AU2004269785A1 - Fault detection and location in power distribution networks - Google Patents

Description

S&F Ref: 684284

AUSTRALIA

Patents Act 1990 PROVISIONAL SPECIFICATION FOR THE INVENTION ENTITLED: Fault detection and location in power distribution networks Names and Addresses of Applicants: The Australian Gas Light Company, an Australian company, ACN 052 167 405, of AGL Centre, 111 Pacific Highway, North Sydney, New South Wales 2060, Australia Victoria University of Technology, (ABN 83 776 954 731) a body politic and corporate established by section 4 of the Victoria University of Technology Act 1990 (Vic), of 6 Ballarat Road, Footscray, Victoria, 3011, Australia Names of Inventors: Ryszard Orlowski and Akhtar Kalam This invention is best described in the following statement: 5805c FAULT DETECTION AND LOCATION IN SPOWER DISTRIBUTION NETWORKS

C.)

Field of the invention t The present invention relates to improving the performance of power distribution 00 networks.

Background Following the disaggregation of the Victorian Electricity Industry in 1994, a regulatory framework implemented a reliability improvement incentive scheme. This scheme rewards or penalizes power distributors depending upon network performance.

A financial incentive called the "S-factor" is defined by this incentive scheme, and encompasses Distribution Network performance indicators such as unplanned System Average Interruption Frequency Index (SAIFI), unplanned Customer Average Interruption Duration Index (CAIDI) and planned System Average Interruption Duration Index (SAIDI).

These performance indicators are weighted to reflect their degree of importance, as expressed by customers. The weightings for SAIFI, CAIDI and SAIDI are defined as 100%, 65% and 25% respectively. Unplanned events on distribution networks, characterized by SAIFI and CAIDI, present the biggest burden for power customers.

Therefore, these indicators are the primary targets on which distribution network owners concentrate.

Due to predominantly overhead bare wire construction, distribution feeders in urban areas are often subject to fauna, vegetation and human interference. Faults caused by urban wildlife, such as fruit bats or possums, are of a random nature and are difficult to prevent.

Preventative measures are not always successful. Re-conductoring of the overhead network with insulated cables and/or undergrounding is typically not a financially viable proposition.

684284 -2- Fault currents, depending on the fault location and the source of the fault, can reach levels of 12kA to 13kA. With high voltage industrial customers, characterised by high load consumption, the load distribution along feeders can be "lumpy". Consequently, protection discrimination on short feeders cannot always be achieved. Where protection discrimination can be achieved, the protection may not be as fast as might otherwise be desirable.

As a consequence of the traditional physical feeder characteristics, unplanned events on the networks, referred to as faults and reflected by worsened SAIFI and CAIDI, inevitably occur.

Their effects, however, expressed in terms of the number of customers affected by the feeder faults as well as the resulting duration of outages, can be minimised by implementing a uniform feeder reliability improvement strategy. To improve SAIFI and CAIDI, this strategy focuses on feeder sectionalisation, to minimise the number of customers affected by the interruptions to supply, and subsequent fault containment within the affected feeder sections. To further improve CAIDI, the healthy feeder section or sections\ located downstream of the faulty section, should be back-fed by the adjacent feeder whenever such an opportunity is presented.

The feeder sections on the network are desirably defined on a standardised load capacity basis. Under this scenario the load being transferred is known and the adjacent feeder has adequate capacity to carry the additional load. In effect, this strategy constitutes a fault location scheme heavily reliant on the fault current breaking capabilities of the feeder switching apparatus installed at the boundaries of the feeder sections.

Accordingly, improved techniques for addressing performance issues associated with power distributions networks are desirable.

684284 -3-

O

Summary

C.)

0 Protection discrimination is difficult to achieve in some urban areas due to heavy loads, a number of protection devices present on a distribution feeder, and the short distances involved. Another application is on distribution feeders where it is required to speed up t protection in general.

00 IND Reclosers are installed "in line" (that is, in series, along a distribution feeder) and are able to break fault currents. These "in line" reclosers can be used to limit the impact of a fault O 10 to a faulty feeder section, saving customers on the "healthy" feeder sections an outage, with the consequence that both SAIFI and CAIDI measures can also be reduced, due to this ability to break fault currents.

Remotely controlled and monitored distribution feeder switching apparatus is installed at the boundaries of the feeder sections including feeder paralleling points. As a result, rapid load transfers can be carried out under fault conditions, as well as during planned network reconfiguration. Remotely controlled and monitored load break switches can be successfully used on the paralleling points. Reclosers with re-configurable protection settings, augmented by reliable communications, can be used at the primary (feeder backbone) paralleling points.

A switching section affected by a fault can be promptly isolated via remote control, provided the fault location scheme correctly identifies the faulty feeder section.

Ethernet-based communications is highly desired for remote control and monitoring on reclosers. Ethernet-based communications provide a flexible platform for legacy protocols and newly emerging communications solutions alike. Implementation at field devices without access to optical fibres remains a challenge. Spread spectrum radio offers field devices access to Ethernet-based communications. As spread-spectrum radio provides limited coverage, a network of local master stations may be necessary.

684284 -4- Description of drawings 0 Fig. 1 is a schematic representation of a configuration for a power distribution feeder, which is described with reference to hypothetical Faults A, B and C.

tFig. 2 is a schematic representation of communications hardware for an upstream recloser 00 controller of a type described with reference to and depicted in Fig. 1.

Fig. 3 is a flow chart of steps that occur in detecting and locating faults in a distribution feeder having a configuration as depicted in Fig. 1.

Detailed description Two types of controllable switching apparatus can be used on distribution feeders. These are load break switches, and fault break reclosers. Both can be remotely controlled and monitored from a central host, typically a network-wide Supervisory Control and Data Acquisition (SCADA) host. Load break switches are used as the primary device for remote control and monitoring of distribution feeders. A significant disadvantage of load break switches, however, is the lack of a fault current breaking capability. Load break switches also rely on the Zone Substation circuit breaker fault clearing capabilities, thus causing all customers on the feeder to experience an outage whenever a feeder fault occurs. As a result, the metrics of SAIFI and CAIDI (as defined and described above) are affected.

Reclosers installed "in line" are able to break fault currents. Thus, "in line" reclosers can limit the impact of a fault to a faulty feeder section saving customers on the "healthy" feeder sections an outage, with the consequence that both SAIFI and CAIDI measures can also be reduced, due to this ability to break fault currents.

Remotely controlled and monitored distribution feeder switching apparatus is installed at the boundaries of the feeder sections including feeder paralleling points. As a result, rapid load transfers can be carried out under fault conditions, as well as during planned network reconfiguration. Remotely controlled and monitored load break switches can be 684284 O successfully used on the paralleling points. Reclosers with re-configurable protection settings, augmented by reliable communications, can be used at the primary (feeder O backbone) paralleling points. A switching section affected by a fault can be promptly isolated via remote control, provided the fault location scheme correctly identifies the faulty feeder section. Network re-configuration can also be attempted, if appropriate.

00 Fault location detection Fig. 1 schematically represents a fault location scheme introduced by the installation of two "in line" reclosers in the backbone of each distribution feeder. A Zone Substation X and a Zone Substation Y have respective Circuit Breakers 110, 110', and are demarcated by vertically-oriented dashed lines. Two reclosers are located at the boundaries of the switching sections, as schematically depicted in Fig. 1. A fault current is detected by the protection circuitry of the reclosers located upstream (that is, closer to the point of supply) from the faulted section. As illustrated, Recloser A 120 and Recloser B 130 are located between Circuit Breaker 110 of Zone Substation X and a normally open Switch 140. Similarly, Recloser C 150 and Recloser D 160 are located between N/O Switch 140 and Circuit Breaker 110' of Zone Substation Y.

Only the recloser that is adjacent to the faulted section is permitted to operate after a fault is detected. This scheme can be implemented when protection discrimination can be achieved between the "in line" Reclosers and/or the Zone Substation Circuit Breaker.

On feeders that are relatively short, heavily loaded and with "lumpy" load distribution, the discrimination between the reclosers' inverse time protection characteristics curves cannot always be achieved. In these cases, blocking schemes are desirably implemented to prevent the upstream Recloser and/or a Circuit Breaker from operation. Single shot reclose operations, preferred on urban distribution feeders, reinforce the requirement for fast blocking schemes.

Due to the cost involved the simplified schemes, involving blocking between the "in line" reclosers only, are more likely to be implemented.

684284 SRecloser communications Sr Reclosers operate in two modes. First, reclosers can operate in a recloser-specific mode when recloser operation is governed by the events on a feeder resulting in actions t triggered by recloser protection relays. Second, reclosers can also be made capable of 00 operating in a remote control and monitoring mode when standard network C, reconfiguration is performed. Fast and reliable communications links are desirable for

(NO

providing protection signalling for the described blocking schemes.

Consequently, there are two aspects of recloser communications, namely recloser protection signalling, characterised by response times in the order of milliseconds, and communications facilitating remote control and monitoring. Both of these communications types are of paramount importance in the area of Distribution Feeder Automation involving reclosers as a rule. However, the requirements for the performance standards for these communications types are significantly different. Due to the high speed and reliability, required by protection applications, the choice of a suitable communications medium is quite often limited to fibre optics. In urban areas, the implementation of communications solutions based on fibre optics can be difficult logistically. Also, use of fibre optics is not cost effective when implemented for recloser communications only.

On the other hand, the performance standards for communications, facilitating remote control and monitoring on reclosers, are more relaxed in comparison with those of protection signalling. Response times in the order of 1 second are quite acceptable when specific control functions are considered. Diagnostic functions, such as retrievals of event log files or oscillographic disturbance records can take up to a few minutes.

Overall, specific communications solutions can only address specific communications needs. As a result, two distinct communications approaches are adopted when considering communications with reclosers. These two approaches involve guided communications for protection signaling, and radio communications for remote control and monitoring applications, when the implementation of guided communications is not justifiable. In 684284 -7- O cases where guided communications are available, joint implementation of both protection signalling and communications for remote control and monitoring is available.

0 Modem recloser protection relays incorporate basic communications functions within the relay. This approach leaves the internal process control with the relay's Central Processing Unit (CPU). Due to the stringent processing time constraints, measured in 00 cycles referenced to the frequency of supply, the CPU desirably gives first priority to Sprocesses catering for protection functions.

Out of the three distinct CPU process categories for relays (that is, protection parameter calculations, (ii) protection signaling, and (iii) communications covering remote control and monitoring) communications has the lowest priority to avoid overburdening the CPU with non-protection tasks at protection critical times.

Communications interfacing is usually implemented via a number of serial communications ports. Protection signalling is implemented over dedicated ports.

Theoretically, this approach eliminates the need for external Remote Terminal Units (RTUs) implying cost savings. In practice, however, a communications platform is used between field devices and the operational centre to transport existing/legacy protocols.

This communications platform accommodates the introduction of newer relay types equipped to run with higher communications speeds. Network diagnostics and network re-configuration are the basic features expected from such a platform. Consequently, the installation of additional communications hardware is required in the recloser controller to fulfil the communications needs of modern remote control and monitoring. A relatively uniform approach to network communications, such as provided by the extension of Ethernet-based communications to field devices including reclosers, can provide a suitable communications platform.

Ethernet-based communications is not the only solution available for communications with reclosers. However, the use of Ethernet-based communications is one solution that has the potential to simplify existing network complexities arising from a wide variety of legacy communications protocols used by existing field devices. Due to the high 684284 O bandwidth requirement of Ethernet-based communications, fibre-optic and spread spectrum wireless mediums are suitable.

O

When fibre optic communications is not available at a recloser location, another solution available for the Ethernet-based communications is the spread spectrum radio. Practical t applications are limited to remote control and monitoring. Protection signalling, over the 00 0 spread spectrum radio, may not necessarily offer a level of reliability considered C satisfactory for protection applications. Ultra high frequency signal penetration can be

(N

inadequate in urban areas as "line-of-sight" cannot always be guaranteed. When fibre optic communications are not available the co-ordination of protection systems on reclosers can rely on current and time discrimination techniques.

Recloser remote control and monitoring benefits The primary benefit of remote monitoring and control on reclosers is network switching, whether under fault conditions or during planned network reconfiguration, commanded from the operational centre. Reliable status and alarm monitoring allows for informed decision making under any circumstance. Remotely controlled and monitored reclosers also increase the degree of knowledge about network parameters along the feeders, and can allow for proactive as well as reactive remedial actions to be taken by a Distribution Network owner.

One feature enabling proactive action is sensitive earth fault (SEF) monitoring. Increasing SEF current levels, compared with the bench-marked SEF current levels, can be met with remedial action before reaching the pre-set trip level. A typical example is when a tree grows into power lines and the resultant intermittent current leakage raises an SEF alarm.

The tree branches can be cleared well before the SEF trip level is reached, averting customer outages and reducing the risk of fire. Recloser operations and post fault network switching are examples of reactive remedial actions.

Without remote control and monitoring on reclosers the only option for the distribution network owner is to rely on time consuming reactive remedial actions. A typical example is a feeder recloser that trips and goes to lockout due to a fault that took place during the 684284 O night. Because of light feeder loading at night, the recloser trip may go unnoticed by the operational centre staff until notified by customers. This notification may come hours O after the fault occurrence.

Post-fault analyses are desirably based upon reliable remote event log file downloads.

t Retrieving event reports may use manufacturer specific protocols for each recloser relay 00 type. When the number of reclosers on the network is significant and when they are IO spread over a large area, efficient downloading, file storage and subsequent access for analysis purposes becomes a complicated and costly exercise. The issue is exacerbated if io legacy protocols are involved. For this reason the implementation of an Ethernet-based (,i communications, capable of providing a transportation platform for serial protocols and widely accepted by the communications industry, can bring benefits in terms of reduced maintenance costs.

Recloser controller with full blocking Depending on specific distribution feeder requirements, protection signalling can be implemented as a limited blocking scheme, or as a full blocking scheme. The limited blocking scheme involves blocking of the upstream recloser by the downstream recloser, or blocking of the Zone Substation Circuit Breaker by the upstream recloser. The full blocking scheme involves blocking of the upstream Recloser by the downstream Recloser, as well as blocking of the Zone Substation Circuit Breaker by the upstream Recloser, as depicted in Fig. 1.

The communications hardware and software configuration of the upstream recloser (Recloser A 120) is the most elaborate as this configuration incorporates two limited blocking schemes. "In line" reclosers are used with remote control and monitoring facilities and protection signalling over fibre optic cables.

Fig. 2 schematically represents the full blocking scheme communications hardware and the Ethernet-based communications hardware for remote control and monitoring.

Ethernet-based communications covering remote control and monitoring is implemented over a fibre optic loop comprising two fibres. This fibre optic loop begins and ends at the 684284 Zone Substation X Ethernet switch, which is part of the inter-Zone Substation Wide Area Network (WAN). The WAN facilitates communications with the operational centre.

O Along this fibre optic loop, Recloser A 120, Recloser B 140, remotely controlled switch, Recloser C 150, Recloser D 160 and Zone Substation Y 110' are connected. If need be, other devices may be connected anywhere on the loop. The return path from Zone t Substation Y 110' is physically separated from the forward path to ensure "self-healing" 00 0 of the entire loop.

Under normal circumstances, incoming data arrives at the controller of Recloser A 120 via the forward fibre connected to the receiving terminal of the optical/electrical media converter M1 260. Then, the data is passed onto the local Ethernet switch 250, where the data packets are switched according to their encoded address. Data destined for Recloser B 130, Recloser C 150, Recloser D 160 and Zone Substation Y 110' is forwarded through the transmitting terminal of M2 270. Data destined for Recloser A 120 is passed to the Internet Protocol (IP) addressable Terminal Server 240 from where the data is communicated serially to the appropriate port of the Recloser relay 220.

As depicted in Fig. 2, there are two relay ports of the Recloser relay 220 that are assigned for non-protection duties. One port is for remote control and monitoring, using a protocol such as DNP3. The other port is for event log downloads using the relay manufacturer's specific protocol.

A reverse data path is used in the event of a fibre optic cable breakdown between Zone Substation X 110 and Recloser A 120. In this case, incoming data arrives at the Recloser A 120 controller via the reverse fibre connected to the receiving terminal of the optical/electrical media converter M2 270. The data is then passed onto the local Ethernet switch 250 where the data is switched according to its encoded address. Data destined for Recloser A 120 is passed onto the Terminal server 240 from where the data is communicated serially to the appropriate relay port of the Recloser relay 220.

Two limited blocking schemes are implemented through the dedicated optical/electrical media converters M3 210 and M4 230. M3 210 facilitates circuit breaker blocking and 684284 -11- O M4 230 facilitates the downstream recloser blocking. These blocking schemes utilise Stechniques developed specifically for protection applications.

0 C Absent fibre optic capability, the communications hardware of Recloser A 120, depicted in Fig. 2, can be modified for radio remote control and monitoring needs. A two port n spread spectrum radio can be connected directly to the recloser relay 220 to provide 00 0Ethernet or other wireless capability.

Basicfault scenarios The operation of a full blocking scheme can be explained in conjunction with Fig. 1.

Three basic fault scenarios are presented by: "Fault "Fault B" and "Fault and each are considered in turn following the overview described below with reference to Fig. 3.

is Fig. 3 flow charts the steps involved in detecting and locating these different Faults A, B, and C. The scheme operates only if the configuration of the distribution feeder is normal; that is, if the N/O Switch is open, and the reclosers and circuit breakers are all closed.

Accordingly, a check is first made in step 310 of whether or not the distribution feeder is normally configured. If so, then a check is made in step 320 of whether each of the Circuit Breaker 110, Recloser A 120 and Recloser B 130 all detect a fault. If so, then Fault A is diagnosed in step 330. Section S3 is faulty, and thus Recloser B 130 sends a blocking signal to Recloser A 120 and Circuit Breaker 110. Recloser B 130 automatically opens in step 340.

If, instead, only Circuit Breaker 110 and Recloser A 120 detect a fault (and not Recloser B 130), then Fault B is diagnosed in step 360. Section S2 is thus faulty, and Recloser A 120 sends a blocking signal to Circuit Breaker 110. Recloser A 120 automatically opens in step 370.

The remaining alternative is that the Circuit Breaker 110 alone detects a fault, in which case Fault C is diagnosed in step 380, and thus section S1 is faulty. The circuit Breaker 110 opens automatically in step 390.

684284 -12- Fault A scenario O The scenario presented by Fault A illustrates that customers on two-thirds of the feeder Ccan be saved from an interruption to supply if a blocking scheme between Recloser A

(N

120, and Recloser B 130 and the Circuit Breaker 110 at Zone Substation X is Simplemented. This results in a corresponding improvement in SAIFI. Under this scenario, 00 0Recloser A 120, Recloser B 130 and the Zone Substation X Circuit Breaker 110 all detect C the fault current, but only Recloser B 130 opens, thus blocking Recloser A 120 and the

IND

Zone Substation Circuit Breaker 110 from operation.

Consequently, the Feeder Section S3 135, between Recloser B 130 and the normally open high voltage switch, is isolated. The impact of this fault is localised within this feeder section. Overall, two feeder sections are not affected by the fault and one section is permanently affected. An improvement in CAIDI can also be expected, as the location of the fault is narrowed to one Feeder Section, rather than the three illustrated Feeder Sections.

Fault B scenario The scenario of Fault B illustrates that customers on one third of a feeder, namely Feeder Section S1 115, do not experience an interruption to supply at all. This is due to the blocking scheme implemented between Recloser A 120 and the Circuit Breaker 110 at Zone Substation X.

The fault is localised to Feeder Section S2 125 after Recloser B 130 is opened via remote control. Customers on the Feeder Section S3 135 do experience an interruption to supply.

This interruption, however, is as short as the time needed to back-feed Feeder Section S3 135 via the normally open switch. The improvement in SAIFI is less significant than in the Fault A scenario, as more customers experience the interruption to supply. CAIDI performance however, is further improved as although more customers are affected by the fault, as supply is restored quickly, the CAIDI measure for the overall network is improved. In scenario Fault B, one feeder section is not affected by the fault, one section is temporarily affected, and one section is permanently affected.

684284 -13-

O

Fault C scenario The scenario presented by Fault C illustrates that customers on all three Feeder Sections s are affected by the feeder fault. This is because the fault is cleared by the Zone Substation t Circuit Breaker 110. Accordingly, this particular class of fault scenario represents 00 0 instances in which the described architecture offers no specific benefit in contrast to IND existing architectures. Once Recloser A 120 is opened via remote control, the fault is localised to Feeder Section S1 115. Following the fault containment, Feeder Sections S2 lo 125 and S3 135 are back-fed via the normally open switch. CAIDI performance is better than in the Fault B scenario as more customers have their supply restored quickly following the fault. There is no improvement in SAIFI as all feeder sections experience the interruption to supply. Overall, two feeder sections are temporarily affected and one section is permanently affected.

Conclusion An improvement in SAIFI and CAIDI measures is likely with implementation of the described fault detection and location scheme described herein, and can be considered as savings achieved in comparison with distribution feeders using load break switches in place of reclosers. With only load break switches in place, entire distribution feeders can experience an interruption to supply, due to a circuit breaker fault clearing at a Zone Substation end whenever the likely fault scenarios occurs. Various alterations and modifications can be made to the techniques and arrangements described herein, as would be apparent to one skilled in the relevant art.

Claims (2)

1. 4. The apparatus of claim 3, further comprising a communications links between 2 the central host and each of the circuit breaker and the respective reclosers. 1 5. The apparatus of claim 1, wherein the blocking scheme is a full blocking 2 scheme.
2. 6. A method for detecting and locating faults on a distribution feeder of an 2 electrical power distribution network, the method comprising: 3 determining which of a circuit breaker, and a pair of reclosers, detect a 4 fault on a distribution feeder; locating the fault depending upon which of the circuit breaker and the 6 reclosers detect, the fault; and 7 sending a blocking signal to an adjacent recloser depending upon the s location of the located fault. 684284 1 7. An electrical power distribution network comprising: 2 distribution feeders supplying power from first and second zone O 3 substations; 4 a zone substation circuit breaker for breaking electrical power supply to the switching zones of the distribution feeder; 6 a pair of reclosers located in series between the circuit breaker and a 00 7 normally open point; and 1 8 communications links amongst the circuit breaker and the respective 9 reclosers for implementing a blocking scheme. Dated 29 October, 2004 The Australian Gas Light Company Victoria University of Technology Patent Attorneys for the Applicants/Nominated Persons SPRUSON FERGUSON