Module suppliers stunned the industry last year with the introduction of modules with ratings exceeding the 500 W and even 600 W mark, made possible by utilising the new 182 mm M10 and 210 mm G12 wafer formats. This marks a significant leap from the previous year’s 350-400 W nameplate power ratings, and brings with it some electrical challenges.
Smaller cell formats that dominated the market in 2019 produce a short circuit current (Isc) of around 10-12 A. But the larger M10 format cells produce an Isc of around 13-14 A, and G12 cell-based modules can go even higher. The 500 W+ version of Trina’s Vertex modules, for example, produces short circuit currents of 13 A, while the 600 W version produces 19 A. This is nearly double what inverters are used to handling.
“PV current development has only known one way for years, and that was up,” says Michael Viotto, system architect of power conversion home and business at SMA Technology AG. “But this year the leap was a lot more significant. This is something module suppliers have also noticed and they are working on technical solutions on their side.”
With that in mind, it is imperative to look at the maximum short circuit current that inverters can take to avoid damage or danger. The product sheets of the largest capacity string inverters of market heavyweights such as SMA, Huawei, Sungrow, Ginlong, or Goodwe show that most have an Isc of 40 A, with a maximum MPPT input current (Impp) of 26 A. Not quite enough to fully accommodate all big modules without compromise. It is an issue that can and will be fixed with the next product generation update, for sure, but it is also a problem that requires some attention from project developers now.
Ginlong and GoodWe have already brought their latest string inverters to market, with an MPPT input current for each string at 15 A and an Isc of 45 A in preparation for the higher current modules. All new utility-scale string inverters lined up for release in the coming months will likely raise these levels by equivalent margins. But until that is the case, there are a few considerations to bear in mind.
“As a first step it is important to check whether the connectable short circuit currents per MPPT are within the margins. The device is only tested and certified for guaranteed safety up to that current level,” explains SMA’s Viotto.
The Impp of M10 cell-based modules, Viotto assumes, is something to the tune of 14 A and remains constant in a string, as it is connected in a series circuit. However, the current is adding up when two strings are connected to one MPP tracker – commonplace for utility-scale string inverters – leading to 28 A per MPPT. An inverter that can take up to 40 A could take two strings per MPPT in such a configuration. However, the MPPT can only handle 26 A, which means that if the modules are operating at high irradiation, which drives currents up, the inverter has to operate the PV generator slightly off the maximum power point.
“I’ve calculated that yield loss for Freiburg … Germany,” he says. “To connect 12 A rated modules to an MPPT with a maximum input current of 10 A, the mean annual yield loss would be 0.5%, for a south orientation array and no yield losses for an east-west orientation array.”
Higher currents are only achieved when the module produces at its maximum. Depending on location, this will not happen the majority of the time, so Viotto suggests that a mild MPPT input current exaggeration is not necessarily breaking the bank. However, in sunnier locations such as the MENA region, the mean annual yield loss would end up being more significant.
Central inverters or string inverters with one main string input, on the other hand, work differently and are therefore less susceptible to the problem. SMA’s Sunny Central UP, for example, has an Isc of 6400 A and the Sunny Highpower Peak 3 has an Isc of 325 A. To connect the modules, 16, 24, 32, or 36 substrings can be combined to a main string, with the central inverter allowing the connection of up to 24 main strings. In this example, one could connect 355 modules at 18 A Isc to the central inverter. It is important, however, to configure the strings so that the combined strings don’t surpass the 6,400 A.
Viotto notes that even with today’s 26/40A inverters, it is possible to use modules with a higher Isc. SMA’s inverters allow more than the MPPT input current if installers connect only one string to a two-channel MPPT. Instead of two 12 A strings, it is possible to connect one 18 A.
A more important consideration is the voltage level as it adds up when modules are connected in series, which dictates string length. If the inverter capacity is relatively small, it could prove somewhat tricky to hit the sweet spot for string voltage. Smaller modules prompt 30-40 V increments, making it more manageable. There are also modules in the 500-600 W class that have just over 40 V open-circuit voltage, though there are many at 60 V.
Rather than increasing the input and short circuit current levels, the PV industry could also opt to increase system voltage once again. In the past, system voltages of PV arrays have increased from 800 V to 1,000 V and even 1,500 V. Why stop there and not go on to 2,000 V to reduce the input current, as a higher system voltage would allow? There are first trials with such system voltages in China, as BlombergNEF energy specialist Xiaoting Wang disclosed at pv magazine’s virtual Roundtable USA event in December. But it will be a long way before 2,000 V becomes a new system standard, if at all.
Going above 1,500 V will require a revision of the IEC 62109-1 standard, which prescribes clearances and creeping distances for medium voltage systems. Such clearances and creeping distances would increase with higher voltages. It is not impossible to change standards and still build safe inverters – there should be good reasons to go down that path.
“A higher voltage allows conductive losses to go down, and to reduce the conductor diameter,” says Andreas Falk, a systems architect of energy systems at SMA. “But electricity transmission on site is DC and thus the conductive losses are Ohmic, which at 150 m distance from string to inverter should be manageable.”
Additionally, Falk says, there is no inherent system efficiency improvement due to a higher system voltage. While high voltages can reduce conduction losses, semiconductor performance drops at the same time.
Marco Jung, a professor for e-mobility and electric infrastructure at the Bonn-Rhein-Sieg University of Applied Sciences – and the department head for converter and electric drive technology at Fraunhofer IEE – says that he is concerned about higher voltages being fed into the inverters on average. “The higher the input voltage that is permanently applied, the higher the error susceptibility of the semiconductors,” explains Jung.
Already, studies on solar array reliability suggest that inverters are a component with a comparatively high failure rate. And it is not just the semiconductors that would be under more stress in a higher voltage environment. Connectors are also routinely found to be the cause of power losses and even failure. An EU study on the bankability of solar PV arrays from 2015 to 2020 concluded that cabling- and connector-related issues are among the main drivers of bankability costs. Their susceptibility to errors – particularly due to mishandling during installation – similarly increases with the voltage level.
Another aspect comes to mind when mulling to increase the string voltage from 1000V or 1500V to 2000V. Power conversion steps are more efficient if the DC/AC factor is kept as small as possible. Jung adds to this that with a higher DC input voltage, a higher AC voltage can be achieved, which would improve system efficiency. While efficiency gains on the DC side might negligible, the longer AC transmission could benefit from this.
This upsizing of PV modules has prompted a range of new questions regarding whole solar system optimisation, and more solutions and developments are to follow. For now, an alliance of module manufacturer BoS suppliers and power electronics manufacturers has formed to agree on specific system parameters universally. It looks like 15 A string current and 45 A Isc will be the first go-to solution.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: firstname.lastname@example.org.