It should be buck boost (isolation reqd) with mppt voltage of 300 - 1000 VDC to output 650VDC. if not single 50 KW preferred, how many trackers can be used?
Edit : Output DC is converted to three phase AC. It is battery less system
Eventually you may consider a boost stage for the MPPT (DC rail voltage eg. 1 kV) - followed by a buck controller to 'create' the 650 V. To keep ripple low I'd suggest a 4- phase design for each stage. This would result in 12.5 kW per phase branch.
Depending on the availability of appropriate semiconductors, you may eventually consider more phases.
Depends. With a multi-phase design, it is much easier to start with individual stages that have well-defined (and rather constant) set-points. These are relatively easy to optimize.
An "all-in-one" design - including isolation - will be quite "tricky" and much harder to optimize as everything is interlinked with everything.
Thus my suggestion to start with individual stages. Once you got this design 'perfect', you are free to try an all-in-one approach. Then you will also have a "benchmark" system in terms of cost, performance etc.
At first i would like to greet the respected colleague U. Dreher for his proposals. From the principle point of view, you have first to define your load precisely. Then you can size the photovoltaic system and then you can specify the poer electronic circuits for power conditioning, that is the charge controller . I wonder that you speak of terms like full bridge configuration. Such configuration is used normally for rectification and for DC to AC inverters. Here your source is DC and your your source is a DC load the batteries so you need to transfer the power from the pv source to the batteries.
Normally the batteries are divided into battery banks. So you can build a separate charger for each bank. In this case you increase the reliability of your system.
So divide the the whole storage into uniform battery banks and feed every bank with the suitable PV generator.
What is the suitable operating current and voltage of the banks can be optimized for reliable operation of the system. Operating at high voltage will save the ohmic losses in the cupper and it will save also cables. On the other side the maximum operating voltage of the modules is restricted with defects rising because of the high voltage operation.
I think the distributed modular system is the most appropriate for such applications.
I haven't posted the whole project here. It's actually solar Inverter what I'm working on. Since, I would like to prove my Charge controller part first, I haven't added the rest. So, basically the load is three phase AC and battery less system. Sorry for the confusion.
Adding to the point above, full bridge isolated DC-DC converter is widely used at high powers. That's the reason I suggested. Please see attached link. Resonant type is used in the application
Isolation is only implemented when and where necessary. For automotive applications, isolation is mandatory. For stationary applications, other means of protection are less costly.
When driving an isolation transformer the "classic" way, a full-bridge design is appropriate. When considering "transformer-less" topologies (CuK and alike - or "coupled inductors"), full-bridges add non-necessary complexity.
The Cree design is still some way from you target power of 50kW. AND it implements isolation to isolate the vehicle from the grid!
When you implement the targeted DC/AC converter (grid connection) in your design, it would be easy to remove the DC/DC buck converter (leaving the PFC stage as it is - functional and intact) and add the DC/AC stage.
While an integrated PFC+DC/DC stage would have to be removed completely.