Robin Vincent

Inverter loss over nominal power usually represents the energy lost due to the inverter max AC power. It can also include the energy lost due to the grid maximum power if the option "Account as separate loss" in the power limitation window is not selected. Having optimizers in your system should have a very limited impact on the inverter clipping losses, as long as they are not limiting the max string power. With SolarEdge, this will mostly depends on the number of optimizers per string. In the example below, the max string power cannot be reached. The string operates at constant voltage (850V), but the max optimizer current is 15A so the max output power is limited to 12.75kW. In this situation, the clipping that used to be necessary at inverter level will occur at optimizer level. You can check this by looking at OP_Imax or OP_Pmax values, respectively, the clipping loss due to the max optimizer current and max optimizer power. NB: There is a known issue in PVsyst 8.0.15 to 8.0.21 that overestimates the current clipping losses. If you are in this situation, please use PVsyst 8.1.0 or later. In the case of the inverter, the percentages are expressed based on EArray, not Egrid. Could that explain the discrepancies you are seeing ?

PVsyst will use the value specified as the inverter AC power, in this case 33.3kW. If a difference of 0.1% on the total AC power is a problem, you could copy the inverter, change the max AC power to 33.4kW, and replace 1/3 of your inverters with this new one. As long as the inverter name is unchanged, it will be usable with the same optimizers.

Hi, I've contacted Huawei, I will let you know of the answer.

Dear Nikhil Lokhande Thanks for your feedback! As you correctly analyzed, the battery is currently shared across the whole system, between the final MV transformer and the injection point. You could find some very far-fetched workaround (splitting your system into two variants and upscaling your inverters so as to not clip and use a variable grid limitation to reintroduce the clipping after the battery), but I'm not sure about the reliability of the results you may obtain that way. We have planned for more options in terms of battery connection, but it won't be before the next major version at the very least. Secondary Reporting Observations: SLD Visualization: The single loss diagram is currently under heavy rework and should be available in the next major version. System Overview: The chargers (Dynapower DPS-500) are omitted from the primary System Overview tab, which focuses only on the grid-feeding inverters. As mentioned, the battery can only be set on the AC side for the moment, so it is not possible to simulate a system with DC-DC converters. Questions for the Community/Developers: Is there a workaround to "un-assign" a specific sub-array from the DC-coupled storage bus within the same variant? : Not for the moment (DC coupling is not possible either) If we must run two separate projects and combine the hourly data in Excel to respect the 2.49 MW POI limit, how can we best account for the shared "Grid Power Limitation" within PVsyst’s energy management logic? When applying the grid limitation to individual inverters during the simulation, PVsyst is dispatching the total value based on the estimated AC power of each inverter. You could reproduce a similar strategy if you split your system into two variants. run the two variants without grid limitation Compute the relative production of each system Compute the grid limit for each system at each time step based on the total grid limit and the relative production Use the possibility introduced with PVsyst 8.1 to import a time series for the grid limitation Run the simulation again with the grid limitation activated. You can also create alternative strategies to reflect your system behavior if needed (e.g : prioritize one side of the system for example).

Inverter/optimizer combination rules are directly sent by Huawei to PVsyst. Huawei's latest update indicates that this specific inverter can only connect to a total of two strings. As it seems different from what the data sheet in your screenshot suggests, I'd recommend that you contact Huawei to report the issue, and ask them to update their PVsyst database.

With Peak shaving, the strategy is relatively simple: charge the battery with excess energy and discharge the battery into the grid up to the grid limitation value. You could set up your system to start discharge during the night (i.e with PV zero production). In this case, you will discharge at the max inverter power (20MW) until the battery is empty. To ensure your battery is fully charged each day, you can use the power shifting strategy instead Using the "Fixed charging power before exporting" will prioritize battery charging before grid injection. This won't guarantee a full 40MWh/day of charging if the PV production is lower than that, but you should be closer to your target. This will probably lower your overall PV production though, since the battery will be more likely to be fully charged when your PV production reaches the grid limitation.

With the Peak shaving strategy, you will indeed put any excess production into the battery with respect to the charger max power and battery state of charge. In these conditions, you can charge your battery up to 20MW, but only if the total AC power is 50+20MW. Your system will always output less than 70MW (max clear sky production being estimated to 68.23MW), but will also charge for more than 2 hours a day, so your daily charging energy will be somewhere between 0 and 40MWh/day. The actual energy will depend on the weather and battery state of charge at the beginning of the day, so you cannot predict the charging/discharging energy based on the max charging power.

From your table, I understand you have in one case 2*2 MPPT with 2 strings each, and on the other one only 2 MPPT with 2 strings each, is that correct ? You may or may not see any difference in your results depending on how you've set your simulation parameters. Having two strings per MPPT may result in a lower production when : The combined strings power or current exceed the MPPT limits. The voltage mismatch value in the detailed losses in not null. Please also note that the maximum MPPT power depends on the power sharing option. If you have unbalanced MPPT input power, you can redistribute the total inverter power accordingly. This will let the MPPT connected to two strings to have a higher power limit, and strongly reduce the difference between your two configurations

There is a known issue where the total AC power in the report is not accurate with SolarEdge subarrays. This is due to the way the total is computed in the report, based on individual MPPT powers. With SolarEdge and its specific way to attribute (and distribute) strings to inverters, the computation is sometimes incorrect. This issue is on our roadmap, but as it does not impact the simulation results, it has been scheduled for a later version.

In both versions an issue may happen for systems using the power shifting strategy with transformers losses. The issue happens when most of the PV production is diverted to the battery or most of the grid injection comes from the battery. In these cases the AC losses computation may lead to unrealistic values. This issue will be corrected in the next release. In the meantime, users are strongly advised to use PVsyst 8.0.17 for projects using power shifting.

Optimizers won't necessarily increase the system production. Optimizers are especially useful when a part of the string is partially shaded, and has a different Impp (current at max power point) from the rest. Without optimizers, the whole string have to use the current of the worst module, leading to a lower overall production. With optimizers, the shaded and unshaded modules can operate at their own optimum current, the optimizers ensuring the output current to be the same. But when you don't have partial shadings, optimizers cannot improve anything. And since optimizers have their own efficiency, they end up increasing the system losses, leading to a lower PR.

For SolarEdge system there is two different advanced parameters limiting the Pnom ratio, one for inverters with less than 25kW nominal power, on for more than 25kW. Since the 50kUS is set to 16.7kW (1/3 of the physical inverter), the <25kW parameter. Please check these parameters in the advanced parameter menu (Settings => edit advanced parameters) You should be able to simulate your system if you change the parameter 404 to something higher than 39%

This is a known, but benign, issue where the total AC power is not always well recomputed when you have inverters with MPPTs connected to varying DC power. This is due to the total AC power being the sum of each MPPT equivalent AC power, itself based on the power sharing options. This approach is necessary to be able to deal with fraction of inverters, but may lead to small discrepancies in total. We will solve that in a future PVsyst version.

Since PVsyst 8.0.17, you have the possibility to use the "power shifting" strategy in Grid storage. If you do so and select "Fixed charging power before exporting" you can set it to charge as much as possible between 8 and 10. If the charging power is higher than the PV production, nothing will be injected to the grid. The main limitation for the moment is that any value you set will be applied every day for the whole year. Time series import should be available starting from PVsyst 8.1.0

Again, if you change any figure for the battery, it WILL impact your simulation. If you change the number of cells or the number of blocks or the number of racks, the battery characteristics will be impacted (voltage for the components in series, capacity for the components in parallel). The only issue I can see is the display of the number of UNITS in the grid storage window, which is not changed as it should when you change the number of blocks in the battery component. Now, if you have this values in your report (10*2 Units) but did not set 10 racks in series and 2 racks in parallel, please contact us at support@pvsyst.com with your project so we can have a better look at what's happening.