Tom's Blog

kinetic energy recovery systems

2009-10-22T20:18:00.004-05:00

(edit: after a bit of googling it seems I am not the only one with this idea. and I am starting to think it will also be beneficial in full electric and gas/electric hybrids. also, this would be great in many other application such as trains/cranes/etc)

In all present road vehicles, braking is achieved by an absurdly crude and wasteful method: The kinetic energy of the vehicle is turned into heat by friction in the brakes, where it is (literally) thrown into the wind.

The main benefit with battery electric hybrids is regenerative braking. And with electric vehicles the mechanical energy has to be converted to electrical and then back to mechanical. There are losses in efficiency in this conversion.

What I am curious about is why I cannot find very much information about mechanical kinetic energy recovery systems (eg storing the energy used in braking in a flywheel). I found some stuff about F1 using something similar with a cvt controlling the flywheel. however I cannot find any information on efficiency gains nor can I find anything pointing to anyone (major manufactures or small startups) trying to develop such a system. edit: http://green.autoblog.com/2007/10/31/more-details-about-the-flywheel-kinetic-energy-recovery-system/

I cannot help but think that the efficiency gain would be huge (eg energy would not be wasted starting and stopping, essentially highway mpg would == city mpg). Not to mention the possible improvement in performance (eg have power delivered to the wheels from the flywheel AND the engine) while reducing the wear on the brakes. While yes, you might be carrying around an extra 30-40lbs, the energy needed to move that weight I would imagine would be far surpassed by the amount of energy it could store and reuse eg http://en.wikipedia.org/wiki/Rotational_energy

If the efficiency gains are large it might not be a bad idea for myself to work on creating such a system that can be easily adapted to any automobile (would not make sense for gas/electric hybrids since they already have regen braking).

Still, it looks like the small companies that are working on it do not actually have a product that is out to market yet (just F1, nothing for consumers yet). Probably quite a bit of room for some competition. I need to get a crappy car and a few thousand dollars to develop this now...

EDIT: here are a few high level mock ups of current systems (without a kinetic energy recovery system) and with a kinetic energy recovery system using a flywheel. Ideas on how it can be improved... let me know in a comment.

Eg when you first start driving or when there is no energy in the flywheel:

Used for braking/slowing down:

Reusing only the energy stored in the flywheel from braking/slowing down:

Using energy from both the engine and the flywheel for added efficiency and power:

other notes:
"A flywheel system offers up to twice the efficiency of a battery storage system. The overall in-out efficiency of 65 to 70 percent compares to 35 to 45 percent for a hybrid battery-electric system. That is because it's an entirely mechanical system. In a battery system, kinetic energy has to be converted back and forth to electrical and chemical energy, thus the efficiency losses."

"The energy stored in a lightweight steel and composite flywheel spinning at over 60,000 rpm can be used to accelerate the vehicle. The FIA defined the amount of energy recovery for the 2009 season as 400 kiloJoules per lap, providing an extra 80 horsepower over a period of 6.67 seconds. Its low weight, about 55 pounds, is also very important in motorsports."

article from 1993

what would be cool would be to also implement is a type of cvt. eg totally replace the friction brakes and depending on how hard you press on the "brake" pedal the higher ratio the cvt has thus the more quickly the energy will be placed into the flywheel. likewise the "brake" pedal could also control how quickly the energy can be displaced back to the wheels. granted this can already be done with the regular transmission.

edit: it's probably best to leave the engine in gear while decelerating
"When coasting with the engine running and manual transmission in neutral, or clutch depressed, there will still be some fuel consumption due to the engine needing to maintain idle speed. While coasting with the engine running and the transmission in gear, most cars' engine control unit with fuel injection will cut off fuel supply, and the engine will continue running, being driven by the wheels. Compared to coasting in neutral, this has an increased drag, but has the added safety benefit of being able to react in any sudden change in a potential dangerous traffic situation, and being in the right gear when acceleration is required."
yes, there will be more drag, but I doubt it would take more energy than the energy used in fuel to keep the engine idling.
The optimal solution would be similar to the VW Lupo (I think) that would on the fly be able to turn the engine off while decelerating and was able to start the engine automatically within like .5 seconds... No reason to run the engine when it is not needed.

Very unscientific (and knowingly incorrect) ROI analysis for a block heater

2009-10-07T00:01:00.003-05:00

I recently purchased a block heater. I bought it for the purpose of increasing fuel mileage and also to have heat working sooner when I drive off in the winter.

To me it seemed like basic physics. An engine runs at its peak efficiency when it is at it's optimal operating temperature which is around 195F. I did not initially do any research as what the ROI would be for this purchase. But I am doing some rudimentary analysis now.

I could not find any ROI analysis for a block heater so I figured I might as well create one. Some of the data I used is incorrect since said data does not exist and I do not have the tools to properly collect said data.

I have also found some data (not really 100% credible, but there is no real reason for them to lie) that suggests that the efficiency gains (in terms of reduced fuel consumption) of a block heater remain true in the summer as well as winter.

The unit cost $30 from amazon.

There are 365 days in a year and I estimate that I will use the heater 95% of the time so about 347 days each year I will use the device. In Illinois electricity is about $0.09/kwh. The device consumes 0.2 kw/h and most forums online say it only needs to be plugged in about 2 hours before driving to heat the engine to an optimal temperature (maximization problem with the curve that I am ignoring). So for each day I use it I will probably only have it plugged in for 2 hours. So it will probably take less than $13/year in electricity costs.

Here is where my data is non existent and I used another model which I assume is similar. I am assuming that driving my moms 2002 chrysler 300m takes about five minutes of driving to reach it's optimal temperature (for efficiency). This may be a bit conservative, it probably take a bit longer.

Then I used http://www.metrompg.com/posts/warm-up-idling.htm to get an average of how inefficient the engine would be for the first five minutes (again, no correct data was available but I can see the numbers on that site being reasonable).
So from their data (1.5+1.1+1+.8+.7+.7)/6=.966 so .5(fuel consumption at optimal temp)/.9666(avg of fuel consumption for first 5 minutes of operating)=.517 so 51.7% efficiency.

My mothers car gets about 25mpg when running at optimal temperature, full tires, 10w30, no wind...

so 51.7% of that is 12.93mpg for the first five minutes of driving. I would estimate my average speed for my first 5 mins of driving is about 40mph=.66 miles per minute. so in 5 minutes I travel 3.33 miles where I am only getting about 12.93mpg.

so 347 days*3.33 miles=1156 miles I was driving per year where I was getting 12.93mpg. With gas at $2.55/gallon and assuming that having the heater plugged in for 2 hours will get the engine to it's optimal temperature where it gets 25mpg it would save about $110 in gasoline per year. This is not subtracting the $13/year in electricity costs nor is it taking into consideration the initial $30 purchase price.

I am too lazy right now to do a breakeven analysis to find out how long it would take to break-even and start saving money.

some other info I would like to take into consideration http://priuschat.com/forums/gen-ii-prius-main-forum/23749-some-hard-numbers-block-heater-operation.html#post296616 possibly later (eg, using the heater I think would only provide 19mpg on average for the first 5 mins thus only $73 saved on gas per year)

This is ignoring the obvious benefit of the comfort of having a warm car right away when you drive off on a blisteringly cold day and also ignoring the benefit it would have on the engine (oil is more viscous and it is a lot easier on the engine to start it warm vs cold thus it will probably last longer). It also ignores the heat dispersion/etc and a bunch of mathematical and scientific stuff for now. It also ignores time value of money stuff. investing an extra $5/month for 25 years at a 10% return would result in almost $7k at the end of the 25 years. Again, this ignores the fact that the car will have a longer useful life (hopefully) and yes I know, it's not mine.

Let me know if you find a proper ROI analysis for block heaters.

Been Awhile

2009-10-06T23:57:00.002-05:00

It's been 8 months since my last post and I feel that this trend of infrequent posts will continue but not to the same extent. Possibly I will update this blog once every 1-2 months or whenever I feel I have something important/interesting enough to post.

I used to post here to keep friends and family up to date and I was posting about quite a bit of frivolous stuff. I have found that twitter helps to fill this niche. I'm still not sure what would be classified as important/interesting enough to post here vs twitter. I guess I will figure it out as I go along.

wow

2009-02-28T16:28:00.002-06:00

http://seattlepi.nwsource.com/local/401779_schene28.html

6'2 195lb cop on "administrative leave" (still getting paid)

1. beats the shit out of a tiny 15 year old girl
2. shoots a mentally ill man 11 times
3. Gets a DUI

omg he was "assaulted" with a shoe...

why did the cop who was with him not arrest the other cop? at the very least he could have used means to stop his partners abusive behavior.
all cops are the same. if there were any good cops, there would be no bad cops.

fucking cowards.

why/why not?

2009-02-17T22:59:00.002-06:00

why bailout the auto industry? If there was value to it, private investors would back it. Why make the taxpayer have to pay for something with no value?

why not cut spending rather than "stimulate" the economy? the only way we are going to pay for this "stimulus" is debt/print money/taxation, all of which will be like not using lube...

while I agree that spending money on infrastructure is great, why not keep it at the state level like it's suppose to be. The fed should not get involved in that.

i could go on... but hopefully this provides *some* clarity.

25 things about me

2009-02-13T09:46:00.000-06:00

1. I get bored very easily causing me to never finish anyt

chromium on debian/os x

2009-02-11T22:35:00.004-06:00

thanks to codeweavers

http://www.codeweavers.com/services/ports/chromium/

works fairly well (well... heh, screenshot might suggest otherwise).

Cannot wait until google finally releases an official port for linux/mac (cannot believe they did not make it platform agnosticish from the beginning).

One thing I hate (but I understand the reasoning, complexity=bugs/problems, but i think soon [probably even now] the benefits will outweigh the costs) about firefox/epiphany/etc is that they are all single threaded and do not take advantage of multicore cpu's very well.

this becomes somewhat of a big deal if you are like me and use google for just about everything (docs/calendar/reader/mail/blogger/finance/search). most of their stuff is very ajaxy and requires a bunch of cpu power (yay for php+javascript+whatever else for ajax to work).

Working on a google docs spreadsheet in firefox is aweful. you see the one core pegged at full load (the one firefox process eatting up 100% making the entire browser unusable) and the other core sitting idle.

Even though firefox 3.1 is much quicker at javascript (on par with chrome) it sucks not being able to do anything in your browser while you are waiting for 10+ seconds for your spreadsheet to finish applying some formula.

chrome is great, and while yes, it takes about the same amount of time to update formulas, I can actually do something else like check my gmail while I am waiting.

I understand it is probably going to be very difficult to make firefox multithreaded, but would it hurt to look towards the future and create a multithreaded branch for developers to start working towards a multithreaded firefox?

All I can say is, thank god for competition. if it were not for chrome, it would probably be a much longer wait for firefox to implement.

and now I go back to accounting homework...

Barack Obama is tired of your motherfucking shit

2009-02-06T15:27:00.000-06:00

not sure if this is real, if it is, lol

Differential Synchronization - google tech talk

2009-02-06T15:21:00.001-06:00

http://www.youtube.com/watch?v=S2Hp_1jqpY8

microsoft, this is fucking brilliant

2009-02-01T04:08:00.001-06:00

Songsmith is creating beautiful art. sheer genius.

billy idol's White wedding interpreted via songsmith

good/bad

2009-01-27T23:48:00.002-06:00

well, cannot get internet access in my room anymore, neighbors ap won't even connect anymore :(
hopefully this awesome neighbor gets it fixed up soonish.

on the plus side I found a faster network (400kbps vs 50kbps), but it can only be accessed from downstairs where it's really cold and I don't have my table/whiteboard/monitor/keyboard/etc

both essid's were "linksys" thus the compy seemed to be confused about which it was connecting to.

and thanks to http://64.223.189.234/node/808 (change settings on router w/o password) I was able to move to the clear channel 1 (probably why I get 400kbps). I am surprised by how huge of an exploit that is, only discovered this past year and so many millions of linksys ap's with the stock firmware from before 2008...

update your firmware if you care about security!!!

i would ask for permission, but i'm not harming anything, and it's their signal on the property. not to mention how weird it would be to go door to door to strangers asking if you can use their wifi to browse the web. might get shot...

dammit, miss my desk so bad. so damn cold...

I know I am retarded

2009-01-26T04:32:00.003-06:00

but is this really all that difficult?

"prove the identity of each of the following boolean equations, using algebraic manipulation:
wy+w'yz'+wxz+w'xy'=wy+w'xz'+x'yz'+xy'z
"
what I have so far (not even sure if it's right)
w(y+xz)+w'(yz'+xy') (the yz' +xy' would both=0 thus drop out)
wy+wxz+w'

any help?

*EDIT
turns out it is very difficult and took our professor ~4-5 hours just to do this problem. nobody else got it.

pie is good

2009-01-24T02:09:00.002-06:00

Proofreading is not a good place to cut the budget

2009-01-24T00:31:00.002-06:00

the lab book for my computer organization class is horrendous.
I have only gone through about 25 pages and have already come across:
1. grammatical errors. This would not be a problem typically (heck I do it all the time), except it is since some key points are uninterpretable.
2. Spelling errors. Not as big of a deal, usually you can figure out what they are attempting to convey.
3. Telling you "the following LED's should then light up" and then being like "wtf, they are not lighting up" only to find out that they tell you 20 pages later how to wire the lights. I shit you not.

The first two I would normally shrug and go on, but every 3rd paragraph contains errors.

these tardbuckets expect people to be able to "easily learn assembly" when they cannot even be bothered to proofread?

What the fuck happened to the world when I was not looking?

*cracks open another beer^
^out of whiskey :(

waiting game

2009-01-22T10:18:00.003-06:00

CH 12 ][ Elapsed: 3 mins ][ 2009-01-22 10:17

BSSID PWR Beacons #Data, #/s CH MB ENC CIPHER AUTH ESSID

00:1F:33:46:8A:76 0 2 0 0 1 54 WPA TKIP PSK TSU1998
00:18:F8:B8:99:29 0 44 5 0 6 54 WEP WEP mike
00:22:A4:20:6F:99 0 41 0 0 7 54 WPA TKIP PSK 2WIRE957
00:18:F8:C0:58:6D 0 350 2 0 6 54 WPA TKIP PSK soling family
00:21:7C:AD:53:29 0 106 0 0 5 54. WPA TKIP PSK 2WIRE942
00:18:39:62:0E:A8 0 1986 23 0 6 54 OPN linksys
00:1B:5B:69:94:D9 0 2242 0 0 6 54. WEP WEP 2WIRE354
00:18:3F:0E:69:E9 0 2219 0 0 6 54. WEP WEP Wilmot
00:18:F8:C3:8E:D6 0 874 2 0 6 54 WEP WEP Black Label Society Network
00:22:6B:02:A1:88 0 2237 0 0 6 54 WPA TKIP PSK @MikeAsh
00:1D:5A:61:D8:41 0 1351 13 0 6 54. WEP WEP PATRICIA
00:1E:E5:B3:02:FA 0 2183 0 0 6 54 OPN @Home02FA
00:18:3F:89:55:C1 0 1515 0 0 6 54. WEP WEP 2WIRE997
00:18:3F:86:CE:51 0 15 0 0 6 54. WEP WEP 2WIRE163
16:3D:9B:02:7C:6D -1 294 0 0 6 11 WEP WEP
00:18:3F:0D:8B:21 0 4 0 0 6 54 WEP WEP 2WIRE515
00:1F:B3:72:D6:51 0 10 0 0 7 54. WPA TKIP PSK 2WIRE405
00:1B:5B:69:97:D1 -1 0 0 0 11 -1
00:18:39:E9:E9:CA 0 13 0 0 6 54 OPN linksys

BSSID STATION PWR Rate Lost Packets Probes

00:18:39:62:0E:A8 00:13:02:A6:8F:9A 0 1- 1 0 26 linksys
00:18:39:62:0E:A8 00:16:E3:C8:05:9C -1 1- 0 0 2
00:1D:5A:61:D8:41 00:1D:7E:05:69:B8 0 12- 1 0 33
16:3D:9B:02:7C:6D 00:00:00:00:00:00 0 0- 1 8 294
(not associated) 00:13:72:C1:D8:31 0 0- 1 0 3 RickJames
(not associated) 00:14:A5:79:5E:47 0 0- 1 0 30 Secret,ConklinRE,linksys,Beezelbub
(not associated) 00:17:3F:FD:14:D7 0 0- 2 0 4
00:1B:5B:69:97:D1 00:40:F4:FF:F7:D7 0 0- 1 0 14

All of these AP's and all are running on overlapping channels except for one. it's like a bunch of people shouting in an echo chamber. the one (one other, on 1, but it's WPA which could take years to crack) that is on a non overlapping channel (11... unique channels are 1,6,11) says
"" which means it has no idea what the essid is (not broadcast). if I wait for a host to connect to it, I should be able to snag the essid (from my understanding, it will be sent in cleartext).

no idea what, if any, kind of encryption there is (please be WEP).

flash 10: min system reqs: 128mb vram

2009-01-18T20:07:00.000-06:00

that's fucking nuts

Ronald Jenkees!

2009-01-07T10:48:00.000-06:00

http://www.youtube.com/watch?v=ZQ-FC3DLKwc

buy his album

Once you go Debian, you don't go back

2009-01-05T06:59:00.003-06:00

Ubuntu is a great distro, and they do put quite a bit of polish into it. But in all honesty, it's basically Debian's Sid (unstable) branch.

Nothing wrong with that, but after awhile, the whole "whoa, new feature" thing gets a bit old and the idea of staying about a year behind so the majority of the bugs can get ironed out kicks in. Bleeding edge is usually not worth it.

Summary:
Ubuntu/Debian Sid: damn whipper snappers. Firm release dates releases often contain critical bugs. releases are more like RC1 releases which they attempt to fix after it is pushed out the door.
Debian Testing: packages that have 1-2 years worth of testing that have (for me) a perfect balance of features, and stability. Very few major bugs.
Debian Stable: 2-3 year old packages, about as close to bug free and stability as you can possibly get. It's ready when it's ready, even if the target release date is missed by years.

conclusion:
choice is good.

web 2.0 ide

2009-01-04T03:12:00.002-06:00

almost everything is being taken online.
see: google docs/google mail/etc

one area that has been overlooked is taking an IDE online.

would be really nice to have the eclipse platform integrated right into google code in a fashion similar to google docs.

code anywhere, seamless collaboration/bug tracking/etc

guess i am not alone in my thinking

2009-01-03T07:54:00.000-06:00

"Word to the wise, if your girlfriend or wife is a nurse and you claim that your engineering degree was harder then their nursing degree because they never took calculus, be prepared to spend the night on the couch. Just a tip.

Still, my $TYPE engineering degree makes me more then qualified to do any profession. Why, with a few books from the library and maybe a couple Google searches I could probably give your friend that kidney transplant they need. How hard could it be anyway, those overpaid doctors never had to work with Laplace transforms!"

-104614

"Well, there is something to what you say. Having worked both in an engineering capacity and as a skilled tradesman, I've noticed that there is a distinct difference between between the two. Doctors and nurses are skilled tradespeople, like highly trained auto mechanics. No one is ever going to ask a doctor to design a better human being, any more than anyone is going to ask an auto mechanic to design a better car. This is not to say that it's easy to be able to instantly recognize the symptoms of disease (x), or the bad interaction of drugs (y) and (z); just that it's not a particularly creative field of endeavor. Engineering and the hard sciences (including programming) are less about being able to instantly reference huge volumes of memorized information, and more about taking a small amount of basic knowledge and putting it together in new ways."

-230075

wow, people really are idiots

2008-12-29T16:38:00.002-06:00

http://change.gov/openforquestions
is full of really bad questions by people who have no idea what the governments role is. very sad.

there are a few good questions that make me somewhat proud (sadly I have not figured out how to link directly to them) like "why are we wasting trillions of dollars occupying countries which we have no business occupying? we are just pissing people off."

or questions regarding the reasoning for marijuana being illegal.

and they are very few and far between.

most are "I want the government to do X for me" X being tell private companies how to run their businesses/give tax credits for things the government should not get involved in... etc

i really hate school

2008-12-28T00:53:00.001-06:00

need to get out of school, but looks like I will be stuck for another 3 or so semesters...

how to get straight A's

2008-12-21T18:25:00.003-06:00

This is how to handle the situation:

Schedule an appointment with your professor(s), then walk into his office (shutting the door behind you) and grab his tie. Yank it down so hard that it chokes him and his head slams into his desk and say with your other fist clenched,

"You punk motherfucker - I'm not going to waste my time on this bullshit and you're going to give me an A as if we didn't have this little discussion, capiche?"

If he says anything other than "yes" then grab his stapler and pistol-whip him with it. Directly after this break into his home and hang his pets from his ceiling fan, but take one of the pets, behead it, and place its head on your professors' bed. Then write, "I see you" above his bed using his favorite pet's blood.

If the plan outlined above doesn't work, you just might have to play hardball.

(somewhat stolen from some anonymous entity).

for gods sake, use generics

2008-12-19T20:37:00.003-06:00

I could attempt to explain it, or I could just sit here and post a link to someone who already has.

really, it's more useful than you might think and few people know how/where it is appropriate to use generics.

having to cast when writing something that is object oriented generally means you are doing something horribly wrong and you need to rethink it.

such is the case for using generics. less possibility of doing something which could get you in trouble later.

oh that is just too awesome

2008-12-19T19:50:00.003-06:00

totally pwns timemachine and is on par with vistas "earlier version" thing.

a lot more polished than timevault (which it seems the developer has not touched in like a year).

apple, please, versioning for backups. you're literally over a year behind windows/linux/solaris