

(eq. 5, not shown)

It can be seen from (eq. 5) that BC2 (k) is also a monotonically decreasing function, that is, the more FTUs installed, the lower the cost-benefit ratio is.

The net benefit of installing k FTUs is

(eq. 6, not shown)

The net cost-benefit ratio of installing k FTUs is

(eq. 7, not shown)

It can be seen from (eq. 7) that BC3 (k) is still a monotonically decreasing function, that is, the more FTUs installed, the lower the cost-benefit ratio is.

In conclusion, from the viewpoint of cost-benefit ratio, the scheme of installing only one FTU has the highest cost-benefit ratio. But the scheme of installing FTUs should also meet the requirement of service reliability. Thus, the number of FTUs installed is the minimum amount to meet the requirement of service reliability, which may have the highest possible cost-benefit ratio.

3.3 Planning the Amount of Terminal Units to Meet the Requirement of Service Reliability

Terminal units can be classified into two types; the FTUs with the function of tele control, tele-metering, and tele-indication, and the fault indicators with the function of tele-indication and inaccurate tele-metering.

To simplify the problem, we suppose that each terminal unit controls or monitors only one node. In fact, although the pole mounted FTU on overhead lines may control only one node and the fault indicator may monitor only one node, the FTU installed in a ring main unit cabinet sometimes may control more than one node, which will be discussed later.

Assuming that the amount of customers is evenly distributed and the number of terminal units to be installed is k, which divides the feeder into k + 1 regions, the number of customers in each region is n/(k + 1).

The duration time of fault processing consists of three parts shown in Equation (eq. 8)

(eq. 8, not shown)

Where, t 1 is the time of searching the fault position, t 2 is the time to isolate the fault region and restore the service of the healthy regions by manual work, and t 3 is the repair time of the failed devices corresponding to the fault.

3.3.1 Only-FTU Mode

For the only-FTU mode, all the nodes needed to equip terminal units are installed with FTUs with the function of tele-control, tele-metering, and tele-indication. Also, each node with FTU should be a switch with an electric operating mechanism. Thus, the cost of the only-FTU mode is much higher but all the operations in fault processing are automatic by remote control. Thus, service restoration of the healthy regions is fast.

For the only-FTU mode, both t 1 and t 2 are approximately zero, then we have

(eq. 9, not shown)

In case of the N - 1 criterion being satisfied and with k FTUs installed, the Average Service Availability Index (ASAI3 ) to only include the outage due to a fault can be deduced as Equation (eq. 10) according to the definition by the ASAI


(eq. 10, not shown)

Where, f i is the fault ratio of the i-th region.

Assuming that the total fault ratio of the feeder is F and each region has the same fault ratio of f, we have

(eq. 11)

Equation (eq. 10) can be re-written as


(eq. 12)

If the desired ASAI to only include the outage due to fault is no less than A, that is,

(eq. 13, not shown)

Besides, all of the loop switches should also be installed with FTUs, we assume that the amount of the FTUs installed on loop switches is k0 It can be seen from Equation (eq. 13) that the number of FTU needed to be installed depends on the desired ASAI to only include the outage due to fault, the repair time of the failed devices corresponding to the fault, and the fault ratio.

3.3.2 Only Fault Indicator Mode

For the only fault indicator mode, all of the nodes needed to equip terminal units are installed with fault indicators with the function of tele-indication and inaccurate telemetering. A fault indicator may be installed to monitor a switch with or without an electric operating mechanism; it also may be installed on a certain position of a branch or lateral without any switch. Thus, the cost of the only fault indicator mode is much lower. But all the operations in fault processing must be done by manual labor, causing the service restoration of the healthy regions to be slow. The fault location process is as fast as that of the only-FTU mode.

For the only fault indicator mode, t 1 is approximately zero, then we have

(eq. 14, not shown)

In case of the N - 1 criterion being satisfied and with k fault indicators installed, the Average Service Availability Index (ASAI2) to only include the outage due to fault can be deduced as Equation (eq. 15) according to the definition from the ASAI


(eq. 15)

In Equation (eq. 15), nFt2 is the loss of outage (hour·customer) due to service interruption of the whole feeder, which lasts the time t2 before the fault is isolated, because fault indicators lack the tele-control function.

Assuming that the total fault ratio of the feeder is F and each region has the same fault ratio of f, we have


(eq. 16)

If the desired ASAI to only include the outage due to fault is no less than A, that is,


(eq. 17)

From Equation (eq. 17), we may obtain the number of fault indicators needed be to installed as

(eq. 18, not shown)

It can be seen from Equation (eq. 18) that the number of fault indicators needed depends on the desired ASAI to only include the outage due to fault, the time to isolate the fault region, and restore the service of the healthy regions manually, the repair time for the failed devices corresponds to the fault and the fault ratio.

It is worth noting that Equation (eq. 18) is deduced in the condition of k 10, thus, once the value of 1 is less than -1, it means that no feasible solution exists.

cont. to part 2 >>

Top of Page PREV.   NEXT Similar articles HOME