From what I see, it seems like going in with a technology focus, simplifying the payment process for doctors, and destroying the highest-margin procedures with competition could have a pretty significant impact. The for-profit insurance companies don't have any real interest in getting end-user costs down, just their costs. The non-profits seem to have other challenges (those little spoken about for-profit divisions as an example).
You could save 10% over night pretty easily in my mind... maybe 25%.
You look at the real-time waveform rather than the integrated frequency and get a slew rate for each PWM interval, and marginally increase or decrease output to assist in maintaining the target waveform and thus frequency. (An AC generator watches phase angle in order to load or unload itself.)
At a grid level, frequency a more reliable measure than anything else, since increasing your generator's excitation current will temporarily boost output, until the point that the prime mover cannot keep up. Most voltage regulators are programmed with a V/Hz operation mode that lets them reduce voltage and frequency in tandem when overloaded, and increase them when overpowered.
The control networks will only act predicatively or reactively-- not really in a real-time mode. A number of things can be predicted - increased cloud cover with a weather pattern moving towards a solar array, same types of things with wind power. Demand-side, there is generally enough information in aggregate to see variations between day-ahead forecasts and real-time status. But, when you get transient events like a power plant or transmission line tripping offline, real-time interactivity of a battery system can mean the difference between major sub-systems tripping and needing to be slowly re-closed, and zero impact.
You should be able to fully control the device via the hotel interface system available on many TVs. You can brute-force the same thing with an IR Blaster covering the IR receiver on the TV. A little bit of scripting on a Raspberry Pi to glue it all together and you are done.
As much as I do hate the MM abbreviation, the alternative “m” that is commonly accepted for million just pisses me off. The ambiguity of a single M likewise makes it a poor choice. I guess you could be pedantic and write 100M$... but what the hell...
You can handle it on a sub-cycle level, shifting your output slightly forward on the waveform to the capacity of your system. You are just limited by your own capacity as to what you can do to stabilize things.
I am not sure if you have cause and effect reversed. I’m also not 100% sure if we are going to start seeing a drop in the AWS strategy based on current security concerns.
The layers of security get obscured with hosted and virtualized services, and I am seeing some activity that suggests more companies are starting to move more services in-house on dedicated hardware. Too soon to know for sure though.
From what I can tell, they seem very business-unit centered with basically workgroup IT strategies. It seems depressingly half-assed. Might explain their products some...
The wholesale structure is different than the retail structure; while I doubt the premise of earning $1MM in a week, being paid to absorb energy for a few minutes and possibly up to a half-hour is common for grid stabilization as plants spool down. It averages out to a rounding error in the scale of the grid and given the limited duration.
Supposedly it was US$100MM construction cost (which seems high; I would expect closer to $50MM). Environmentally and economically, it is a great buffer when coupled to a wind farm like it is; the majority of wind energy is produced at night when demand is lower, and when you are at peak output it is difficult to sell/use all the power.
If you assume a 3,500 cycle life, 90% round-trip efficiency, and just time shifting, you need US$0.25/kWh delta between charging price and discharging price. If these are actually good for 10,000 cycles then you are closer to a more reasonable $0.08. Presumably they are also paid for other grid stabilization services, which would reduce the needed price delta by up to 30% or so.
I would love to understand how they optimize operation; I get the sub-transient and short-time operation logic, and at least at a high level predicting real-time price swings-- but as a whole I can't quite wrap my head around how they control it.
Do they control it based just on what they are paid to do at a given point in time, or does it simply act as a "good citizen" of the grid? Does it work on 24-hour look-ahead (or longer), or is it more responsive real-time? What is the minimum charge level they target?
It will be interesting to see how these large batteries work when there are multiple units controlled independently.
Aside from needing double track in more locations and electrification, the US has a robust rail network, and I fail to see why “more” is needed. While it is impressive that some places can run rail to within a mile of the destination, it seems like an unrealistic strategy for the US. I do find it ironic that most of the logistics centers seem to forego rail spurs, but there is a reason.
What I think the US railroads really need though is micro switching and intermodal transfer yards to better support local transfers.
Those things are already well known; apparently they didn’t hit their stride until the end of December, where they were at a rate of 1000 model 3’s per week in the last three days. Timing now seems designed to hit the stock before earnings.
Also, the base $35k model is a random reference... of course the lowest margin version will be last.
Based on the fact that I have seen a few model 3’s on the road this past week (first ones for me), I am guessing production is consistent now and possibly accelerating beyond 1,000/week.
It will be interesting to see where we end up after the Spectre/Meltdown lawsuits. I cringe at the resolution (or lack thereof) for my 6-month old iMac.
While your request is perfectly reasonable, it can quickly become unworkable. Some products would no longer be possible (for better or worse), and "easily" is pretty darn subjective... unless you add the phrase without the use of tools.
I'm not sure where we should be going on these things; my iPhone X that replaced my iPhone 6 uses up its battery significantly faster, and I am fairly certain it will be significantly degraded by the time I want it replaced. But, historically batteries were not the thing that made me replace my phones-- maybe one case in 15 years/10 phones (the old days when you had a separate phone from Blackberry included). Most were replaced because of feature requirements, and two were broken when dropped. The only phone I ever purchased a replacement battery for was a RAZR, and that turned out to not really be my problem.
You also have to remember that most people are not criminals, and are unlikely to be investigated for a crime (presuming proper due-process), as they are unlikely to be the victims of a crime for which evidence from a phone is likely to make a difference to their outcome.
I am sure there is an intelligent comment in there somewhere; does anybody have any idea what he means? I can see how someone could apply "too geeky" to certain languages which makes them a poor choice for your first useful program... but talking to 7th graders it doesn't really make sense to talk about C, at least to me.
The guy walked away from the crash, so the car did its job. I guess they isolate the following and stopping logic because there are too many motionless things that cannot be logically tracked or anticipated until you are on an otherwise irreversible collision course with them.
Room for improvement to say the least; something blocking the traffic lane should be considered prior to impact.
Software companies like Adobe and AutoDesk can now squander even more money with their near-monopoly control over their core industries. Thus far, I don't object to Microsoft's pricing nearly as much with 365, but it will hit a limit quickly with any increases.
With AutoDesk, it approaches 3% per year of an employee's direct labor cost for us now (plus a 20% productivity hit for using BIM). It isn't sustainable long term, and they will kill their golden goose.
I completely fail to understand SalesForce, but I guess the attraction is lower barriers to entry for enterprise functionality easily linked to legacy systems.
One of my few gripes last month was that it was hard for me to unlock in bed laying on my side with faceid. That seems to have improved dramatically since then.
The issues with the screen and home zone you describe are pretty limited for me; most of my apps don’t take advantage of the notch yet (and there are a number of great ways to improve the UI by adding in support, like putting the menu in the top left so it disappears), but not that big of a deal.
I have almost gotten over the animated poop, but there is a time and place for everything.
I don’t trust FaceID yet or my banking passwords, but for locking and unlocking the phone it is great— a transparent means of unlocking vs a more conscious move. Apple seems to have made a few tweaks that addressed my edge case issues on the phone, and I have to agree with GP: the X is the future. I also happen to think they sold many more than the pessimists and it is a blockbuster.
Yeah— it seems like a very odd list with more than half the locations completely unworkable. NY/Philly (3), DC (3), LA, Chicago, and even Denver seem almost impossible to work effectively. I see the logic of Dallas/Austin, and Atlanta. I think Indy would be a great city for it, as might Nashville— but my money would be on Columbus; it is the hub of their air operations.
Slashdot Mirror
Well stated.
From what I see, it seems like going in with a technology focus, simplifying the payment process for doctors, and destroying the highest-margin procedures with competition could have a pretty significant impact. The for-profit insurance companies don't have any real interest in getting end-user costs down, just their costs. The non-profits seem to have other challenges (those little spoken about for-profit divisions as an example).
You could save 10% over night pretty easily in my mind... maybe 25%.
You look at the real-time waveform rather than the integrated frequency and get a slew rate for each PWM interval, and marginally increase or decrease output to assist in maintaining the target waveform and thus frequency. (An AC generator watches phase angle in order to load or unload itself.)
At a grid level, frequency a more reliable measure than anything else, since increasing your generator's excitation current will temporarily boost output, until the point that the prime mover cannot keep up. Most voltage regulators are programmed with a V/Hz operation mode that lets them reduce voltage and frequency in tandem when overloaded, and increase them when overpowered.
The control networks will only act predicatively or reactively-- not really in a real-time mode. A number of things can be predicted - increased cloud cover with a weather pattern moving towards a solar array, same types of things with wind power. Demand-side, there is generally enough information in aggregate to see variations between day-ahead forecasts and real-time status. But, when you get transient events like a power plant or transmission line tripping offline, real-time interactivity of a battery system can mean the difference between major sub-systems tripping and needing to be slowly re-closed, and zero impact.
You should be able to fully control the device via the hotel interface system available on many TVs. You can brute-force the same thing with an IR Blaster covering the IR receiver on the TV. A little bit of scripting on a Raspberry Pi to glue it all together and you are done.
As much as I do hate the MM abbreviation, the alternative “m” that is commonly accepted for million just pisses me off. The ambiguity of a single M likewise makes it a poor choice. I guess you could be pedantic and write 100M$... but what the hell...
Thanks!! Great info.
You can handle it on a sub-cycle level, shifting your output slightly forward on the waveform to the capacity of your system. You are just limited by your own capacity as to what you can do to stabilize things.
I am not sure if you have cause and effect reversed. I’m also not 100% sure if we are going to start seeing a drop in the AWS strategy based on current security concerns.
The layers of security get obscured with hosted and virtualized services, and I am seeing some activity that suggests more companies are starting to move more services in-house on dedicated hardware. Too soon to know for sure though.
From what I can tell, they seem very business-unit centered with basically workgroup IT strategies. It seems depressingly half-assed. Might explain their products some...
The wholesale structure is different than the retail structure; while I doubt the premise of earning $1MM in a week, being paid to absorb energy for a few minutes and possibly up to a half-hour is common for grid stabilization as plants spool down. It averages out to a rounding error in the scale of the grid and given the limited duration.
Supposedly it was US$100MM construction cost (which seems high; I would expect closer to $50MM). Environmentally and economically, it is a great buffer when coupled to a wind farm like it is; the majority of wind energy is produced at night when demand is lower, and when you are at peak output it is difficult to sell/use all the power.
If you assume a 3,500 cycle life, 90% round-trip efficiency, and just time shifting, you need US$0.25/kWh delta between charging price and discharging price. If these are actually good for 10,000 cycles then you are closer to a more reasonable $0.08. Presumably they are also paid for other grid stabilization services, which would reduce the needed price delta by up to 30% or so.
I would love to understand how they optimize operation; I get the sub-transient and short-time operation logic, and at least at a high level predicting real-time price swings-- but as a whole I can't quite wrap my head around how they control it.
Do they control it based just on what they are paid to do at a given point in time, or does it simply act as a "good citizen" of the grid? Does it work on 24-hour look-ahead (or longer), or is it more responsive real-time? What is the minimum charge level they target?
It will be interesting to see how these large batteries work when there are multiple units controlled independently.
Aside from needing double track in more locations and electrification, the US has a robust rail network, and I fail to see why “more” is needed. While it is impressive that some places can run rail to within a mile of the destination, it seems like an unrealistic strategy for the US. I do find it ironic that most of the logistics centers seem to forego rail spurs, but there is a reason.
What I think the US railroads really need though is micro switching and intermodal transfer yards to better support local transfers.
Those things are already well known; apparently they didn’t hit their stride until the end of December, where they were at a rate of 1000 model 3’s per week in the last three days. Timing now seems designed to hit the stock before earnings.
Also, the base $35k model is a random reference... of course the lowest margin version will be last.
Based on the fact that I have seen a few model 3’s on the road this past week (first ones for me), I am guessing production is consistent now and possibly accelerating beyond 1,000/week.
It will be interesting to see where we end up after the Spectre/Meltdown lawsuits. I cringe at the resolution (or lack thereof) for my 6-month old iMac.
While your request is perfectly reasonable, it can quickly become unworkable. Some products would no longer be possible (for better or worse), and "easily" is pretty darn subjective... unless you add the phrase without the use of tools .
I'm not sure where we should be going on these things; my iPhone X that replaced my iPhone 6 uses up its battery significantly faster, and I am fairly certain it will be significantly degraded by the time I want it replaced. But, historically batteries were not the thing that made me replace my phones-- maybe one case in 15 years/10 phones (the old days when you had a separate phone from Blackberry included). Most were replaced because of feature requirements, and two were broken when dropped. The only phone I ever purchased a replacement battery for was a RAZR, and that turned out to not really be my problem.
You also have to remember that most people are not criminals, and are unlikely to be investigated for a crime (presuming proper due-process), as they are unlikely to be the victims of a crime for which evidence from a phone is likely to make a difference to their outcome.
I am sure there is an intelligent comment in there somewhere; does anybody have any idea what he means? I can see how someone could apply "too geeky" to certain languages which makes them a poor choice for your first useful program... but talking to 7th graders it doesn't really make sense to talk about C, at least to me.
The guy walked away from the crash, so the car did its job. I guess they isolate the following and stopping logic because there are too many motionless things that cannot be logically tracked or anticipated until you are on an otherwise irreversible collision course with them.
Room for improvement to say the least; something blocking the traffic lane should be considered prior to impact.
Software companies like Adobe and AutoDesk can now squander even more money with their near-monopoly control over their core industries. Thus far, I don't object to Microsoft's pricing nearly as much with 365, but it will hit a limit quickly with any increases.
With AutoDesk, it approaches 3% per year of an employee's direct labor cost for us now (plus a 20% productivity hit for using BIM). It isn't sustainable long term, and they will kill their golden goose.
I completely fail to understand SalesForce, but I guess the attraction is lower barriers to entry for enterprise functionality easily linked to legacy systems.
If it is fast enough it would not need to try to be stealthy in attack mode, but the propulsion method seems to be the key logistics issue.
My guess is that it would be more of a manned submarine that has a 100MT warhead built in, and the crew is sacrificial.
One of my few gripes last month was that it was hard for me to unlock in bed laying on my side with faceid. That seems to have improved dramatically since then.
The issues with the screen and home zone you describe are pretty limited for me; most of my apps don’t take advantage of the notch yet (and there are a number of great ways to improve the UI by adding in support, like putting the menu in the top left so it disappears), but not that big of a deal.
I have almost gotten over the animated poop, but there is a time and place for everything.
Power off first— not that hard.
I don’t trust FaceID yet or my banking passwords, but for locking and unlocking the phone it is great— a transparent means of unlocking vs a more conscious move. Apple seems to have made a few tweaks that addressed my edge case issues on the phone, and I have to agree with GP: the X is the future. I also happen to think they sold many more than the pessimists and it is a blockbuster.
Honest question— with or without kids?
We live in one of the most (empirically) desirable spots in Southern California, without kids, and could live comfortably at 1.75x CReimer’s salary.
No, they just need to move to circular runways. Odd proposal, but could be interesting.
Yeah— it seems like a very odd list with more than half the locations completely unworkable. NY/Philly (3), DC (3), LA, Chicago, and even Denver seem almost impossible to work effectively. I see the logic of Dallas/Austin, and Atlanta. I think Indy would be a great city for it, as might Nashville— but my money would be on Columbus; it is the hub of their air operations.
Or Toronto to piss off/on Trump.