Transformer protection is essential for reliable power
Transformer protection is vital to ensure that one of the most expensive and important pieces of equipment in a power system can stay in service
- Understand the purpose of transformer protection.
- Learn about the different types of transformer protection.
- Know the codes and standards pertaining to transformer protection.
Understanding the stresses and faults that can damage a transformer and how to protect against them is an important design consideration for electrical engineers. Without proper protection, one of the most important components of an electrical system, often a single point of failure, can be left vulnerable to damage.
The use of the word “transformer” refers to electromechanical power transformers, except where potential transformers or current transformers are directly discussed. Autotransformers or emerging transformer technologies, such as solid-state inverter-based transformers, present their own unique opportunities and challenges and as such, are not discussed here.
Standards and codes
All electrical designs and construction must adhere to applicable standards and codes. The following standards and codes relate to transformer protection requirements and industry standard recommendations.
- IEEE Standard C37.91-2008: IEEE Guide for Protecting Power Transformers.
- IEEE Standard C37.95-2014: IEEE Guide for Protective Relaying of Utility-Consumer Interconnections.
- IEEE Standard 242-200: IEEE Recommended Practice for Protection and Coordination of Industrial and Commercial Power Systems, Chapter 11: Transformer Protection (Buff Book).
- NFPA 70: National Electrical Code, Article 450, transformers and transformer vaults.
- Maximum overcurrent protection requirements (more than 1,000 volts): NEC Table 450.3(A).
- Maximum overcurrent protection requirements (less than 1,000 volts): NEC Table 450.3(B).
While still a valuable resource, IEEE 242 was last published in 2001 and since then, relay technologies — especially concerning differential protection schemes — have advanced greatly. IEEE has been updating and reorganizing the various color books into the 3000 series of standards with a standard related to transformer protection planned as IEEE 3004.9.However, the project authorization request for that standard was withdrawn in 2017 and it is not known if there are current plans to reinitialize the project.
The purpose of a transformer is to step up or down the voltage of an alternating current. In doing so, the current will decrease or increase inversely proportional to the voltage change. This capability allows for a more complex and efficient system. For instance, the electricity generated by a power plant can be transported at high voltages, low currents and once at a load, stepped down to a lower voltage. This strategy is more efficient because larger currents are more lossy over long distances.
A transformer changes voltage and current through magnetic inductance. Magnetic inductance passes through a medium to transfer electrical power. The medium of choice is an iron core, as iron has ideal magnetic characteristics for concentrating a magnetic field. To generate magnetic fields, primary and secondary coils are wrapped around their respective iron cores. These coils are typically made using copper or aluminum because they have good conductivity.
Transforming power through magnetic inductance is efficient, but there are still some losses, usually between 1% and 2%. These electrical losses primarily appear in the form of heat. To prevent the heat from becoming an issue, transformers have several types of cooling systems, such a soil-filled transformers and fan cooled transformers.
However, there are cases in which the heat can become too much for the transformers’ cooling system, slowly or quickly damaging the transformer. This can happen due to several reasons, including a lightning strike falling close to a transformer and causing a voltage surge or an overload of demand power forcing the transformer beyond its rated kilovolt amperes.
These will cause excessive heating, which will cause premature deterioration of equipment insulation. This deterioration is within the transformer and not easily identifiable. As the transformer deteriorates further, it will become more vulnerable to short circuit conditions. An uninterrupted short circuit can cause permanent irreparable damage to the transformer and potentially result in severe injury or loss of life. For these reasons, it is important to implement protection strategies when designing an electrical system with a transformer.
Overcurrent protection is needed to protect the transformer windings from short circuits and overloads. Overcurrent protection works differently based on the type of overcurrent protection device used. The two main types of transformer OCPDs are circuit breakers and fuses.
When using a circuit breaker, the fault is first detected within the device using the heat or magnetism created by the electric current. For larger currents or voltages, protective relays are used to sense a fault and operate the circuit breaker. Once a fault is detected, the circuit breaker will open the contacts to interrupt the circuit, based on its time-current curve, which is dependent on the settings of the breaker chosen. Circuit breaker contacts can be opened using a spring, compressed air, thermal expansion or a magnetic field.
The benefit of using a circuit breaker is the ability to reuse the device after it is tripped simply by reclosing the contacts. A disadvantage comes from the process of opening contacts. At high voltages or currents, an arc is created when interrupting the circuit.
The process of interrupting a fault with a fuse is different from that of a circuit breaker and simpler in function. The function of a fuse is to melt when too much current passes through it, thereby opening the circuit and interrupting the flow of current. The process can be fast and efficient, however, once the fuse melts down and stops the current flow, it is destroyed and must be replaced. The advantage of using fuses comes from how quickly they work, removing high fault currents before they can become a problem and preventing the creation of arcing currents when clearing a fault.
The NEC requirements for these OCPDs for transformers can be found in Article 450.3, wherein specific guidelines are given for the maximum ratings and settings of OCPDs for transformers of different ratings. These are broken down into two tables: Table 450.3(A), for transformers greater than 1,000 volts and Table 450.3(B) for transformers at 1,000 volts and less.
In Table 450.3(A), there is a primary protection column and secondary protection column. The determining factor for the primary protection is:
- What OCPD is being used (fuse or circuit breaker).
- The impedance of the transformer.
- Whether the transformer is in a supervised location.
For secondary protection, the same parameters apply, except the section is split between “more than 1,000 volts” and “1,000 volts or less.” In NEC Table 450.3(B), there is also a primary protection and secondary protection column, but the determining factors are the currents and whether both primary and secondary protection are needed. The determining currents are “currents of 9 amperes or more,” “currents less than 9 amperes,” and “currents less than 2 amperes.
An additional NEC requirement for secondary protection is that no more than six circuit breakers or sets of fuses shall be connected to a transformer in parallel. The total device ratings of these circuit breakers and/or fuses cannot exceed the allowed value of a single overcurrent device.
Another aspect of overcurrent protection to consider when designing a system is time current curves. When putting OCPDs on the transformer secondary and primary sides, the tripping parameters must be set for each device. For instance, if a fuse is on the secondary and a circuit breaker is on the primary, it may be ideal to trip the circuit breaker to prevent the fuse from blowing even though these protective devices are in-series and not required to be coordinated; however, if there are multiple secondary protective devices, it may not be ideal to trip a circuit breaker for a single line fault.
Sometimes the best solution can be a compromise in which the circuit breaker will open and reclose when a fault occurs, giving the fault a chance to resolve itself before blowing the fuse of that line. See Figure 1 for a TCC demonstrating important transformer protection parameters.
The secondary and primary OCPDs are not the only ones to consider. All OCPDs downstream of the transformer and upstream of the transformer must be considered. A TCC helps with the design of this process by having the circuit breaker and/or fuse settings graphed logarithmically. The time, in seconds, represents the Y-axis and the current, in amperes, represents the X-axis. From this, an engineer can visualize the tripping times for each device and adjust them to create a desired TCC.
Other methods of transformer protection
Multiple alternate or supplemental forms of protections beyond overcurrent protection are available for transformers, though these are generally limited to more expensive or more critical transformers. The following methods of protection, especially differential protection, may provide more rapid protection without sacrificing system security or impacting the selective coordination of the system’s overcurrent protection devices. While differential protection and temperature-based protection can be applied to both dry- and liquid-type transformers, gas- and pressure-based protection methods are restricted to liquid-type transformers only.
Differential protection — At its most simple, differential protection is the comparison of two or more currents from CTs installed around a given piece of equipment. If the currents entering the zone of protection established by the CTs doesn’t equal the currents leaving that zone of protection, it likely indicates an internal fault and the differential relay will operate. While differential protection for buses may be straightforward, differential protection for transformers requires substantially more consideration due to phase shifting, difference in current magnitudes between the primary and second, transformer inrush and other parameters.
Traditionally, such challenges would be overcome by selecting the ratios of the primary and secondary CTs such that the current magnitudes matched, wiring the CTs on a delta winding in wye and on a wye winding in delta, incorporating auxiliary CTs, harmonic restraint relays or other methods. An example of such a setup is shown in Figure 2(a).
Modern microprocessor-based transformer differential relays consider many of these challenges and handle them internally. Instead of calculating the perfect CT ratios to match the primary and secondary currents, simply input the CT ratios and transformer primary and secondary voltages into the relay. Instead of ensuring one set of CTs is wired in delta and another in wye, simply input the transformer winding characteristics. An example of such a setup is shown in Figure 2(b).
While in an idealized system, a differential protection scheme could be set to trip for any difference in current, a real–world scheme must contend with common occurrences that cause the primary and secondary currents to be proportionally static. Whenever a transformer is energized, a current spike — commonly referred to as transformer or magnetic inrush — of primarily reactive current with significant second-harmonic content occurs flows into the primary side of the transformer to establish the magnetic fields within the transformer.
While this current will vary depending on the transformers magnetic properties when it was last de-energized, it can be upward of 12 times the nominal primary full-load amperage of the transformer. Additionally, while the core losses of the transformer remain more or less the same after inrush, the copper losses of the transformer will vary in proportion to the loading of the transformer, i.e., that there is a greater expected difference in the primary and secondary current at the currents increase. Adding in other sources of error such as potential transformers or even unavoidable differences in the components themselves, one can see that a simple yes/no question of do the currents match is not sufficient.
The most common way to account for these challenges is to use a restraint current and operation current, often with a slope or curve instead of a direct proportional comparison of the primary and secondary currents. Though different methods are available, operation current is often the vector sum of the measured currents while the restraint current is often the sum of the measured current magnitudes. In such a scheme, during an external fault, the operate current will be approaching zero while the restraint current will be twice the nominal. As noted above, transformer inrush current is composed of a significant portion of second-harmonic content, as such, many differential schemes also include second-harmonic restraint or blocking to increase the sensitivity of the relay.
As with all forms of optional protection, the cost versus benefit of installing differential protection must be considered. Besides the simple costs of the differential relay current transformers, wiring and installation costs, engineers must consider the physical space requirements of six additional current transformers and where they will be mounted, especially if transformer bushings are not available. For this reason, differential protection is generally not applied on transformers less than 2 millivolt amperes in size.
Nonelectrical/mechanical fault detection systems — When liquid-cooled transformers experience a fault, gases are released from the liquid, which in turn affects the pressure of the liquid and other gases within the transformer. While these gases are often analyzed as part of regular electrical testing and maintenance and can yield an abundance of data based upon the types and percentages of gases released, the presence of gases or changes in pressure can also be used for more immediate protection.
How the gases or pressure may be detected vary depending on the physical construction of the transformer, i.e., whether it is a sealed main tank or a main tank with a conservator. The physical construction of a transformer may often decide the type of pressure- or gas-based protection that can be used as well as where the relays or sensors are located.
Gas-and pressure–based protection — The Bucholz relay is probably the most well-known form of gas-based protection for transformers, though it is often used as an alarm to notify personnel of a potential issue and not to directly trip a transformer’s upstream protection. This type of relay — at its most simple, a float/level switch — is installed to accumulate all gas rising from a transformer’s main tank into its conservator tank. Low–level faults or simple overloading will result in a slow accumulation of gas, while high–magnitude faults will cause oil to move rapidly through the relay.
Sudden pressure relays are used to sense the rapid change in liquid pressure that may be caused when an internal fault vaporizes some of the cooling liquid. As the pressure of a liquid is affected by its temperature, these types of relays are often built with multiple stages that may automatically adjust their sensitivity based on the temperature of the cooling liquid.
In recent years, gas- and pressure-based protection for liquid-cooled transformers has evolved beyond mechanical means and periodic dissolved gas analysis to real-time computer-based monitoring systems. These systems constantly analyze the gas content within the liquid of a transformer, allowing for both rapid and, possibly more importantly, targeted detection of faults. While these systems can be expensive, a client may benefit from the reduction in down time from both periodic maintenance and fault repair.
Nonelectrical temperature-based protection — As with almost all electrical components, temperature may be used to detect and identify overloading or faults internal to a transformer. For dry-type transformers, this may be through resistance temperature detectors installed between the transformer windings and enclosure. For liquid-type transformers, a simple thermometer may be installed. Both dry- and liquid-cooled transformers may have individual thermal relays. Regardless of the type installed, temperature sensors may be used to alarm and alert staff, enable another stage of cooling to run or trip the transformer’s primary protection device.
Deploying protection schemes
This case study presents a generic 13.8 kilovolt-480/277 volt 2.5/3 millivolt amperes oil-filled transformer and provides examples of how the 50/51 and 87T protection schemes described may be deployed. For this case study, we will assume the transformer is not installed in a supervised location. The following characteristics are given for the transformer:
Oil natural/air natural stage: 2.5 millivolt amperes
Oil natural/air fan stage: 3.0 millivolt amperes
Primary winding: 13.8 kilovolt delta
Secondary winding: 480/277 volt wye
Inrush current: 12 x full-load amperage, 0.1 seconds in duration
Primary full–load amperage: I_(PRI,ONAN)= S_(MVA,ONAN)/(V_LL × √3) =(2.5×10^6 VA)/(13.8 ×10^3 V × √3) =104.6 A
Primary full-load amperage, with fans: I_(PRI,ONAF)=S_(MVA,ONAF)/(V_LL × √3) = (3.0×10^6 VA)/(13.8 ×10^3 V × √3) =125.5 A
IPRI,ONAF=SMVA,ONAFVLL × 3–√ = 3.0×106 VA13.8 ×103 V × √3 =125.5 AIPRI,ONAF=SMVA,ONAFVLL × 3 = 3.0×106 VA13.8 ×103 V × √3 =125.5 A
Secondary full-load amperage: I_(SEC,ONAN)=S_(MVA,ONAN)/(V_LL × √3) = (2.5×10^6 VA)/(480 V × √3) =3007.0 A
Secondary full-load amperage, with Fans: I_(SEC,ONAF)=S_(MVA,ONAF)/(V_LL × √3) =(3.0×10^6 VA)/(480 V × √3) =3608.4 A
ISEC,ONAF=SMVA,ONAFVLL × 3–√ =3.0×106 VA480 V × √3 =3608.4 AISEC,ONAF=SMVA,ONAFVLL × 3 =3.0×106 VA480 V × √3 =3608.4 A
Maximum let–through current = I_SC=I_(SEC,ONAN)/(%Z)= (3007.0 A)/0.05 =60140.0 A
Inrush current = I_INRUSH=I_(PRI,ONAN)×12=104.6A*12=1255.2 A
NEC Table 450.3(A) permits a transformer of less than 6% impedance and primary voltage over 1,000 volts to be protected by a circuit breaker rated no more than 600% of the rated current. In this case, 600% of the primary full-load amperage equates to 627.6 amperes. As this value does not correspond to a standard rating or setting, the next higher standard rating is permitted. It is not required nor necessarily beneficial for the circuit breaker to be rated as high as possible. While transformers can handle significant overloads for short periods of time, such situations will lead to a decrease in the transformer’s expected life, even if immediate damage does not occur.
A 600–ampere rating will be selected because most manufacturer’s medium-voltage circuit breakers start at this rating; however, the CT ratio will be selected to match the FLA more closely — in this case, 200:5 should be acceptable. When selecting CTs, engineers should consider the fault currents the CT may be exposed to and the burden of the circuit that it feeds to confirm the CT will perform within an acceptable range of errors.
NEC Table 450.3(A) also provides guidance on the secondary protection of the transformer. For a transformer with primary voltage 1,000 volts or less, a circuit breaker or fuse is permitted to be rated no more than 125% of the rated current. In this case, 125% of the secondary FLA equates to 3758.8 amperes. As this value does not correspond to a standard rating or setting, the next higher standard rating is permitted. For most manufacturers, this would be 4,000 amperes. A CT ratio of 4,000:5 should be acceptable.
The transformer overload protection (50) is normally set over the FLA rating of the transformer, generally in the range of 110% to 125% of the FLA. Higher pickups, provided they do not violate the NEC requirements, may also be acceptable, as transformers may sustain some overload without significant damage. The inclusion of a second or third stage of cooling also allows a higher overload pickup current, though these may appear in TCCs as not providing full transformer protection. For this example, we will assume a pickup of 110% of the second-stage FLA, which equates to 138.1 amperes on the primary and 3969.2 amperes on the secondary.
The transformer instantaneous protection on the primary should be set to ensure that it will not pick up upon transformer energization, calculated to be 1,255.2 amperes. A setting of 1,350 amperes should provide adequate buffer for the inrush current and rapid tripping for arc flash or coordination purposes. The relay curve and time delay should be adjusted to ensure it is above the inrush point and that it provides adequate transformer protection and coordination with other overcurrent protective devices.
An instantaneous pickup may not be necessary if the relay curve and time delay are set as described. The differential protection will provide the quickest isolation of the transformer during arc flash events; if the bus being fed by the transformer has differential protection or other protections/procedures to mitigate the high incident energy present at the lineside of the main device, eliminating the instantaneous pickup setting of the primary and second overcurrent protection could ease coordination with downstream overcurrent protective devices. The protection of the conductors and downstream electrical equipment should also be considered when selecting protective device settings.
For this example, the conductors are sized greater than the transformer ratings and the downstream overcurrent equipment selectively coordinates with these settings.
To ensure the CTs used for the more sensitive differential elements do not saturate, the CT ratios for the differential protection scheme are often chosen to be greater than that of the main CTs. It is often recommended that these CTs also have an accuracy rating of C200 or higher. For this example, we will assume that it was confirmed through CT burden calculations that the differential CT ratios could match the main/overcurrent CT ratios. We will also assume the differential relay accounts for the delta-wye polarity shift and the difference in the CT ratios. Different manufacturers implement differential protection schemes in unique ways and the actual relay’s instruction manual should always be referenced when determining settings.