Copyright Material IEEE  Paper No. ESW2016-03    Paper with figures and tables is available from IEEE

                                    Peter R. Walsh, P.E., Life Member, IEEE, Mersen

                                    Michael M Price, P.E., Member, IEEE, American Electric Power

Abstract – Mitigation of arc flash incident energy is important to increase safety. One method is to reduce the duration of the arc flash by using protective relays to sense an arc flash fault. This method requires a supply-side switching device that will respond to the protective relaying fast enough to mitigate the incident energy. In many existing installations, MV transformer primary fuses are already installed to provide transformer protection. Before the development of a MV controllable fuse, the MV fuse would have to be replaced with a MV circuit breaker or circuit switcher to have relay control of the MV protective device. This replacement would have significant equipment/construction costs and would require extensive equipment outage time. The MV controllable fuse method uses the existing fusegear protecting the primary side of the transformer. The protective relaying senses the arc fault and signals the controllable fuse to change to a faster acting time-current response. Incident energies can be reduced from 200 Cal/cm2 to below 8 Cal/cm2 on the transformer secondary. This paper reviews an installation using a MV controllable fuse to mitigate the incident energy on the equipment connected to the transformer secondary.

Index Terms — Controllable fuse, arc flash mitigation, NFPA 70E, IEEE 1584, arc fault sensing, arc flash protective relaying, MV arc flash mitigation, 

Since arc flash incident energy was determined to be a significant electrical safety hazard, engineers have been reducing it by new designs. Arc flash hazards are dependent on several variables, but typically the duration, distance, and the arc current are the major determinants [1]. The arc flash duration is proportional to the total incident energy released. Many new designs take advantage of this phenomenon.

Such techniques as changing relay and time-current characteristics to minimize arc flash duration are common. Application of these techniques varies from the low cost of adjusting existing equipment, to the expensive replacement of entire switchboards, motor control centers, circuit breakers, and fusegear. 

This paper will describe the installation of a new design using MV controllable fuses in a retrofit application to reduce arc flash hazards on the equipment connected to a transformer secondary. The controllable fuse can change its time-current curve response when it receives communication that an arc fault is occurring in a protected zone. It can typically reduce a hazard from 200 Cal/cm2 to below 8 Cal/cm2 with its fast response time by reducing the duration of the event [2]. Characteristically the calculated arc flash current is in excess of the fuse rating.

This application provided arc flash incident energy reduction for the auxiliary equipment at a pumped storage generation facility. This paper will examine the design decisions made including compatibility with existing equipment, projected performance, cost trade-offs, outage time, and technology risk.

A. Plant Description
AEP’s Smith Mountain Pumped Storage Plant was designed in the 1950’s and began producing power in 1965. The plant has 5 hydro-electric generators, three of which can reverse direction and become pumps. The equipment was designed and installed before arc flash was recognized as a significant electrical hazard.

A pumped storage plant generates power by taking water from an upper reservoir through its hydro electric generators to a lower reservoir during peak power demand times.  During off-peak times, the plant pumps water back into the upper reservoir. The Smith Mountain plant has the capability to generate 605 megawatts of electricity for up to 11 hours; however, it is optimally used for short periods of time to meet peak electricity demands.  Most of the time, the plant is either generating electricity or pumping water.

B. Plant Auxiliary Power System
The portion of the plant auxiliary power system supplied by fuse protected transformers is shown in Fig 1. 

Transformers #102 and 103 are both three-phase 13.8 kV to 600V, 1500kVA transformer banks protected by their respective 100A fuses.  These fuses are located in outdoor auxiliary fusegear.  The transformers are equipped with both primary and secondary CTs.

The 600V switchgear breakers 1B1, 1C-B, 1C-11, 1C1-B, & 1C1-2  are all 1600A breakers with A and C phase CTs on the bus side of the breakers and a phase B CT on the line side of the breaker (room does not permit all three CTs to be on the same side of the breaker).

During initial arc flash studies, in accordance with IEEE 1584 2002 [1] the incident energy levels on 600V switchgear buses 1C and 1C1 were found to be over 40 Cal/cm2, and main breakers 1B1, 1C11, and 1C1-2 were over 100 Cal/cm2. It is AEP’s practice to assume that all main and tie breakers that can be energized from both sides will be so energized during an arcing fault.  This configuration results in an arc (or arcs) being supplied by two sources, possibly doubling the incident energy available in the breaker cubicle.  For the purpose of showing the effect of the controllable fuses, the more common practice of considering only a single source will be used when reporting the results in this paper.

Based upon the initial studies, the AEP engineering team began to review changes in procedures and possible designs to mitigate the arc flash hazard [3].  AEP’s typical approach to reducing arc flash in a breaker cubicle is to trip the upstream device faster.  Since the up-stream devices for breakers 1B1, 1C11, and 1C1-2 are fuses, the typical approach is more challenging at Smith Mountain Plant.

Fig 1: Smith Mountain Plant Aux Power System


A. Common Methods
Arc flash detection can be accomplished several ways [2]. One method is to monitor only over current. The detector threshold is set below the expected arc fault current. Care must be taken in making this current setting so it is higher than the inrush currents caused by starting the motors downstream of the detector.

Another method has a sensor for a bright flash. The placement of the sensor is important. It is normally located in the equipment that would have the arc flash. Sometimes it can be prone to giving nuisance trips because of other light sources [4].

Some protective relaying can combine current sensing and the light sensing into one unit. The settings can be adjusted to give a trip signal only when both an overcurrent and bright flash occur at the same time. This can reduce nuisance trip signals.

A maintenance switch is sometimes used to reduce settings of the overcurrent trips. In normal operation the overcurrent settings are set to selectively coordinate within the system. When the maintenance switch is on, the instantaneous setting is lowered enough to operate quickly at arc fault current levels.  This setting must also be set high enough to make sure that it will not trip due to the normal starting of motor loads.

Zone selective interlocking can maintain selective coordination by having the upstream overcurrent protection momentarily wait for a signal from the downstream protection that it intends to open. If none of the downstream protection reports that it sees the fault, the upstream protective device trips. This wait is short enough to reduce the potential incident energy of an arc flash. Communication is required between the overcurrent protective devices.

Differential zone protection checks that the current coming into a zone equals the current leaving the zone.  If more current enters a zone than leaves the zone, the differential protection will detect a problem and trip the protective devices.  This type of protection requires CTs at all entrances and exits from the zone.  Differential protection is extremely fast, and is the preferred protection for most MV (and above) transformers and buses. Due to its near instantaneous tripping speed, it can be very effective in the reduction of arc flash incident energy.

Differential protection is suitable for complex distribution circuits, in addition to radial feeds. It is capable of disconnecting one problem feed from a bus without shutting down the entire bus.


A. LV Main Circuit Breaker
While the LV main circuit breaker is sufficient to reduce the incident energy on its down-stream bus, faults within the main breaker compartment itself must be cleared by an upstream protective device. If the upstream trip device is slow acting, there may be high incident energy levels within the main cubicle.  If the main breaker is in a switchgear assembly, the labeling and work practices become problematic due to high incident energy in a portion of the equipment. Similar problems arise when the circuits are dual fed or have redundancy as in automatic back-up power systems. 

B. Primary Transformer MV Circuit Breaker
Interrupting the arc fault current at the primary of the transformer effectively can reduce the incident energy to the entire switchboard.  The speed of the arc flash detection combined with the breaker interruption speed will determine the effectiveness. If a detection circuit or overcurrent device operation takes too long, the energy levels may be excessive.

Depending upon the age of the facility or the perceived importance of the power feed to the facility, fusegear and MV fuses were the standard MV design when MV circuit breakers were not specified.

C. Primary Transformer Fuse Protection
MV fuse protection for transformer primaries has been a standard practice for many years. They are significantly less expensive for initial purchase/installation, and require little, if any, periodic maintenance.  They are applied to protect the transformer from through faults or to minimize the consequences of an internal fault. They were not originally intended to reduce arc flash incident energy on the secondary side of the transformer.

The correct fuse size and characteristic is determined by the transformer characteristics, National Electrical Code [7], and electrical system requirements [5]. The general principle is to select current limiting fuses for short circuit operation capable of protecting the transformer from the high forces generated by a short circuit current on the secondary. The fuse has to be sized large enough to avoid fuse openings under transformer inrush conditions.

This high fuse sizing generally precludes protecting the transformer from overload damage. The secondary main is sized to protect the transformer from overload. Because the fuse is sized high, it can take much more than two seconds to operate when a secondary arcing fault occurs.

A controllable fuse can change from its normal time current curve to a much faster acting time current curve. This switch in characteristic can be triggered by arc flash detection similar to those used for MV circuit breakers. The rapid response of the faster acting time current curve will reduce incident energies on the equipment connected to the transformer secondary.

V. Controllable fuse technology

A. Basic Operation
The controllable fuse has the functions of the ordinary fuse plus additional capability. The fuse cartridge has two fuse elements and a normally closed switch. There is a control wire connector from the base that plugs into the actuator module.

Functionally, the fuse current goes through the main fuse elements and then through the normally closed contact out to the base of the fuse. In this normal state, the fuse time current curve is similar to the original fuse.

When the fuse actuator confirms that an arc fault is occurring, it sends a signal to the fuse to open its contact. The contact opening routes the current from the output of the main fuse element to the input of the small fuse element. That electrically places the small fuse element in series with the normal fuse element. When the normally closed contact opens, it changes the time current fuse characteristic.

Some providers have self-monitoring to ensure the fuse will open with the selected current curve. These methods vary between suppliers.

For example, the main fuse element of a 100 amp controllable fuse has a similar characteristic to the originally installed 100 amp fuse. At 400 amps, it would take approximately 10 to 25 seconds to clear. Refer to Fig 2.

Fig 2 Controllable Fuse Time Current Curve

When the fuse has been switched to the small fuse curve, it will clear within 0.15 seconds at 400 amps. The total clearing time will be the sum of the controllable fuse clearing time plus the arc fault detection time.

B. Controllable Fuse Modules
The controllable fuse system consists of three modules. The first module is the controllable fuse which is similar in shape to a standard MV fuse compatible with industry standard fuse clips. This contains the standard fuse and small fuse elements along with a one-use contact. See Fig 3. This is the module that is replaced after operation. The other modules are electronic and are not consumed by operation.

Fig 3 Controllable Fuse

The second module is the actuator. It is bolted on to the base of the fuse with a multi-pin cable connecting them together. Refer to Fig 4. It communicates with the fuse through the multi-pin connector and to the third module using fiber optic cables. The actuator is at the voltage potential of the fuse bottom.

Fig 4 Actuator Module

The third module is the interface between the actuator and the arc flash detection system. It uses fiber optic cables to communicate with the actuator and hard wiring to the arc flash detector. See Fig 5.

Fig 5 Interface Module

C. Electronic Operation
The basic system operation switches the fuse time characteristic to a faster one. That faster characteristic is important for the safety of the maintenance person so the electronics are self-monitored.

Two microprocessors, one in the actuator module and one in the interface module, are communicating on a regular basis self-checking for any system problems. An external command can be given to force the system into self-checking just before maintenance work is performed. The control system check is fail-safe and a positive “system healthy signal” is required.

Despite being built and tested to applicable MV fuse standards [6], this is a use of new technology. The new technology is backed up with time proven designs. Because the entire fuse current flows through the main element, the protection is at least as good as the prior fuse. If the electronic system failed for any reason, such as station power loss, the controllable fuse would still protect the transformer as did the prior fuse.

VI. Comparing alternatives: MV Circuit Breakers vs. Controllable fuses

A. Construction Complexity
The existing transformer primary protection consisted of 100E 15.5kV fuses for the 3-phase 1500 kVA transformer banks. Removing the fusegear and installing a MV circuit breaker would be a major undertaking. The footprint of a new switchgear circuit breaker would have to exactly match the existing fusegear since the incoming and outgoing 13.8kV cables are in conduit embedded in concrete.  This would require a custom switchgear design that would increase the cost of the equipment, and possibly make future maintenance more difficult.

B. Required Outage Time
A MV circuit breaker installation would require extensive downtime. The facility is cramped and placing new equipment and removing existing MV equipment would require a total shutdown of the Smith Mountain Pumped Storage Plant.  The outage duration was estimated in excess of a week.

Some controllable fuses are available to fit as replacements into standard fusegear. The 100E 15.5kV fuses used at Smith Mountain were one of the types of controllable fuses available at the time the mitigation engineering was being developed. The new controllable fuses are 3 inches shorter than the original fuses used at Smith Mountain, however this is a standard size used in the fusegear, and the fuse clips only had to be moved to a different set of pre-drilled holes on the fuse support buswork.

The outage required for the MV work using the controllable fuses was one day. The fault detection system and the LV controls were partially installed prior to the outage, during normal facility operations.

Because of the importance of the availability of the Smith Mountain Plant to the power system and the revenues that would be lost during an extended shutdown, the use of a controllable fuse was considered the only practical way to achieve arc flash mitigation goals at Smith Mountain Plant.

C. Installed Controllable Fuse
On September 7, 2014, Smith Mountain Plant went back into service after a 1-day outage with controllable fuses protecting the 102 transformer bank.  The design for this installation uses a microprocessor-based transformer differential relay that will send a trip signal to the controllable fuse should the relay detect a fault within its zone of protection.  The differential relay was connected to the CTs on the high side of the 102 transformer bank and to the CTs at breaker 1B1.  With the 1B1 breaker included in the zone of protection of the differential relay, the incident energy in the 1B1 cubicle was reduced from 65 to 6 cal/cm2. See Fig 7.  

Due to a wiring error, the differential relay tripped as the load was initially being increased on the 102 transformer.  Thanks to a minimum tripping current threshold of 100 amps in the fuse control circuit, the fuse did not trip; only the 1B1 breaker tripped.  The wiring error was found and the differential relay and controllable fuse have been operating without incident since then. As of May 15, 2016, there have been no mis-operations of the fuse or differential relay.

One feature of the original 100E fuses but not included with the controllable fuse is the interface to the fuse logic switches in the fusegear.  These switches are used to detect when a single fuse has blown.  This is not a problem, as the microprocessor-based differential relay can detect this condition and take the appropriate action; to trip the low side breaker(s). 

D. Next steps
Engineering and design to add a maintenance switch to reduce the arc flash incident energy levels on the 1C, 1C1 and the Unit 1 buses has been completed. See Fig 1.  The maintenance switch, when turned to the SAFETY position, will tell the differential relay protecting the 102 transformer bank to introduce a low set instantaneous trip based upon the current flowing through the CTs at the 1B1 breaker.  If the differential relay protecting the 102 transformer bank does not detect a fault within its zone, the FU102 controllable fuse will not be called upon to trip.  If only the instantaneous function of the relay calls for a trip, the relay will only trip the 1B1 breaker.  This will reduce the incident energy levels in breaker cubicles 1C-B and 1C1-B as well as buses 1C, 1C1, and the Unit 1 buses.

Also included in this next step will be the replacement of the fuses protecting transformer 103 with controllable fuses, and the addition of a microprocessor differential relay to the transformer 103 protection.  Both the 1C11 and 1C1-2 breakers are included in the zone of protection of the differential relay.  The differential relay will trip the controllable fuse for a fault within these cubicles.  This will result in the reduction of the incident energy in these cubicles from 51 to 6 cal/cm2, as shown in Fig 7. 

The differential relay used on transformer 103 will be a different model than what was used for transformer bank 102.  This relay will have 3 restraints, since the secondary of the transformer is a bifurcated feed and is connected to the two low voltage breakers; 1C11 and 1C1-2.  This differential relay will also be connected to the maintenance switch and will enable two separate low-set instantaneous trip settings based upon the currents flowing in the CTs at breakers 1C11 and 1C1-2.  If the differential relay protecting the 103 transformer bank does not detect a fault within its zone, the FU103 controllable fuse will not be called upon to trip.  If only the instantaneous function of the relay calls for a trip, the relay will only trip the LV breaker that has the fault current flowing through it.

A feature of the maintenance switch, as implemented at AEP, is the use of an LED lamp located at the switch location to provide feedback from the relay(s). See Fig 6. When the switch is turned to the safety position, the relay is signaled to place the low-set instantaneous setting into effect.  When this part of the programming is executed in the relay, a contact is closed at the relay and the LED at the switch is lit.  This provides positive indication to the operator that the safety system is operational.  In the Smith Mountain installation, prior to lighting the LED, the relay also sends a request to the fuse interface module to perform the self-checking function of the controllable fuse.  The LED at the switch is lit after the “system healthy signal” is received from the fuse. If the “system healthy signal” is not received within a preset time, a relay system failure alarm is presented in the plant control room.

The installation of these next steps is presently scheduled for Fall 2016.        

Fig 6 Typical Maintenance Switch Installation

Fig 7 Incident Energy Results

VII. Conclusions

  There is an additional alternative when selecting arc flash mitigation solutions. A MV controllable fuse located on a transformer primary can significantly reduce the arc flash on the equipment connected to the secondary side of the transformer. When the facility already has existing MV fusegear, there are cost and downtime savings by retrofitting instead of replacing with new MV circuit breakers.


The authors wish to acknowledge the contributions of Mark Samborsky, of American Electric Power, to the success of this project.  Mark produced the detailed electrical design, drawings, relay settings and relay programming for the installation of the differential relay system protecting transformer bank 102.  Mark continues to be a critical resource as we move forward with the next phases of this project, especially in applying the advanced functionality of the microprocessor relays to the project.


[1]      ANSI/IEEE 1584-2002. Sections 4.5 and 9.2
[2]      Tech Topics Arc Flash Note 7, Issue 1: Resources, Literature Library, Tech Topics
[3]      NFPA 70E, 2015 Standard for Electrical Safety in the Workplace, Annex O, Quincy, MA:  NFPA.
[4]      R Hughes et al, ‘‘High-Current Qualification Testing of an Arc-flash Detection System’’, White Paper Copyright  2010 TP6421-01, Schweitzer Engineering Laboratories Support, Publications, Technical Papers
[5]      Tech Topics Arc Flash Note 6, Issue 1: Resources, Literature Library, Tech Topics
[6]      ANSI/IEEE C37.41-2008 IEEE Standard Design Tests for High-Voltage (>1000 V) Fuses, Fuse and Disconnecting Cutouts, Distribution Enclosed Single-Pole Air Switches, Fuse Disconnecting Switches, and Fuse Links and Accessories Used with These Devices, Series 1, 2, 3 and Table 1
[7]      NFPA 70, 2014 National Electrical Code, Article 450, Quincy, MA:  NFPA.


Peter Walsh graduated from Worcester Polytechnic Institute (WPI) with a BS in Electrical Power Engineering and later earned an MBA from Suffolk University, Boston in 1981. He has been with Mersen, formerly called Ferraz-Shawmut and Gould-Shawmut, starting in 1998. He is currently Mersen’s Northeast Solutions Engineer for power distribution. He presents codes and standards in educational seminars relating to safe and reliable design, installation, operation, and maintenance of electrical distribution systems.

Prior to joining Mersen, Peter held the positions of Independent Electrical Consulting Engineer, and as an Electrical Engineer at such companies as GE and Cooper Industries.
Peter is a:
Member of IEEE 1584 and P1683
Member of NFPA 110, Standard for Emergency and Standby Power Systems and also NFPA 111.
Registered Professional Engineer
NFPA Certified Electrical Safety Compliance Professional

Michael Price has over 40 years of experience in the electric power field.  He received his BSEE degree from Northeastern University in 1974 and Master of Engineering – Electric Power Engineering from Rensselaer Polytechnic Institute in 1975.  He started working for American Electric Power (AEP) in 1970 as a co-op student and worked in various departments at AEP throughout his undergraduate years.  Mr. Price worked at Duke Power Company from 1975 through 1979 where he was part of a team that developed software to aid in the design of auxiliary power systems for power plants.  In 1980, Mr. Price returned to AEP where he has written and supported electrical engineering analysis software; designed major auxiliary power system additions at power plants; developed smart drawing systems for power plant design; performed switchgear assessment activities; and presently leads the team performing arc flash studies and mitigation engineering for AEP power plants.  Mr. Price is a Principal Engineer in the Electrical Engineering Systems Section at AEP and a Professional Engineer in the state of Ohio.

Peter R. Walsh, PE, ESCP, AT