Solar generation relies on a discontinuous power source — the sun. Day and night cycles paired with environmental factors like precipitation and cloud cover influence its reliability, so power generation from this type of renewable source is cyclical rather than continuous.
This means your transformer will not run at 100% load for 24 hours. Depending on the time of year, it may be at full load for only 6 of those hours. This brings up two questions:
- Can I reduce the size of my transformer since it is only loaded part-time?
- Can I overload the transformer during the day since it is underloaded at night?
The short answer is no. These scenarios could increase the demand put on the transformer. This article reviews the unique requirements solar applications face.
GENERATION
Let’s look at the generation side of things.
Inverters
Inverters (Figure 1) connect the solar array to the step-up transformer. Inverters convert DC-generated solar power into AC. They handle the wide swings in power supplied by the solar array and steady the voltage supplied to the step-up transformer.
Inverters do all this with special switching that regulates their power output. This switching can create power quality problems in the system that result in additional heating at the transformer.
- Harmonics. The main concern with harmonics at the transformer is overheating. Most inverters have filters to reduce harmonic distortion. With filters, inverters can keep their harmonic output below 5%. This does not account for any interaction with the transformer or other parts of the system. The total harmonic distortion across the system could be higher. You cannot always assume the inverter’s harmonic rating across the entire system.
- DC bias. Some inverters use an additional DC supply to regulate their AC output voltage. This DC component is superimposed on the AC output signal. The DC voltage cannot pass through the transformer to the grid, but it does end up in the transformer’s low-voltage winding. This can easily create overheating in the transformer core and insulation stress. This often shows up as high hydrogen gassing on a transformer DGA test.
- Output power. Some inverters can output more than their stamped nameplate rating. In such cases, transformer sizing should follow the inverter’s peak output capability — not its stamped rating. Sizing a transformer on the nameplate inverter rating alone could result in overloading at the transformer.
Other sources of increased inverter output stem from environmental factors. Solar panel output correlates with ambient temperature, and some seasons will produce more output than others when temperatures change.
A sunny day with patches of clouds can produce output power spikes. When certain types of clouds are high in the atmosphere, they amplify the brightness of the sun’s light. This is known as cloud lensing. Cloud lensing increases the output power of solar panels. The mixture of cloud shadows and cloud lensing at the panels also creates voltage spikes, which causes transformer overloading in longer durations.
Sizing for Overload
Correct transformer sizing allows for overload situations. The kVA should match the inverter’s full output characteristics. Wherever possible, consult transformer and inverter manufacturers for their input. An in-depth power quality analysis of the solar system can reveal which kVA is best. If an in-depth PQ analysis is not in the cards, size for the worst-case scenario. A dual-rated temperature rise (55/65) works well in such cases. It provides an extra 12% of kVA above the base rating.
UNIQUE FEATURES OF SOLAR TRANSFORMERS
Solar transformers feed power to the inverter and also to the grid. This is where the term bi-directional comes in. To satisfy this bi-directional requirement, a few things must be considered when designing the transformer.
Electrostatic Shielding
An electrostatic shield or e-shield (Figure 3) offers bi-directional protection.
This shield serves two purposes:
- Protecting the transformer and grid. Harmonic disruptions from inverters can pass to the utility grid. These power disruptions cause voltage spikes and impulse-like effects in the high-voltage winding. Such power disruptions can wreak havoc at the transformer and downwind on the grid. An electrostatic shield (Figure 4) between the high-voltage and low-voltage transformer windings eliminates this problem. The shield metal is thin to reduce added eddy currents. It is connected to ground at one single point (internally) in the transformer.
- Protecting the inverter. Transient overvoltage spikes on the utility side can also pass to the inverter. These overvoltage events can damage an inverter’s sensitive components. Due to its important function, the electrostatic shield should always be engaged.
HOW ELECTROSTATIC SHIELDS WORK
A grounded e-shield reduces capacitive coupling between the primary and secondary windings. Rather than coupling with the secondary winding, the primary winding couples with the e-shield. The grounded e-shield provides a low-impedance path to ground (Figure 5). Voltage disturbances such as mode noise and transient spikes are redirected away from the secondary winding. This also works from the other end of the transformer (secondary to primary).
Dissipation Factor Tests with Electrostatically Shielded Windings
Dissipation factor (Tan δ) testing is commonly performed on transformers in the field to evaluate the condition of the winding insulation. For most apparatus insulation systems under normal conditions, this value is not expected to be less than zero. However, for apparatus such as those with an e-shield, a result less than zero is possible and potentially even expected. We will take a look at how this is possible.
Dissipation factor tests use the operating principle of a two-plate capacitor. In a perfect capacitor, voltage and current are 90 degrees out of phase with one another (Figure 6).
This would also be the case in a perfect insulation system. In reality, there is no such thing as a perfect insulator. There will always be some lost energy (heat) in the dielectric medium. The dissipation factor test measures this loss to help identify the condition and characteristics of an insulation medium.
Figure 7 shows two angles: δ and θ. Both are affected by the amount of resistive current (IR) that makes up the current measured (IT) by the test set. Where no resistive current is present, the measured current (IT) would equal the capacitive current (IC), and δ would be zero. The more resistive current present, the larger the angle δ. Theoretically, the best insulation system will have the smallest loss angle at δ.
Where the measured current (IT) leads the applied voltage (Vapplied) by more than 90 degrees, the loss angle is negative. On paper, this would seem to indicate the insulation system is generating power, but this is not possible. There are several reasons this could occur. We are mainly concerned with how this may result when an electrostatic shield is present.
Figure 8 shows a capacitance measurement taken between the high and low side windings (CHL) with the ground lead guarded (UST). This should exclude any current measurements from the ground lead. Only current returning to the test set from the low side winding should be measured. However, the e-shield creates a parallel resistive path which capacitively couples with the return current being measured by the test set. This parallel resistive current path produces a reduction in the total current measured by the test set, resulting in an artificially low loss angle. Depending on the magnitude of the resistive component of the returning ground lead current, it can even result in a negative number.
Figure 9 illustrates this phenomenon. The measured current in Figure 9a shows the artificially low loss angle approaching zero degrees. The measured current in Figure 9b shows an example of a negative test result where IT leads Vapplied by more than 90 degrees.
STEP-UP AND BIDIRECTIONAL DESIGN
Renewable generation sources such as solar interact with transformers in a unique way (Figure 10). At startup, power is fed from the utility to the solar inverter. Once the inverter receives a balanced voltage input, the solar side feeds back into the grid. The transformer plays the role of a step-up and step-down unit. This is why the term “bi-directional” often appears on solar equipment.
Power flow in transformers is bi-directional by nature: Current may be fed from either winding. By itself, this term does nothing more than define a normal transformer. It means more in the context of certain applications like solar.
As mentioned already, energization happens on the utility-side winding. The low-side winding is excited after mutual induction is present between the coils. Unless the transformer is de-energized and re-energized repeatedly, inrush current is not a big issue. So bi-directional has more to do with how the transformer gets the grid and inverter to metaphorically shake hands. This is configured in the primary and secondary windings.
Winding Configuration
The utility’s distribution feed determines the high-voltage winding configuration. Utility systems are usually grounded. This means they require a wye connection with a grounded neutral point. Likewise, the inverter’s requirements determine the configuration on the low-voltage winding. Most inverters prefer a connection to a wye service with a solidly grounded neutral point. If a neutral is connected to the inverter, it is usually for voltage sensing only. This is the reason most solar transformers are configured as wye-wye. The most important thing is to match the configuration required by the inverter and grid. A wye-wye connection is not always required, but it is the most common.
Anti-Islanding
When connecting to the grid, the inverter must sense any voltage imbalance from the utility. If one phase of the utility feed is lost from a fault, the inverter must recognize this. If the inverter does not see the lost phase, it could backfeed power to the utility, causing catastrophic damage or harm. This is known as islanding. When a fault occurs on the grid, the inverter must trip and shut down. The vector grouping of the transformer plays a key role in this. A lost phase with a wye connection will yield a voltage imbalance on the other two phases. This is why most inverters favor a wye-connected service. For delta-connected services, additional ground fault detection outside the inverter may be needed. Some inverters will support a wye or delta connection. You must consult the inverter manufacturer to determine what winding configurations will allow the inverter to trip during a grid-side fault.
It is important to realize that not all inverters and grid connections are alike. A truly bi-directional transformer design (Figure 11) will always harmonize the needs of the grid and inverter.
Solar Voltages
Renewable transformers do not use standard industrial voltages. Solar array voltages include 800 V, 630 V, 600 V, 480 V, and 208 V. European inverter manufacturers commonly use 800 V; 630 V is usually found in larger solar arrays; 600 V is the most common voltage for solar inverters.
Monitoring and Gauge Alarm Contacts
Temperature and pressure monitoring is essential because of the unique loading profile of solar transformers. Early detection of overloading and overheating is the best way to prevent equipment failure and unwanted downtime.
Due to the remote nature of many renewable projects, solar transformers are often outfitted with alarm contacts on the gauges. These alarm contacts communicate with the end user allowing them to monitor the transformers remotely (Figure 12).
CONCLUSION
Integrating renewable energy sources such as solar introduces unique challenges for transformers. The cyclical nature of the source can lead to overheating, power quality issues, and overloading. It is critical to size your transformer appropriately for your solar system. It’s also important to understand the unique design of solar transformers when performing certain field acceptance tests such as dissipation factor.
Ben Gulick began working for Maddox Industrial Transformer at its South Carolina location in 2016 and is currently the Technical Sales Engineer. He received his BA in music from Indiana University before starting his career in the electrical industry with a contractor in Indianapolis, Indiana.