Perplexity is amazing.

In this case the answer is only as good as the assumptions going in. The assumption about what the system will generate is in all likelihood wildly inaccurate. Solar systems never generate as much as their theoretical nameplate capacity states. The theoretical capacity is only reached under laboratory conditions. Factors like latitude, average weather conditions, panel tilt, dust, etc. will reduce that amount by a huge amount. Also, you actually have to use the power for it to have value. Since demand varies, much of the generation capacity is wasted so the payoff formula must take daily average usage and the buyback price of the excess current into account.
 
#3 has been debunked pretty throughly at least in the US. This was a sales job from the utilities.
If you are paying generators (doesn't matter if they are called net metered or not) $0.52/kWh it's going to drive-up retail rates, because the other ratepayers still have to cover that cost. This is while demand goes down, so you end up having a higher cost spread over fewer kWh, which drives up the overall supply cost; drives up system cost.

California doesn't have the highest electricity prices in mainland North America because their supply price is cheap, lol.

These schemes were implemented to spur adoption, unfortunately, the long-term impacts weren't well studied and by the time it was dropped (NEM 3, cancellation of Ontario's GEA) the damage was done.
 
In this case the answer is only as good as the assumptions going in. The assumption about what the system will generate is in all likelihood wildly inaccurate. Solar systems never generate as much as their theoretical nameplate capacity states. The theoretical capacity is only reached under laboratory conditions. Factors like latitude, average weather conditions, panel tilt, dust, etc. will reduce that amount by a huge amount. Also, you actually have to use the power for it to have value. Since demand varies, much of the generation capacity is wasted so the payoff formula must take daily average usage and the buyback price of the excess current into account.

Correct.

I'm my case I use the actual number from both bills and consumption monitoring vs any nameplate #'s.
 
If you are paying generators (doesn't matter if they are called net metered or not) $0.52/kWh it's going to drive-up retail rates, because the other ratepayers still have to cover that cost. This is while demand goes down, so you end up having a higher cost spread over fewer kWh, which drives up the overall supply cost; drives up system cost.

California doesn't have the highest electricity prices in mainland North America because their supply price is cheap, lol.

These schemes were implemented to spur adoption, unfortunately, the long-term impacts weren't well studied and by the time it was dropped (NEM 3, cancellation of Ontario's GEA) the damage was done.

Sure. It will drive up the cost. The question is what is that cost really?
Lawrence Berkeley did a pretty good study on this.
Its minuscule - way lower than all of of the other variables like fuel pricing

Originally the number was .03, and it was revised later to .0003.

PG&E (my " man") is the highest in the nation. California proper a bit lower.


I'd be happy to pay your man, but do my best to avoid mine.
 
Heres an interesting question.

Which has more G forces ?

A 24" drum spinning 1600RPM, or a 27" drum spinning 1300.
 
Well, I can give you those numbers for Ontario if you'd like? Don't want to further derail your thread though if you don't want me to.
You've shown them, it's a much bigger impact than here. I think your original nem stuff was way way out of wack, our was too.

This is why I said " in the US" because I know from you the Canadian numbers are different.
 
You've shown them, it's a much bigger impact than here. I think you original nem stuff was way way out of wack, our was too.

This is why I said " for the US" because I know from you the Canadian numbers are different.
Yes, what we had wasn't called "NEM" it was "FIT" (and MicroFIT) but similar impact, paying people $0.60-$0.80/kWh to force solar adoption.

You need total supply cost to figure it out accurately. We have that information for Ontario, as well as individual source supply cost averages, so it's easy to calculate the impact. It's more difficult where that data isn't obtainable or only obtainable in part.

If we assume that the average rate for solar in Cali is $0.45/kWh just for the sake of this exercise, solar generated 41.3TWh and was the 2nd largest source behind gas in 2023.

Total generation was 207TWh, so solar was 20% of supply and at the above rate, at a cost of $18.6 billion

If we assume the cost of gas was $0.04/kWh (which is about what it is here plus the price to be on standby) which is about market rate, then gas at 87.2TWh was 42% of supply at a cost of $3.5 billion.

Diablo Canyon should also be around $0.04/kWh, so nuclear at 17.8TWh and 8.6% of supply was $712 million.

If we just fudge the remaining 60.7TWh at $0.15/kWh we end up with $9.1 billion.

That's a total supply cost of $31.9 billion, or an average cost of $154/MWh; $0.154/kWh. If the 41.3TWh was generated by gas at the margin, total supply cost would be $14.96 billion, an average cost of $72/MWh; $0.072/kWh.

That's rather massive.

Of course we have to look at foregone generation as well; generation avoided by self production under NEM which would also be by marginal rate gas most likely, increasing the overall supply volume at a lower cost, which would make the results even more dramatic.

The other factor is commercial and industrial users which get a preferential (lower) rate than Joe Average consumer, which further drives up the supply cost for retail customers.
 
Last edited:
What would really be nice is if it could figure out the many other variables on it's own. You know, things like the ACTUAL cost of electricity ie. (total monthly bill/kwh, sort of thing) As well as the other associated costs of installation. For example, how much your homeowners insurance went up, what is the cost of removing/installing the panels when you need a roof replacement or repair. You know little things like that.
Otherwise, it's gonna be great one day!
Excellent point! Beautiful point! Spot on!
As humans, we make our decisions based on what we have learned, experienced, etc. We develop patterns. The decision we make may be perfect, or very good right now, but as we know, change is the only constant.

We may become more educated. There are more variables than we currently know (and consider). There is far more data and constantly more new data. Our needs change. Tech changes. So our patterns evolve. This doesn't make our prior decision bad; new science improves old science, right?

The human mind is the AI design and how AI works; it is self learning.

At its core, artificial intelligence is about the ability to use algorithms to define patterns. Once you recognize the patterns, you can eventually predict them. And once you can predict them, you can prescribe meaning. Just like the human brain, that's the self learning part.

That "meaning", aka decision or result, will constantly evolve and improve, just like ours does.
 
Last edited:

My property has one half a miners inch of incoming water that is stored in a 2700 gallon buffer tank. How long can I run a sprinkler zone at 15 gallons a minute before running out of water?​



Answer
To answer your question about how long you can run a sprinkler zone at 15 gallons per minute (GPM) with your given water supply, let's break down the information and calculations:
  1. Water Supply:
    Your property has one-half miner's inch of incoming water. According to standard measurements, one miner's inch is equivalent to approximately 11.22 gallons per minute.Therefore, your water supply is:0.5×11.22 GPM=5.61 GPM0.5×11.22 GPM=5.61 GPM
  2. Buffer Tank:
    You have a 2700-gallon buffer tank for water storage.
  3. Sprinkler Usage:
    Your sprinkler zone uses water at a rate of 15 GPM.
Now, let's calculate how long you can run the sprinkler zone:
  1. Initial Water Available:
    You start with 2700 gallons in the buffer tank.
  2. Net Water Usage:
    While running the sprinkler, you're using 15 GPM but still receiving 5.61 GPM from your water supply. So the net usage is:15 GPM−5.61 GPM=9.39 GPM15 GPM−5.61 GPM=9.39 GPM
  3. Time to Empty the Tank:
    To find out how long you can run the sprinkler, divide the tank capacity by the net usage rate:2700 gallons9.39 GPM=287.54 minutes9.39 GPM2700 gallons=287.54 minutesThis is equivalent to approximately 4 hours and 48 minutes.
Therefore, you can run your sprinkler zone at 15 GPM for about 4 hours and 48 minutes before the buffer tank runs out of water.It's important to note that this calculation assumes:
  • The buffer tank is full (2700 gallons) when you start.
  • The incoming water supply remains constant at 5.61 GPM.
  • The sprinkler system consistently uses 15 GPM.
In practice, it's advisable to stop before the tank is completely empty to avoid potential issues with your water system. Also, remember that refilling the tank completely will take some time after you've used a significant portion of its capacity.
 
And I could have asked AI to code a Linear Regression curve against a large SQL set of data instead of me racking my 2 brain cells coding thousands of lines...
 
Aren't we looking at a Hal9000 type of scenario at some point?
Making machines do more so people can do more.

In time, there will be no set of Doctors that can diagnose symptoms like AI. It can analyze history and up to the moment data including what is going on medically in anywhere on the planet. And it doesn't sleep...
 
Back
Top