Causes of equipment failure
Each piece of electrical equipment on a distribution system has a probability of failing. When first installed, a piece of equipment can fail due to poor manufacturing, damage during shipping, or improper installation. Healthy equipment can fail due to extreme currents, extreme voltages, mischievous animals, severe weather, and many other causes.
Sometimes equipment will fail spontaneously for reasons such as chronological age, thermal age, state of chemical decomposition, state of contamination, and state of mechanical wear. The following paragraphs in this technical article present the most common modes of failure for equipment that is most critical to distribution system reliability.
- Distribution Transformers
- Underground Cables
- Overhead Lines
- Circuit Breakers
- Surge Arresters
- Insulators and Bushings
Transformers impact distribution system reliability in two related ways: failures and overloads. Catastrophic transformer failures can result in interruptions to thousands of customers. When this happens, other transformers are often called upon to pick up the interrupted load.
If there is not enough spare transformer capacity, a decision must be made whether or not to overload in-service transformers and accept the resulting loss of life. Accepting loss-of-life will improve reliability for the moment, but will increase the probability that the overloaded transformers will fail at a future date.
A summary of transformer design temperatures is shown in Table 1. Older 55°C rise transformers are shown to have a hot spot design temperature of 95°C and newer 65°C rise transformers are shown to have a hot spot design temperature of 1 10°C.
The life of a transformer is often defined as the time required for the mechanical strength of the insulation material to lose 50% of its mechanical strength (many other definitions are also possible). Loss of mechanical strength occurs when insulation polymers break down due to heat.
Figure 1 – Transformer failure
The rate of breakdown increases exponentially with temperature, allowing the expected life of insulation to be expressed by the Arrhenius theory of electrolytic dissociation:
insulation life = 10(K1/(273+°C))+K2 hours
Constants for this equation have been experimentally determined for both power transformers and distribution transformers and are documented in standard transformer loading guides. A summary of these values is shown in Table 2. This table also shows the required hot spot temperature rise above normal limits that will cause the rate of thermal aging to double.
Table 1 – Temperatures used for transformer ratings
|Temperature||55°C Rise Insulation||65°C Rise Insulation|
|Average Winding Temperature Rise||+55°C||+65°C|
|Average Winding Temperature||85°C||95°C|
|Additional Hot Spot Temperature Rise||+10°C||+15°C|
|Hot Spot Temperature||95°C||110°C|
Many older transformers have 55°C rise insulation, but most new transformers have 65°C rise insulation. Higher insulation ratings allow transformers to operate at a higher temperature and, therefore, serve higher loads.
Table 2 – Transformer aging constants for insulation equation above, determined by accelerated aging tests that measure the time required for transformer insulation to lose 50% of its initial strength.
|Description||K1||K2||Rise that Doubles Aging Rate|
|Power Transformer (55°C Insulation)||6972.15||-14.133||95°C + 5.9°C|
|Power Transformer (65°C Insulation)||6972.15||-13.391||110°C + 6.4°C|
|Distribution Transformer (55°C Insulation)||6328.80||-11.968||95°C + 6.6°C|
|Distribution Transformer (65°C Insulation)||6328.80||-11.269||110°C + 7.1°C|
Plots of life versus hot spot temperature are shown in Figure 1. These curves can be used to determine the expected life of a transformer and to estimate the loss of life that will occur during an overload. If run constantly at hot spot design temperatures (95°C for 55°C insulation and 110°C for 65°C), power transformers have an expected insulation half-life of about 7.2 years and distribution transformers have an expected insulation half-life of about 20 years.
Transformers are not normally loaded constantly at their nameplate rating and are re-rated by utilities based on weekly load curves to result in an acceptable lifetime (e.g., 30 years).
Loss-of-life will occur if the temperature is allowed to rise above normal ratings, and many utilities will accept a certain amount of loss-of-life during emergency situations.
Figure 2 – Expected insulation half-life of transformers as a function of hot spot temperature.
Typical transformers are designed for hot spot temperatures of 95°C (55°C rise insulation) or 110°C (65°C rise insulation). Transformer life reduces exponentially with hot spot temperature.
Extreme overloads can result in catastrophic transformer failure. The temperature of the top oil should never exceed 100°C for power transformers with 55°C insulation or 110°C for those with 65°C insulation. The consequence of exceeding these limits could be oil overflow, excessive pressure, or tank rupture.
If winding hot spot temperatures exceed 140°C when moisture is present, free bubbles may form and result in internal faults. Due to these considerations, the peak short-duration loading of power transformers less than 100 MVA should never exceed 200% of the nameplate rating.
The resulting fault current passes through the transformer and shakes the windings with a mechanical force proportional to the square of the fault current magnitude. If thermal aging has caused the insulation to become sufficiently brittle, a crack will form and an internal transformer fault will result.
Low impedance transformers tend to fail more often than high impedance transformers since they will experience more severe fault currents. Autotransformers generally have very low impedances and tend to fail more often than multiple-winding transformers.
Recommended Reading – My worst experience in testing and commissioning power transformers (and how I fixed things up)
Extensive research has been done in the area of transformer condition assessment. The goal is to determine the health of the transformer and identify incipient problems before they lead to catastrophic failure. Simple methods include visual and audio inspection. More complex methods include thermal load tests, power factor tests, high potential tests and dissolved gas analysis, and other electrical, mechanical, and chemical techniques.
Some of these tests can be performed continuously, referred to as condition monitoring, and can automatically notify operators if monitored values exceed warning thresholds.
Transformers have many accessories that can also fail. Failures associated with pumps, fans, and blocked radiators reduce the ability of transformers to dissipate heat but do not cause outages directly. Other failures, like cracked insulators and broken seals, may result in outages but can be fixed quickly and inexpensively. Oil-filled load tap changers have historically been prone to failure and can substantially reduce the reliability of a transformer.
Manufacturers have addressed this problem and new devices using vacuum technology have succeeded in reducing load tap changer failure rates.
Recommended Reading – The mystery of nuisance tripping incidents in transformer protection
A major reliability concern pertaining to underground cables is electrochemical and water treeing. Treeing occurs when moisture penetration in the presence of an electric field reduces the dielectric strength of cable insulation. When moisture invades extruded dielectrics such as cross-linked polyethylene (XLPE) or ethylene-propylene rubber (EPR), breakdown patterns resembling a tree reduce the voltage withstand capability of the cable.
When insulation strength is degraded sufficiently, voltage transients caused by lightning or switching can result in dielectric breakdown.
Figure 3 – 33kV underground cable insulation breakdown
The severity of treeing is strongly correlated with thermal age since moisture absorption occurs more rapidly at high temperatures. Water treeing has been a widespread and costly problem for utilities with aging XLPE cable. To address utility concerns, cable manufacturers have developed both jacketed cable and tree retardant cable (TR-XLPE, see Figure 4).
Cable jackets protect the insulation from moisture ingress and protect concentric neutral conductors from corrosion. Tree retardant insulation slows the development of water trees after moisture is present. Utilities can also install surge protection devices on riser poles to limit the magnitude of voltage transients seen by old cables.
Figure 4 – Insulation breakdown versus age for 35-kV cable at Houston Lighting and Power
As you can see from the above Figure 4, tree retardant cable retains much more strength than EPR or XLPE cable. Actual results are shown up to 10 years and are extrapolated to 100 years.
Treeing is largely attributed to bad manufacturing. Insulation impurities and insulation voids accelerate moisture absorption and can substantially reduce the life expectancy of cable. To minimize the probability of installing inferior cable, utilities are encouraged to test all reels before acceptance.
Some common cable testing methods include:
Cable Test #1 – Steady-State Voltage Withstand Test: a destructive test that applies a constant AC or DC voltage (about 3 times nominal) to see if a dielectric breakdown occurs.
Cable Test #2 – Impulse Voltage Withstand Test: — a destructive test that applies a voltage transient about equal to the BIL rating of the cable to see if a dielectric breakdown occurs.
Cable Test #3 – Partial Discharge Test: a small high-frequency signal is injected at one end of the cable. Signal reflections detected by sensors indicate the number and location of partial discharge points.
Cable Test #4 – Power Factor Test: the power factor of cable impedance is measured and compared to cables in known states of deterioration.
Cable Test #5 – Dielectric Spectroscopy: measuring power factor over a range of frequencies (e.g., 0.1-Hz to 20-kHz). Graphs of power factor versus frequency allow cable samples to be sorted by relative health.
Cable Test #6 – Degree of Polymerization (DP): analyzing physical insulation samples in a lab for polymeric breakdown. The average polymer length, referred to as the degree of polymerization, is directly related to the strength of the insulation. DP testing is suitable for celluloid insulation such as oil-impregnated paper but is less applicable for extruded dielectrics.
Cable Test #7 – Indentor Test: using special cable indentors to test the hardness of the cable insulation.
Sometimes this inconvenience is so great that utilities forgo testing and install new cable whenever an old cable is de-energized. An alternative to cable replacement is cable rejuvenation. In this process, special elbows are placed on the ends of cable sections. A special fluid is injected into one end and allowed to slowly migrate through the cable over a period of weeks, filling voids and restoring dielectric strength.
Cable rejuvenation can add decades of life to old cable and maybe a cost-effective solution in areas where cable removal and replacement are expensive. For most utilities, a majority of cable system failures occur at splices, terminations, and joints rather than at the cable itself.
In addition to water treeing, cable accessory failures are caused by incipient mechanical stresses due to water reacting with aluminum and liberating hydrogen gas. Water ingress in cable accessories is largely due to poor workmanship and can be mitigated by proper training and the use of waterproof heat shrink covers.
Due to high exposure, most overhead line damage is caused by external factors such as vegetation, animals, and severe weather. The bare conductor is able to withstand much higher temperatures than insulated conductors and damage due to high currents is less of a concern.
Regardless, high currents do impact the reliability of overhead lines in several ways. High currents will cause lines to sag, reducing ground clearance and increasing the probability of phase conductors swinging into contact. Higher currents can cause conductors to anneal, reducing tensile strength and increasing the probability of a break occurring.
Fault currents, if not cleared fast enough, can cause conductors to fuse and burn down.
Due to thermal inertia, the conductor sag will not occur instantaneously. Typical lines have thermal time constants between 5 and 20 minutes, allowing temporary overloading without sag concerns. If not cleared fast enough, short circuit currents can cause lines to melt and fall to the ground. The maximum short circuit current depends upon many factors including clearing time, conductor resistance, thermal capacity, initial temperature, and fusing temperature.
Conductor damage curves for various sizes of aluminum and copper wire are shown in Figure 5 and Figure 6, respectively.
Figure 5 – Damage curves for bare aluminum wire with a pre-fault temperature of 90°C. The steel core of an ASCR conductor will be damaged much later than the aluminum and may prevent conductor burndown.
Figure 6 – Damage curves for hard drawn bare copper wire with a pre-fault temperature of 90°C. If the fault is not cleared in time, the wire will burndown.
Many reliability problems on overhead systems are associated with auxiliary components rather than the actual wires. This includes energized equipment such as hot clamps, splices, switches, cutouts, arresters, capacitor banks, and voltage regulators. It also includes non-energized equipment such as poles and cross-arms.
Potential reliability problems related to these devices can be identified by various inspection methods including:
Inspection Method #1 – Visual Inspection: Feeders can be visually inspected by following their routes on foot, bicycle, car, or truck. Doing so can identify advanced stages of damage to cross-arms, hardware, and conductors.
Inspection Method #2 – Radio Frequency Interference: Damaged overhead equipment such as cracked insulators, loose connections, or broken conductor strands can emit radio frequency noise. After a radio receiver picks up noise, further investigations can be made to identify the source.
Inspection Method #3 – Infrared Inspection: Nonferrous line hardware in good condition will operate cooler than the conductor16. Unreliable or loose connections tend to operate hotter and can be identified by infrared inspection equipment.
Inspection Method #4 – Aerial Inspection: Visual and infrared inspections can be effectively performed from a helicopter or small airplane. This allows many miles of circuit inspection to be performed quickly, facilitates the inspection of inaccessible areas, and can identify problems that will not be noticed from a ground inspection.
Inspection Method #5 – Switch and Cutout Testing: Switches that have not been recently operated are subject to many reliability problems. They could be rusted shut, fused shut, frozen shut, or have warped or broken components.
These problems will not result in a failure but will prevent the switch from being used after a contingency occurs. Operational testing can prevent these situations, but is costly and may result in customer interruptions.
Inspection Method #6 – Fuse Testing: Distribution protection systems are coordinated based on specific fuse sizes. After a fuse blows, crews may replace it with a different size and cause coordination problems.
Inspection Method #7 – Wood Pole Testing: Wood poles lose strength as they age. Most problems typically occur near ground level where trapped moisture causes rotting. Additional damage can occur due to singular events such as automobile collisions and fires.
A simple way to test a wood pole is to strike it near the base with a hammer and listen for suspicious hollow sounds. More sophisticated methods include core sampling, sonograms, indentation testing, and electrical resistance testing.
Low impedance faults typically have contact resistances of less than 2Ω. If the wire falls to the ground without contacting or arcing to another conductor, the contact impedance will limit current to low levels and a “high impedance” fault occurs.
Tests have shown that faults to surfaces such as asphalt, grass, and gravel are typically less than 50 amperes while faults to reinforced concrete can approach 200 A.
Circuit breakers are complicated devices that can fail in many different ways. They can spontaneously fail due to an internal fault, spontaneously open when they should not, fail to open when they should, fail to close when they should, and so forth.
The table below lists the most common circuit breaker failure modes and their relative frequencies of occurrence.
Table 4 – Typical failure modes of circuit breakers
|Failure Mode||% of Failures|
|Opened when it should not||42%|
|Failed while in service (not opening or closing)||32%|
|Failed while opening||9%|
|Damaged while successfully opening||7%|
|Failed to close when it should||5%|
|Damaged while closing||2%|
|Failed during testing or maintenance||1%|
|Damage discovered during testing or maintenance||1%|
As it can be seen from Table 4, the most common failures occur when circuit breakers open when they should not (false tripping).
The next most common failures are due to spontaneous internal faults. A circuit breaker opening when it should not is referred to as false tripping. False tripping is typically associated with miscoordinated protection devices or problems with relays and associated equipment. False trips can be reduced by testing protection device coordination, relay settings, CT/PT ratios, and control wiring.
Circuit breakers can fail to open or close due to faulty control wiring, uncharged actuators, or simply being stuck. The probability of these types of operational failures occurring can be reduced through periodic exercising and testing all circuit breakers.
In addition to insulation aging, circuit breakers are subject to contact erosion. Each time the contacts are used to interrupt current, a small amount of contact material is vaporized.
Continuous monitoring can be used to estimate circuit breaker condition including the cumulative loss of contact material. Typical monitored values are:
- Opening time,
- Closing time,
- Contact speed,
- Contact bounce, and
- Recharge time.
After excessive values or negative trends are identified, circuit breaker maintenance or replacement can be performed before failures occur.
Recommended Reading – How to perform diagnostic testing of HV circuit breakers
Surge arresters come in two basic forms: silicon carbide and metal oxide varistors (MOVs). Silicon carbide is the older of the two technologies and requires an air gap to avoid excessive currents during normal operation. When the voltage exceeds a certain threshold, the air gap arcs over, and voltage is clamped across the arrester.
MOVs have a highly nonlinear resistance (as a function of voltage), do not conduct excessive currents during normal operation, and generally do not require an air gap. When the voltage exceeds a certain threshold, the resistance of the MOV drops sharply and clamps voltage across the device (similar to a Zener diode).
Further, thermal expansion of water vapor under heat can result in damaging mechanical stress to the interior of the arrester and can result in a catastrophic failure during normal or overvoltage situations.
Other failure modes include bad or aged blocks and direct lightning strikes.
Figure 7 – Blown porcelain surge arrester and damaged bushings on power transformer
There are four other major failure modes associated with surge arresters:
- Thermal runaway,
- Cracking under tension, and
- Cracking under compression.
Cracking and puncture are caused by a localization of the current, which causes concentrated heating leading to nonuniform thermal expansion and thermal stresses. Puncture is most likely in varistor disks with a low geometrical aspect ratio when the current density has intermediate values.
For low and very high current densities, the most likely failure mode is thermal runaway—the surge arrester simply is not able to handle the energy levels flowing through it.
Insulators and bushings are made from three basic materials: glass, porcelain and polymeric. Glass and porcelain are the oldest technologies, but polymeric materials are gaining popularity due to their increased strength and reduced brittleness.
Insulator and bushing failures are associated with dielectric breakdown. A dielectric breakdown in an insulator allows current to arc across the device. A dielectric breakdown in a bushing allows current to arc from the internal conductor to the outside of the device. Sometimes these currents will be small or self-extinguishing.
At other times these currents lead to a low impedance arc, resulting in a short circuit and a potentially catastrophic failure of the insulator.
Figures 7 and 8 show the degradation of dielectric strength with salt contamination for various types of insulator designs and weather shed materials. Voltage withstand capability drops dramatically with contamination but can vary substantially for different insulator designs. This can correspond to a much higher probability of flashover for certain designs.
Figure 7 – Decrease in voltage withstand with increasing contamination for two types of polymer insulator design
Figure 8 – Laboratory test results for different weather shed material types under fully wetted conditions. Contamination is measured in equivalent salt deposit density (ESDD).
Other important factors include uniformity of contamination, ratio of soluble to nonsoluble contaminants, wetting agents, altitude, and mounting position.
Regular washing of insulators will reduce the probability of flashover by keeping contamination density low and, consequently, dielectric strength high. The frequency of washing will depend on a variety of factors including the accumulation rate of contamination, the impact of contamination on particular insulator designs, cost, and overall system impact.
Source: Electric Power Distribution Reliability by Richard E. Brown