André Mermoud

Yes, in the present time PVsyst is not able to import POA values measured with a tracking sensor. And I don't see any possibility to do this "manually": the transposition factor depends on the GlobHor, but also on the DiffHor irradiance values. When importing POA data, PVsyst performs an adjustment for retrieving a set of (GlobHor, DiffHor) values, which will exactly represent your POA data when transposed to the tilted plane (with the Hay model). We intend to extend the importing tool for tracking systems in the future. But sorry, I don't know when we will have time to do that. This requires that in the importing format, all Tracking parameters (including mode ans stroke limoits) should be carefully defined.

Each Power optimizer works on a different mode, so that it needs a specific development in PVsyst. I have worked closely with Solaredge for elaborating the SolarEdge architecture. I'm on the way just now to develop other power optimizers, with the help of their manufacturers. The TIGO technology should be available soon (within 1-2 months).

In the present version, the electrical shading effect is supposed to be null (or negligible) with the SolarEdge architecture. This is one of the main advantages of the SolarEdge technology, due to the fact that all modules operate with a common current within a string, and all strings operate at fixed voltage. Therefore I don't think I will extend the Module Layout possibilities to the SolarEdge system: I would not know how to do that. However I'm working on other power optimizers, and there may be some residual electrical shading effects, that I'm just investigating now. The results of these studies will be published with a future version of PVsyst, within hopefully 2-3 months.

Don't try to find an explanation. This is indeed a bug. I big bug, that I have never seen. Please check the Efficiency curve in your Inverter's definitions. However this should not arise of course. Please send me your whole project, using "Files" / "Export projects" in the main menu. (andre.mermoud@pvsyst.com)

The OK button stays grey until you have entered full valid data for defining your site (i.e. at least valid Global horizontal and temperatures monthly values). This prevent defining erroneous sites in the database. Now you can define these values either by importing them, or by hand (you can also paste them from EXCEL). The dialog should tell you what is incorrect. Now the sites for the US are stored partially under "USA" and partially under "United states" according to the country definition which have been made in each SIT file.

You can get the old installation packages of PVsyst V6 at the following address (you should adjust the 6_1_9 to your desired version): http://download.pvsyst.com/index.php?page=download&prod=true&key=PVsyst-prod%2Fbin%2FPVsyst6_1_9_setup.msi

This is not the right place for performing PVsyst simulations. You should use (on the main page) : "Project design" / "Grid connected". Here you are in a special (advanced) tool for the analysis of data measured on a PV system. DAM files are meteo files including additional measured data about your system (for example output energy).

This is not the right place for performing PVsyst simulations. You should use (on the main page) : "Project design" / "Grid connected". Here you are in a special tool for the analysis of data measured on a PV system.

These are quite different calculations. The "Unlimited sheds" option uses a simple analytic shading factor calculation, which supposes that the sheds are very wide (neglects the edge effects). The 3D shadings are based on the full 3D geometry calculation, but can use (according to your choice) a pre-calculated shading factor table, which may induce interpolation uncertainties. Both use also an integral for the diffuse and the albedo, which also depend on these calculation modes. The difference between both calculatioons is 0.4% in your case, which may be considered as acceptable. We usually observe a little bit lower difference, but please check that you did use the exact similar geometry (i.e same "inactive bands" for example).

Yes it is still difficult. I have established a model (i.e identified 3 main modifications of the standard one-diode model) which works quite well, and matches the real modules behaviour with a very good accuracy. The problem when using this model in PVsyst is to find the good parameters, as the information provided by the manufacturers is usually not sufficient. However the default parameters proposed by PVsyst seem to give good results. See my paper Performance Assessment of a simulation model for PV modulesof any available rechnology, ( 25th EPVSEC – Valencia, 2010), available on our site http://www.pvsyst.com.

In the version 5.73, the night losses are the iron loss of the external inverter. In this version, night loss of the inverter output or auxiliary losses had not been implemented yet.

For the PAN "proprietary" format: if you want to reproduce the old DELPHI encoding of reals on 6 bytes, and modify/update your reading according to all specific parameters defined (or not) for each PV module, I can give you the format. You will probably spend several days (or weeks) for using it. I really don't know any model specifying the effects of the humidity on PV modules, and I can't imagine which phenomenons could affect the PV module behaviour (except morning condensing effects). Please explain. However the meteo data of PVsyst don't include the Humidity information in the present time, and I don't intend to include it unless very strong reasons. Now for the comment of the W7 cell, you should indeed read: "For crystalline modules: if not specified, will be a result of the one-diode model, which is very close to the usual reality".

For the shading loss changes: If your system involves unlimited sheds: in the version 6.13 to 6.18, there were an error when reading the VCi file. The plane tilt used for the evaluation of the shading effect on the diffuse part was not read correctly, which underestimated the shading loss. This error didn't occur with new calculation versions, or if you opened the "Orientation" dialog before the simulation. This problem has been corrected in the version 6.19. On the other hand, the diffuse model is the same in both cases: just the denomination on the report has been changed. "Measured" was used previously for any values read on an hourly file (whatever the origin, which is not known by PVsyst). Now for most meteo data, the diffuse is indeed issued from a model. Therefore this denomination is not correct. In the new versions we have tried to put some more accurate description when we have an idea of the origin of the data. When creating synthetic data (which uses the Meteonorm algorithm), the diffuse is constructed by the Meteonorm software, using a Liu-Jordan or similar (Erbs) model.

The PAN files are written in a specific binary format, only readable/writable by PVsyst. If you want to manage PAN files explicitely, you can use the Import/Export facility to an EXCEL document. You have a template of this EXCEL document in your \PVsyst6_Data\UserData\. The Import/Export procedure is explained in this document. However it is of course not possible to add your own parameters in the file, besides the "official" format. How would they be recognized by PVsyst ???

If so, this is indeed a bug, and I will investigate this. However in the "Detailed Losses" part, if you define the IAM explicitely by unclicking "Uses definitions of the PV modules", the IAM is now specified for the sub-array itself, and will be kept if you change the PV module.

The 6.1.9 version is indeed the 6.19. I don't know why you cannot install it. Please explain what happens.

Each array connected to a given MPPT input should be "homogeneous" (i.e. identical PV module and number of modules in series). You should define one sub-array for each kind of "homogeneous" array.

Yes, if there are several inverters, the section managed by PVsyst is indeed the sum of all individual sections of all inverters. The minimum section proposed by the program corresponds to the section allowed (by usual practice or regulations) for the concerned current.

I don't know. Which version are you using ? Please send me you full project, using "Files" / "Export whole projects" inthe main menu (mail to andre.mermoud@pvsyst.com).

You are right, the grid limitation is applied to the PNom(ac) of the inverters in the present time. The losses after the inverter are not taken into account. In a first try, I just foresee this for a simple case. This is the easiest way for programming. However I have to do some improvements and give some further possibilities to this grid limitation for a next version. - As you propose, the limitation at the injection point: not easy because PVsyst has to perform the calculation twice for each concerned hour. And I'm not sure that the inverters will operate in this way in the reality: this would require a continuous measurement of the injected power, and a return of this information to the inverter software. - If you have several sub-systems (not in the same orientation) the limitation should indeed be applied to the injection, and will act differently on each sub-array. Again this involves the re-calculation at each concerned hour, and is not simple to include in the simulation.

This is an error within your shading construction file: a parameter set at an unexpected value (I don't know why). I have protected the program against such cases, for the next version 6.19.

The wiring resistance is defined for each sub-array separately. It is the equivalent resistance of all wires, as seen from the input of all inverters of the sub-array in parallel. You have just to put all the string resistances in parallel, and add the resistances of all connexions from the roof junction boxes to the inverter inputs. This is straightforward for any electrician engineer. The tool "Detailed computation" helps you for this calculation. Now the resistance is around 5.5 mohm/m for 4 mm2, and 8.8 mohm/m for 4 mm2 (listed in PVsyst, button "Wires"). As an example, If you have 40 m of 4mm2 cable for one string, this will represent 220 mOhm/string. I.e. 22 mOhm for 10 strings in parallel. I don't see how you can get 0.005 mOhm except for a multi-100 MWp instalation (this would represent 44'000 strings).

Further information. First of all, on your table you don't mention the differences (in percentage), so that the visual analysis is very difficult. Now in PVsyst the PR is computed from the whole simulation process. This involves other losses which also vary along a day. These are mainly the IAM loss, eventual shading effect, module efficiency according to the temperature, wiring resistance loss. Other effects of second order may also occur.

For a row installationm if the cell strips are in landscape (not necessarily the modules: in some modules the cells are along the little side) this is indeed the worst situation and you will have a full electrical loss as soon as the bottom of the modules is shaded.

This may probably be due to a dissymmetry between morning and evening in your meteo data. Or an horizon shade if any. I don't see other explanations. However such low differences are not necessarily very significant