recursinging
New Member
- Joined
- May 14, 2020
- Messages
- 1
Hi everyone,
My employer (in Germany) just had another 200KWp PV array provisioned on the roof of a new building in our complex. We're now producing north of 500kWh excess per day, at a pretty unattractive buy-back price.
I got called into a meeting recently to discuss ways to make better use of this excess. One of the ideas floated was a hydrogen PEM/Fuel Cell + storage system, but the round trip efficiency of ~30% is not very attractive, despite the significant subsidies. A small fleet of EVs with V2B support was another idea I had, but after looking around, there seem to be only a couple of EVs and wall-boxes that support it.
So I did a little napkin-math and checked the Vmp temp coefficient of our panels against the typical 10-90% SoC voltages of LiFePO4 cells and found the efficiency to be worst case 95% - comparable with the DC-DC converters in MPPT trackers. So that got me thinking... why not try and scale up the DET (Direct Energy Transfer) concept of connecting the PV array directly to the battery, and build a giant bank of raw LiFePo4 cells to best match the estimated average MPP voltage of the solar array (across our expected temp range) - for example 190s. Then control charging/discharging with a modular BMS like foxbms (https://foxbms.org/), with diodes on the 20 panel strings, and contactors for the high and low voltage disconnect, keeping everything on the DC side of things.
Assuming we can safely and reliably handle the high voltage switching, the main problem I see is how to get energy out of the thing. A 200kWp PV array + 190s LiFePO4 bank that will connect and disconnect itself depending on its SoC, will probably not play nicely with our MPPT inverter infrastructure. We currently have four Fronius Symo 20.0-3-M 20kW MPPT grid-tie inverters hooked up to the PV strings. I can imagine that these would have issue with the low impedance of a battery bank connected on the tracker input, likely triggering some short-circuit protections. They would probably also sink as much current as they can from the battery when it is connected. Replacing them with a large grid tie inverter is a consideration, but we'd still need a MPP tracker when the battery is disconnected.
Yes, I know that hybrid tracker inverters exist. I'm waiting on a couple of quotes.
In the meantime I'd like to entertain this idea. Opinions are welcome.
My employer (in Germany) just had another 200KWp PV array provisioned on the roof of a new building in our complex. We're now producing north of 500kWh excess per day, at a pretty unattractive buy-back price.
I got called into a meeting recently to discuss ways to make better use of this excess. One of the ideas floated was a hydrogen PEM/Fuel Cell + storage system, but the round trip efficiency of ~30% is not very attractive, despite the significant subsidies. A small fleet of EVs with V2B support was another idea I had, but after looking around, there seem to be only a couple of EVs and wall-boxes that support it.
So I did a little napkin-math and checked the Vmp temp coefficient of our panels against the typical 10-90% SoC voltages of LiFePO4 cells and found the efficiency to be worst case 95% - comparable with the DC-DC converters in MPPT trackers. So that got me thinking... why not try and scale up the DET (Direct Energy Transfer) concept of connecting the PV array directly to the battery, and build a giant bank of raw LiFePo4 cells to best match the estimated average MPP voltage of the solar array (across our expected temp range) - for example 190s. Then control charging/discharging with a modular BMS like foxbms (https://foxbms.org/), with diodes on the 20 panel strings, and contactors for the high and low voltage disconnect, keeping everything on the DC side of things.
Assuming we can safely and reliably handle the high voltage switching, the main problem I see is how to get energy out of the thing. A 200kWp PV array + 190s LiFePO4 bank that will connect and disconnect itself depending on its SoC, will probably not play nicely with our MPPT inverter infrastructure. We currently have four Fronius Symo 20.0-3-M 20kW MPPT grid-tie inverters hooked up to the PV strings. I can imagine that these would have issue with the low impedance of a battery bank connected on the tracker input, likely triggering some short-circuit protections. They would probably also sink as much current as they can from the battery when it is connected. Replacing them with a large grid tie inverter is a consideration, but we'd still need a MPP tracker when the battery is disconnected.
Yes, I know that hybrid tracker inverters exist. I'm waiting on a couple of quotes.
In the meantime I'd like to entertain this idea. Opinions are welcome.
