Maybe a wheel rim generator, along with a battery, and some kind of charge controller.

Another possibility would be a flywheel coupled via a CVT (continuously variable transmission).

A very special and elegantly simple pancake motor/generator, a temporary electrical storage device, and a controller. ----- That's kind of what I had planned.

I've never used this manufacturer and this is NOT an endorsement.

http://www.goldenmotor.com/

I only suggest you review the site for ideas to get your project started.

Good Luck!

Thanks !! Great link. Regen capable motors already available!!! Now I just need to solve the slow acceptance problem in order to increase regen efficiency!!!

Gav

I see this application also involves a motor driven bicycle (I was only thinking of generation of electric current using kinetic energy from the moving cycle) As someone who bought his wife an "e-bike" I must say to please be sure that the drive wheel is on the rear. My wife has had several nasty wrecks (much road rash) from small amounts of gravel using a front wheel driver. A front wheel with torque input creates very erratic handling when the tire breaks loose on a curve.

Yes Phys; that is indeed the goal. In an electrical version you would be recovering the kinetic energy and re-launching using a motor/generator. In a nut shell what you would have is a human/electric hybrid. The key is keeping it at safe voltage, simple, and efficient.

Another possible approach is a human/hydraulic hybrid.

When you look at the physics, once the system is charged to its operating energy, mass becomes less of a factor if the regenerative braking is at high efficiency.

As in all self-contained regenerative capable transportation systems the key to efficient regenerative braking is storage acceptance rates; perhaps more conventionally called - power density in acceptance. Common re-chargeable electric storage devices have relatively high power density in discharge but very low power density in acceptance; in a nut shell they can't accept the power that is generated in fast braking from high speed.

An efficient storage device would have to accept power at .5m(v^2) /t = fs/t where m is system mass, v is maximum recoverable velocity, t is braking time, f is the maximum friction force between the wheel and road, and s is braking distance.

Because the energies are relatively low in a human powered bicycle it becomes a bit less problematic but still challenging; especially the power density in acceptance issue.

I have some ideas in motor/generator design, storage, and control. Perhaps I will take the time to put them out into the public domain; but for a guy like me who doesn't know how to use CAD everything has to be done with pencil and paper.

There is a very high probability that there is nothing new here in terms of component design; just a matter of finding the right off the shelf technologies.

I think you may be overestimating the importance of regen braking.

Energy recovered from braking is not a large percentage of the energy used during travel. Most of the energy expended goes into overcoming mechanical and air drag. Neither of these can be recovered. I've run many calculations/simulations and cannot recover more than ~15% energy using "ideal" regen braking. It is important to note that this % only occurred during heavy stop & go travel. Continuous point A to point B travel basically gets "0%" energy recovery from regen braking.

Assuming it is done correctly and with reasonable cost, regen braking is just one small step towards improving overall efficiency. Reduction of rolling resistance and air drag losses are usually more important. Motors and controllers are already very efficient, but may still have some room for improvement. Battery technology is where we could use some significant advances.

Ideally we want batteries that have the following characteristics:

small
light (low mass)
capable of fast charge/discharge
safe (non-toxic, non-explosive, non-combustible)
long life
low cost

The ideal battery doesn't exist... yet. For now, lithium technology seems best. Cost is coming down and performance is incrementally increasing.

I suggest you work out some calculations using realistic losses for mechanical drag and air drag so you don't overestimate the significance of regen braking. Learn where all the energy goes and you have a better chance of increasing efficiency and improving the whole EV as a system.

Best wishes :-)

MJB;

Thank you for that very comprehensive reply.

Since you have done many calculations regarding kinetic energy recovery I have some questions.

If a car has a combustion energy to work efficiency of 25 percent how many units of joule combustion energy must be used to generate 1 joule unit of kinetic energy?

If I am able to recover 1 unit of kinetic energy and reuse it to relaunch the vehicle how many units of combustion energy have I saved?

Is it possible that Prime Mover to work efficiency is a very significant variable when considering regenerative capable systems?

Was that factored into your calculations?

Gav

Chemical to mechanical energy conversion of 25% is a reasonable working value for an Internal Combustion engine.

>Assuming 25%, you need 4 joules of fuel energy for every joule of mechanical energy.

>Theoretically, 1 unit of recovered kinetic energy saves 4 units of fuel energy.

>If you start with an initial primer mover efficiency of 25%, then (regen_braking_energy/total_system_energy) is far less than the 15% I'm calculating.

>Your post is titled "Building a Regenerative Capable Bicycle" and my simulations are for "electrified bicycles" so IC efficiency isn't a factor. For E-bike sim I use theoretical battery capacity and motor+controller efficiency of 90%. I further assume that the motor_controller_battery system can recover 90% of the kinetic (braking) energy. I'm also using classic air drag force calculations with a small additional loss for rolling resistance. This is intentionally optimistic.

I'm NOT claiming regen braking is useless. As previously stated, it is just one factor we can implement to increase overall system efficiency. I AM claiming that I see too many people, who have not thoroughly reviewed all the system losses, overestimating the importance/significance of regen braking from an Engineering/Physics perspective. Of course regen can be a superb marketing ploy if sales/profit is the primary goal.

I've reviewed other literature and empirical data regarding recovered braking energy and see values ranging from "0%" to "~17%". My sim numbers are in line with published research.

Please try to calculate the energy needed to overcome rolling resistance and air drag over a given distance and then compare it to the kinetic energy recovered during braking.

MJB;

Thank you again for your comprehensive and thoughtful replies.

It was a breath of fresh air to see someone who understands how prime mover efficiency is a primary factor in the economics of regenerative braking.

In your E-bike sim; can you describe how many stop and go cycles you used for the given displacement?

For the bicycle commuter this would be the most significant variable; more significant than the road drag or even the aerodynamic drag given the low velocities and small aerodynamic cross-section.

The work done in any transportation cycle would approximate the (aerodynamic drag X displacement) + (road drag X displacement) + (∑ KE dumps) + (mgΔh).

What was your value for the coefficient of dynamic drag? What was your aerodynamic cross-section, operating velocity, and total system mass?

The ratio of KE recovery to total energy consumption may define the fraction of energy recovered; but without factoring the system efficiency into the equation it really is a meaningless ratio.

If you are attempting to calculate energy savings then the ratio should be Kinetic and gravitational energy recovered /(applied energy/input energy) = energy savings; where applied energy (work) / input energy (energy used) is the efficiency fraction of your process.

I AM claiming that too many people who analyse Kinetic Recovery systems fail to factor prime mover efficiency and underestimate the number of stop and go cycles.

Did you write your own algorithms? I don't suppose I could get you to share your formula for road drag with us could I?

Gav

For any practical round-trip bicycle travel I'm going to claim mgh recovery is pointless. The extra energy used up-hill is saved during a coast down-hill without any regen capture losses. I'll consider mgh important if your travel destination is another planet.

For my e-bike sim I use a rolling loss of 2 Watts/meter/second. This is my estimate with no reference for confirmation.

Other calculation refs:

http://en.wikipedia.org/wiki/Drag_coefficient

http://en.wikipedia.org/wiki/Density_of_air

http://en.wikipedia.org/wiki/Work_(physics)

For a human adult on a typical street bicycle, I use a drag coefficient of 1 with a frontal area of 0.75 square meter. Average travel speed is 15 miles/hour (~6.7 meters/second) and bike+rider mass is 125kg. I just ran a quick sim for a 1 hour trip at 15 miles/hour with 20 stops. This would be a realistic ride cross city for my location. Best case energy recovery from regen braking is just under 10%. Significant, but not a large factor.

I suggest anyone interested should attempt to setup and run all the calculations themselves so they understand all the variables and constants. I strongly suggest that anyone attempting this should understand the proper use of "units". They should carry units fully through ALL calculations to verify their results.

I disagree on a couple points. If regen energy is a small portion of an ideal system, it is an even smaller part of a lossy system. I'd claim that improvements are needed to make the system less lossy first. Regen brake energy takes a lower priority until the other more important factors are addressed. As an example, the focus on improving the energy efficiency of an IC vehicle system should start with the engine conversion efficiency, then drag reduction, rolling resistance reduction, and lastly braking energy recovery.

Thanks for the further information.

Let's take one part at a time.

You stated - "For any practical round-trip bicycle travel I'm going to claim mgh recovery is pointless. The extra energy used up-hill is saved during a coast down-hill without any regen capture losses."

The above would only apply if there were no braking required to control speed on the downhill side. Gravitational potential is a very significant factor; especially when you are the power source.

You stated - "If regen energy is a small portion of an ideal system, it is an even smaller part of a lossy system."

Yes the ratio of kinetic recovery to energy use is smaller in a lossy system; but again, are you factoring applied power efficiencies into the whole energy savings equation?

Regenerative payback increases as the ratio of applied energy to consumed energy decreases not the other way around. In a single stop the energy savings approximates the energy recovered/ efficiency fraction of the power process. No --- this is not over unity; it is simply factoring the power process losses.

You stated - "For my e-bike sim I use a rolling loss of 2 Watts/meter/second. This is my estimate with no reference for confirmation."

It might be valuable to revisit this. Although road drag is somewhat dependent of velocity, it is not an increasing linear relationship as your calculation infers. If I am reading this right you are estimating road drag to be consuming about 13.4 watts at 15 mph; and half that at 7.5 mph.

Based on the sim found here - http://www.exploratorium.edu/cycling/aerodynamics1.html your road drag equation consumes nearly half of total power at 15mph. Is it possible that this is significantly overstated?

What is the total displacement in the 1 hour period? Since average speed in constant acceleration is only ½ that of the final velocity is it possible that the air pressure drag may be significantly overstated in your calculations? Is one stop every 5 minutes a realistic scenerio for an urban commuter in busy traffic?

Having said all of this; even a 15% increase in process efficiency is HUGE!!!

You stated - "For any practical round-trip bicycle travel I'm going to claim mgh recovery is pointless. The extra energy used up-hill is saved during a coast down-hill without any regen capture losses."
The above would only apply if there were no braking required to control speed on the downhill side. Gravitational potential is a very significant factor; especially when you are the power source.

I agree that energy to maintain velocity on positive grades IS significant. However, mgh energy recovery is dependent on geography. On a typical bicycle commute I would not benefit from the process. My one small grade change does take more effort uphill, but the free 25mph coast (no braking necessary) on the way down is a nice respite.

You stated - "If regen energy is a small portion of an ideal system, it is an even smaller part of a lossy system."
Yes the ratio of kinetic recovery to energy use is smaller in a lossy system; but again, are you factoring applied power efficiencies into the whole energy savings equation?
Regenerative payback increases as the ratio of applied energy to consumed energy decreases not the other way around. In a single stop the energy savings approximates the energy recovered/ efficiency fraction of the power process. No --- this is not over unity; it is simply factoring the power process losses.

We have different priorities here. Since drag accounts for most of the energy consumed, aerodynamics would be my first target for improvement. If the cost, complexity, and reliability are acceptable, I'd also implement regen braking to recover some energy.

You stated - "For my e-bike sim I use a rolling loss of 2 Watts/meter/second. This is my estimate with no reference for confirmation."
It might be valuable to revisit this. Although road drag is somewhat dependent of velocity, it is not an increasing linear relationship as your calculation infers. If I am reading this right you are estimating road drag to be consuming about 13.4 watts at 15 mph; and half that at 7.5 mph.
Based on the sim found here - http://www.exploratorium.edu/cycling/aerodynamics1.html your road drag equation consumes nearly half of total power at 15mph. Is it possible that this is significantly overstated?

My results indicate power to maintain 15 miles/hour is about 160W or 4x what their calculator indicates. I couldn't find their values for the drag coefficient or the frontal area. As stated on the linked site, a racing cyclist will modify body position (head-down-crouch) to reduce the drag coefficient and frontal area. Each one of these could be halved (Cd=0.5 and A=0.375), resulting in the 4x difference in their calculator. I'm intentionally NOT using racing class cyclist numbers for city commuter applications.

A premium racing bike with thin high pressure tires will have lower rolling resistance. On my pot-hole ridden city streets, these beautiful bikes wouldn't last long. My calculations are based on the knobby fat lower pressure tires similar to those found on mountain bikes. While the rolling resistance is higher, these tires are better suited for the simple commuter application I'm investigating.

Rolling resistance is a complex process since the coefficient actually varies with velocity and several other factors. My approximation is reasonable for my current use. When I can gather some empirical data, I will attempt to refine my approximation.
http://en.wikipedia.org/wiki/Rolling_resistance#Rolling_resistance_coefficient

This calculator appears to generate some interesting values:
http://www.mne.psu.edu/lamancusa/proddiss/bicycle/bikecalc1.htm

Is one stop every 5 minutes a realistic scenario for an urban commuter in busy traffic?

My conditions were:
1 hour = 60 minutes
60 minutes/20 stops = 1 stop every 3 minutes
This would be realistic in my area.

Since average speed in constant acceleration is only 1/2 that of the final velocity is it possible that the air pressure drag may be significantly overstated in your calculations?

From velocity errors, no. I get up to speed in 10-12 seconds. Short quick stops (5-8 seconds) are realistic for city traffic. The 180 seconds between stops is at the stated velocity of 15 miles per hour. My simplified calculation is optimistic in that it assumes no drag or rolling resistance during the acceleration/braking periods.

What is the total displacement in the 1 hour period?

Using 15 [miles/hour] and 20x 20[s] acceleration+brake periods, ΔX~14.2 [miles].

Having said all of this; even a 15% increase in process efficiency is HUGE!!!

It seems we don't share the same frame of reference, especially for the definition of huge. I don't consider 15% here "huge".
Good luck with the huge energy savings and I hope your project is hugely successful.

Thanks for the great Bicycle Power Calculator link!

I ask you to consider the variable power required to do your ride.

Now imagine a system where the sum energy of the system remains nearly constant over the entire cycle if the Prime Mover puts in "Average Power." Perhaps now that hill doesn't't look so long on the uphill side.

I like your analysis of your commute; probably not much different for millions of folks on this planet who commute by bike.

When I have refined my description of the process to the extent it would support a ruling of "public disclosure" I will get it out.

Gav

You are going to take all the fun out of cycling. I ride around 4000 miles/year, so I know a bit about cycling. I, and others I ride with, look forwards to long fast scary descents. Other than the occasional very busy highway intersection (a bit of a rarity in my remote location) on many road rides I never touch the brake levers except for the mid ride break or final stop. The whole premise of the other side of cycling, mountain biking, is to climb as high as possible so you can fly back down the mountain trails. Now you are going to have the cycle try to use up a portion of that energy to generate electricity? Won't go over with the typical cyclist. When we point the bike downhill we want to go too fast.

The goal is not electrical generation - it is Kinetic Energy Recovery.

Renerative braking is of no advantage in a transportation cycle where there is no stopping.

But to a commuter in Amsterdam or any other major city where the transportation cycle includes stop and go driving, a regenerative capable system would significantly reduce the energy input required to get from home to work. I recently visited Amsterdam and when I asked a rider there what she thought of the idea of being able to recover the the kinetic energy for relaunching the bike she thought it was a great idea.

I guess desirability is a matter of application and taste.

Such a system would also have the benifit of encouraging more people to ride.

Also, the American experiance with bicycles is not what the rest of the worlds is.

Improve someone else's efforts?

A regenerative capable Bicycle that uses:

A low voltage single sprocket gear Pan Cake motor/generator rear axle.

A single gear crank set.

Four modes of mg operation.

1. Coasting

2. Infinitely variable transmitted power from the rider.

3. Infinitely variable electric/human transmitted power from storage and rider.

4. Regenerative Braking at high efficiency

Electrical Storage - Low Voltage Super-caps by Maxwell