Any ideas for an embedded renewable/energy project?

Maybe create some sensor boards to monitor central heat/A/C and it's motors/compressors. Another for your refrigerator and it's compressor. Track, set warning thresholds etc. for out-of-norm operation. Maybe add 'docking' of your car and upload data from it at the end of everyday?

Or create a lawn watering system that tracks wind/humidity/insolation and ground moisture. Maybe find some way to add precipitation forecasts to aid in its decision making.

Alex expounded in news:4d220f9d-0536-46e5-8265-e2e28b2fba50 @v41g2000yqv.googlegroups.com:

One idea I've had, would be to use a MCU to control the input to an automotive alternator, used in a wind generator. When the wind dies down below a threshold, you want to cut the rotor input power, since you're not going to get anything back. When it picks up, apply some input power to help it "generate".

Beyond the threshold test, the MCU could adjust input power levels to the rotor to maximize the output generation. If the wind is weak, then you need to supply less of a field. But to get more out of a strong wind, you could add more rotor input power.

Now of course you need to store that "juice" in batteries, and having your alternator generating peak output will probably be bad for your battery(ies) if they're fully charged already. So some sort of battery management would also be required.

Warren

To avoid this hassle, wind turbines often use permanent magnet rotors. :-) It's also more power efficient (no exictation) and less wear (no brushes) Car alternators can be modified to permanent magnet.

Yes, battery management is a must, that could be your project.

My inner Diogenes wants to say that any manager who's ditzy enough to think that working on some random "green" project is going to qualify you to work on his _specific_ "green" project is a manager who's too ditzy for me.

But maybe you have to get your resume through HR first, or you just want to do a "green" project.

Isn't there any product you've always wanted, but never had?

I'd like a programmable house fan that comes on whenever the inside of my house is farther away from some target temperature than the outside of the house, to provide mostly-passive climate control. Basically it would be something that mechanizes "open the windows tonight, dear, and we won't have to run the air conditioner tomorrow".

I think they're made already, but there's little that you could do that's not made already, so you can't let that be a barrier.

I'd like an intelligent lawn watering system.

Something that would sense soil moisture content, ambient temperature and humidity, and the presence of a car in the driveway and then adjust sprinkler on time and duration appropriately.

Better yet, detect if the guy in the car is a traveling salesman, or "the other man" and if so, turn the sprinkler on just as he gets out...

How about a contraption that squeezes the maximum power out of solar panels by setting the best operating point on the current insolation's UI curve?

Hello Victor,

do you happen to have links to whatever is currently considered the best algorithm(s) to achieve this? Thanks M'

of solar

I think the simple way is by changing the current of the panel up or down, measuring the resultant U and I, comparing U*I to the previous value and then either changing the direction of the next delta_I or not. Basically, you keep making small hops along the UI curve. That's all I know.

I've wanted to do this for ages, just haven't found the time.

I've started on just such a project, here. Xeriscape so no "lawn", per se. And, I've never been able to come up with a reliable "moisture detector" given the alkalinity of the soil, here (a friend suggested a potentially viable sensor but I think it would be too expensive to fabricate in small quantities).

The larger problem is sorting out evaporative/transpirational losses and how they correlate to environmental factors. My current plan is to let the system "watch me" for the next year or so (weather variations in the five seasons) and *try* to correlate my actions with observations of the environment. I don't expect to be able to "fit curves to the data" but hope there will be enough correlation that I can write an expert system to meet those goals.

I wonder if you could watch the field in the windings to detect RPM (without having to add a tacho). Use that as a "wind sensor" (instead of watching the actual output of the alternator). You could also use this to protect the rotor in *excessive* wind (think microbursts, hurricane, etc.) -- perhaps "dynamic braking" to lock the rotor in place (I suspect this would be less stressful than running the risk of spinning a bearing, etc.)

I think a more realistic project might be a controller to dynamically fit a (charger) load to the output of a solar panel. This is "existing technology" yet shows a variety of issues that have to be addressed in the design of the hardware and software to do so economically and reliably.

But, you can do many of these things *better* (or, at least,

I think a better approach is to design a system that washes the car when it determines the lawn is in need of watering...

ant

nd

I thought about that.

But then I would have to move the car to the lawn. Moving in is easy; Moving out would be tough with wet soil. Assuming that the wheels won't kill the lawn in the first place.

I can rebuild my drive way to catch and redirect the water.

Or I can build a lift to move the car between the lawn and driveway.

Nevermind, too much work to save some water.

(sigh) Sorry, I guess my humor was missed (think: "it always rains just after you wash your car...")

I hope the system is intelligent enough to turn the water off during the day and only operate during the night.

When using a sprinkler during a hot day with low relative humidity, will evaporate some of the water before it hits the ground.

During the night, the water will actually penetrate the ground (and not evaporate directly from the ground) and actually reach the roots of the grass.

Here (desert southwest), the evaporative losses from classic "sprinklers" are touted as *40%*!

Stef expounded in news:b1298$4c73dd0d$54f63171$ snipped-for-privacy@abuse.newsxs.nl:

The MCU makes it a non-hassle ;-) The downside is the brushes naturally, but the upside is that you can vary power output with the strength of the wind. IIRC, you can get up to 60 Amps from a little alternator, perhaps more with some models, though probably not continously.

The trick is to monitor the rpm and not to exceed maximum continous ratings. If you increase the field, does output continue to increase (while maintaining rpm), or does it tend towards stalling? If it tends to stall, you back off the field and acheive a compromise for maximum generated power.

Putting a permanent magnet in, just destroys all the fun!

Warren

D Yuniskis expounded in news:i53bvq$sbs$ snipped-for-privacy@speranza.aioe.org:

Even though the alternator rectifies the current, it is not pure DC. So the charging current will look like a series of bumps. Feed that into a opamp circuit and shape that into something a MCU can count.

This is the "whine" component of the alternator. At least this is how I would attack it, as it requires no addional parts for monitoring (rpm).

I think there are many smaller lighter components that would better serve as a sensor. You could probably make one out of an old fashioned (ball) mouse, that has those optical encoder wheels in them. Attach that to a tiny plastic fan.

Related to this idea, I have another rolling around in the back of my mind:

- Provide "load tests" and measure batter performance on loads over time (until they are exhausted).

- Feed results into a PC (or memory stick), for later performance analysis.

The idea is to test various brands of batteries under "load". A simple voltage test is a "no load" test (the volt meter is virtually no load at all).

Then you could plot "bang for your buck" charts for batteries by brand and price. Then you can answer questions like:

If I spend extra on the "copper top" am I getting more than if I go cheap on some "no name" brand, which I know will run out sooner?

I'd also perform these tests on a light vs heavy load basis, so you can answer the right question for the right application. Presmably the bang for the buck would be similar, but the test would answer the question "are they the same?"

Light loads might be for electronic toys, calculators etc. while the heavy load performance is important to toy cars etc., that tend to eat batteries.

For load testing, you need a power transistor activated by the MCU periodically. Turn on the one with the appropriate load resistance attached.

After this, you could publish your findings for the rest of us. ;-)

Warren

Most alternators hav a 'W' contact. This is connected to the AC side of the rectifier, giving you a higer amplitude rpm signal.