Whats more energy efficient: Off-grid or Grid-tie ?

[meetyg]

A bit theoretical, but has practical implications:
What's more energy efficient: Pushing PV power to the grid, or storing in battery and using it ?

Let's say our PV array produces 5kwh a day. For the sake of the discussion let's also assume that we are on a fixed grid tariff, and grid-tie is also 1:1 usage/payment ratio per kwh consumed/produced.

Now we all know that there are conversion losses in the way, whatever method we choose (grid-tied or off-grid with battery storage). My question is which one will save the most money on the monthly grid bill?

Let's get a bit technical:
Grid-tied: PV is DC, and this needs to be converted to AC. I suppose that this DC power is either stepped up to a higher DC voltage, or stepped down to a lower DC voltage, before being inverted to AC. Whether it's being stepped up or down, is probably related to PV voltage, the equipment being used and the grid system voltage (110v US or 220v European, for example).
Either way, we have some losses during this step-up or step-down phase. Are there more conversion stages in the way?

So our 5kwh PV in DC, might translate to say 80-90% AC output, being fed back to the grid.

Off-grid:
PV is DC and battery is DC. I'm assuming PV voltage is always stepped down because common battery voltages (12/24/48v) are usually under the PV voltages commonly used. But I know off-grid inverters often have an internal high-voltage DC bus. Is the PV voltage stepped up to this HV DC bus and then stepped down to the battery voltage?
Again, either way we have losses here too.
But there's more: After storing this PV power in the battery, we have more inefficiencies in the way:
1. Battery efficiency (how much power is put into battery vs. how much you can actually pull from it). Lifepo4 is probably somewhere around 90% efficient. Lead acid probably less. Let's assume Lifepo4.

2. Converting DC from the battery back to AC, for powering loads (inverting).

Intuitively, it seems that are more conversion steps with battery storage vs. grid-tied.
But am I correct?
If so, how much are we loosing here compared to grid-tied?

I'd like to hear your thoughts. Has anyone actually calculated or measured the end-to-end efficiency of thier grid-tied/off-grid system?

Thanks.

 

[AntronX]

So our 5kwh PV in DC, might translate to say 80-90% AC output, being fed back to the grid.

More like 97.6% if using Ephpase IQ7 micro inverters. Some string inverters get higher efficiency.

 

[1201]

with 1:1 net metering, having a battery will cost you money, not save you money.

a good grid tie inverter will export 98% of the pv directly to AC.

with a battery system, you save that to a battery at 98% efficiency ,if DC coupled with mppts, and then convert from battery to Ac at maybe 88-95% efficiency.

so, gt pv for grid export is going to be way more efficient

 

[jpwhit]

Generally speaking, storing power in batteries for later use will tend to be less efficient overall. Because as you say, there are typically more power conversions involved. And then you have the charge and discharge losses on top of that, which vary quite a bit between battery technologies. Lead acid charge / discharge efficiencies are typically somewhere between 75 and 85%. Lead acid also have non-trivial self-discharge rates that factor into the overall energy efficiencies. Where Li based batteries typically have round trip charge / discharge efficiencies of 95-97% with extremely low self-discharge rates to the point of being negligible.

Another non-trivial factor is battery system voltage. And that because high voltage battery systems have a significant advantage over low voltage battery systems. That's primarily because of the dramatic difference in currents involved. Things like wiring losses and FET RDS-on losses are directly proportional to the current levels involved in the system. But is also to a lesser extent that there is fixed voltage loss in electronics such as 0.7v diode drops that have less impact on a high vs low voltage system. Lastly, it's much less expensive to build an efficient AC inverter starting with HV DC than it is with low voltage DC.

But, in practice the real-world answer often depends on the exact equipment and the quality of that equipment when you compare within the scope of the same battery technology and high vs low voltage battery systems. And that's because the efficiency of the equipment can vary significantly enough to change the answer. And the reason I say within the same battery technology and voltage level, is because equipment can't do much about the efficiency of Li vs Lead Acid based batteries and to a limited extent between high and low voltage. I can spec out a grid tied system that is less efficient overall using low cost and lower quality components than a high end HV Hybrid system that stores excess power in batteries using high end costly equipment.

If you're in the fortunate situation where you have 1:1 cost recovery for power sent to the grid, it's probably always to your advantage to take as much advantage of that as possible for as long as it's available to you.

While not quite on topic of your original post, but I think in a similar spirit. Is the falling net metering payback rates and shift to time of use billing. The days of 1:1 payback on net metering is ending quickly. Solar production in the US is getting to the quantity now, that it's completely impractical for power companies to continue 1:1 payback rate. That's due to the glut of power now during the peak solar hours of the day, which doesn't match up with peak demand periods. Many states have already revised the net meeting riders to much less than 1:1 payback and almost all have also required net meeting solar customers to move to time of use rates as well. It’s only a matter of time before it changes everywhere.

IMO, battery based ESS will be required by almost everyone as time moves on to make the economics of residential solar work. Time of use rates are also an opportunity for additional savings. In North Carolina where both my houses are, anyone signing up for Net Metering now only get around 15% payback on power sent to the grid. It's also reconciled monthly so you don't have the opportunity to shift excess production in the summer to cover your higher power usage in the winter. Also, everyone with a Net Metering rider is also required to be on a time of use rate plan. But, the standard non TOU rate is 11.661¢ per kWh which is what I would pay without a solar system. The lowest discount rate in the TOU plan that I'm on with solar is 6.814¢ per kWh. That's a 42% savings.

I save as much by only buying power in the lowest discount period as I do with the 10kw of solar at my main house. I basically go off-grid and only run off the ESS except during the discount TOU rate period at that house. I only connect to the grid during the discount period and fully charge up the ESS and my EV cars during the night.

At my lake house, where I have nearly unlimited high production roof space for solar. I'm pretty much completely energy independent.

 

[wpns]

If you have 1:1 net metering you have a virtual battery at 100% efficiency, infinite size, and zero cost. What was the question again? Now you know why the power companies don’t like it.

 

[JBertok]

jpwhit made all the points I had in mind. In the simplest context, every component power goes through that gives off heat as a byproduct of what it's doing represents a loss in total efficiency equal to that heat. FLA inefficiency compared to LiFePO4 is observable as massive temperature increases during discharge and recharge comparatively.
The most efficient solar harvest possible is only compatible with 1:1 net metering, and that would be a grid-tied battery-less string inverter(s). If your net metering is even slightly less, or you can take advantage of time-of-use metering, it opens the door for a battery-backed system netting you more efficient use of your power. Grid reliability is an influential factor too. It's far cheaper to add panels to a grid-tie system than to build a battery system to meet needs if you've got a reliable grid, so it depends on your situation.

 

[wattmatters]

Now we all know that there are conversion losses in the way, whatever method we choose (grid-tied or off-grid with battery storage). My question is which one will save the most money on the monthly grid bill?

This ↓

with 1:1 net metering, having a battery will cost you money, not save you money.

Put another way, a battery is a net consumer of energy.

No point losing 10-15% of the energy going via a battery when you can buy it from the grid at the same rate you get as credit for exporting.

1:1 net metering is ridiculously generous.

If you have it then the only reason to have local storage is for grid outage backup. And then the storage may be in other forms (e.g. fuel for a generator). It all depends on the nature of outages in your area and what sort of backup needs you have.

 

[Ampster]

If you have 1:1 net metering you have a virtual battery at 100% efficiency, infinite size, and zero cost.

That is the way I look at it except for Non Bypassable Charges which cost me $0.025 per kWh which is still less expensive than a battery. I do use the grid as a seasonal battery because my winter loads are high when my production is low. I also use a traditional battery to do peak shaving because the differential between peak rates and my cost of battery usage is greater.

 

[meetyg]

1:1 net metering is ridiculously generous.

OK, so now comes the practical part:
I have a hybrid inverter with option to grid feedback. Unfortunately it doesn't support zero export, but I can limit how much power it pushes to the grid.

Now I don't really have 1:1 net metering, it was solely just for the sake of the discussion.
But, as I have a "dumb" meter, that will spin backwards (yes, I checked), I consider it 1:1 net metering.
Of course it's not legal to export power to the grid without a contract with the power Co. (which I don't have either).
But, as I can limit the export, I can basically do "zero export" as long as the limit is under regular/constant house loads.

As I'm from the Middle-East, not many homes have smart meters yet, so I'm less worried about getting in trouble with the local power Co. if a small feedback should happen past the meter, for some odd reason.

They are slowly rolling out smart meter upgrades, but currently there is no due date in our area.
But that day will come and I wanted to figure out how much I will be losing using batteries, disabling grid export.

Of course having batteries as backup is a great resource and that was one of my goals when building my small system.

Local grid is fairly stable, with the occasional power outage for 2-3 hours, when a grid fault occurs.
But as for now, these faults occur rarely.

So my batteries are usually sitting fully charged on sunny days and I want to make better use of my PV power.

 

[wattmatters]

So my batteries are usually sitting fully charged on sunny days and I want to make better use of my PV power.

You need to find some discretionary loads which can be shifted to daytime. The most obvious candidate is heating water / hot water storage. Another is an electric vehicle.

 

[wattmatters]

For example, this my last 3 days showing load (red) and PV output (yellow) showing how my load control makes good use of PV:



I have a smart PV diverter for the electric water heater, an off-grid battery where charging is managed, and an EV with smart charge control, also to maximise use of solar PV.

 

[DIYrich]

They are slowly rolling out smart meter upgrades, but currently there is no due date in our area.
But that day will come and I wanted to figure out how much I will be losing using batteries, disabling grid export.

If you "disable grid export" without batteties, then you loose any power generated that you don't immediately consume.

Batteries help you save the power you produce during the day, and use it when the sun is not shining. Typically, it takes about 10kwh of generation to use about 9kwh later on. It can be as high as 9.6kWh, or as low as 8.5kwh, depending upon the efficiency of your system.

You will still loose seasonality differences. Your produce a lot more excess in the summer that you can't store for later, and in the winter you may not produce enough to get you through the night.

You can size a smaller system, say 5kw panels with 15kwh of batteries that you 100% self consume all year and buy X kWk from the grid each year, a medium 10kw panels with 20kwh of batteries that provide 100% of your power all year and you buy Y kWh during the year for bad weather, or a large 15kw panels with 45kwh of batteries that lets you be grid independent 99% of the time, including during bad weather. (Sizes are purely hypothetical). Only you can decide if the incremental cost is worth saving "X-Y" or "Y".

 

[meetyg]

For example, this my last 3 days showing load (red) and PV output (yellow) showing how my load control makes good use of PV:

View attachment 215512

I have a smart PV diverter for the electric water heater, an off-grid battery where charging is managed, and an EV with smart charge control, also to maximise use of solar PV.

I happen to own an EV, but I'm usually at work during sun hours, so unfortunately it's less relevant.
We actually have a "solar" water heater, but it doesn't run on electricity to heat the water but rather a black panel with pipes, catching the heat.

But I suppose I'll think of something useful to do with the extra power when batteries are full...

 

[wattmatters]

But I suppose I'll think of something useful to do with the extra power when batteries are full...

Search the forum, there are several threads on the topic of using excess capacity.

 

[Ampster]

Generally speaking, storing power in batteries for later use will tend to be less efficient overall.

In addition, and another poster mentioned that storing energy on the grid is 100% efficient, with grid tie there is another aspect that there alway somewhere for excess solar to be stored. Whereas with off grid when batteries are full and loads are satisfied the charge controller just ramps down the solar production. In a sense this is a form of clipping of which there is a great deal of controversy. From a revenue model, I am saying that grid tie is always going to be more efficient given the assumption made in the first post that there is 1:1 net metering.

 

[wattmatters]

Physical efficiency and financial efficiency are different.

1:1 net metering has 100% financial efficiency.

But the grid has losses too, depends on the distance the power need to travel. Typical transmission and distribution losses on our grid average out at ~7%. And here in middle of the day there is often a lot of curtailment of grid production, with power plants often having to operate at lower power levels and that often means operating at lower efficiency.

 

[wpns]

1:1 net metering has 100% financial efficiency.

For the homeowner. As a matter of public policy, we force the financial losses onto the power company (and their ratebase), which was useful to encourage residential solar, but the pendulum is starting to swing the other way.

 

[wattmatters]

For the homeowner. As a matter of public policy, we force the financial losses onto the power company (and their ratebase), which was useful to encourage residential solar, but the pendulum is starting to swing the other way.

Whatever commercial / financial structures are put in place, in the end physics has the final say.

As I said earlier, 1:1 net metering is incredibly generous. As an incentive to kick start the addition of some renewables it might have made sense but it really needs a sunset clause/date.

In Australia that sunset clause was invoked nearly 10 years ago.

But such a change also needs to be accompanied with a harmonised, uniform and streamlined approval process for homeowners to install grid tied solar PV. Make it simple and far cheaper but also well regulated and designed to work well with the grid.

 

[wpns]

designed to work well with the grid.

Agreed, though the grid operators are still trying to work out what that means and how it’s implemented.

 

[wattmatters]

Agreed, though the grid operators are still trying to work out what that means and how it’s implemented.

True, although it's not like there are not plenty of locations with a lot of rooftop PV to learn from.