electromagnetism - Putting far more mechanical energy into a generator than its electrical load

Funny how I was pondering about this question for a long part of the day yesterday, and today I was at stackoverflow and in the box with the featured questions from other SE sites this very related question appeared: Will a 3 phase induction generator fight its own rotation?. So, I decided to ask my related question here.

I know that a generator will fight its own rotation, but what I am wondering is this: suppose I put far more mechanical energy into a generator than its electrical load; what is going to happen then? Picture this scenario: we have a 1KW generator hooked up to a 100W lamp and a watt-meter. The generator has a crank. If I put about 110W of mechanical energy on the crank, the watt-meter will tell me that the generator supplies my lamp with about 100W of energy. (Assuming 10% loss.)

But now suppose that I put a lot more power into the crank; Suppose I go close to the 1KW mark. (Without reaching it, or exceeding it, so as not to burn my generator.) Where will all the excess energy go?

And one more thing: If the answer is by any chance that the generator will only resist the crank by 110W of mechanical energy, so putting 1KW of energy will be impossible, and I will just spin the crank so fast that I will either reach my own limit of how fast I can spin, or the generator will disintegrate due to centrifugal force, then please answer this to me: Is it possible to build a generator that will provide as much resistance as necessary to the mechanical force driving it so that the generator will consume only as much mechanical energy as required by its electrical load?

Let me tell you the reason why I am asking so that you can have more context to take into consideration, if needed, to answer the question: I was wondering about the possibility of storing energy in a mechanical arrangement such as a coil. So, on windy days, when my windmill power generator has surplus energy to give, I would be redirecting the surplus energy to a motor which winds a pretty huge coil, while on windless days I would allow the coil to slowly unwind, driving a generator. (There is, of course, a gear between the coil and the generator, so that the coil unwinds very slowly, while the generator turns fast.) Now, the coil will have a certain amount of mechanical energy stored in it, and it will want to always apply the same amount of torque to the generator, regardless of how much energy my household requires. So, I was wondering if there is anything that can be done to prevent the coil from needlessly unwinding, other than applying a break on it. (A solution which sounds very bad to me.) The ideal would of course be if the generator by itself could resist the unwinding of the coil in such a way that the tendency of the coil to turn minus the resistance of the generator would at all times equal the exact necessary force to turn the generator to produce the exact amount of energy required by its load. But I do not know whether this is electrically possible.

UPDATE Okay, it turns out that the answer to the first question in bold is what the sentence immediately following it anticipates, (the one which says "And one more thing: If the answer is by any chance that...") so the real question then becomes the second question in bold. Unfortunately, this was left unanswered. I was thinking about it, and I came up with something that according to my simplistic understanding of electricity and electromagnetism might (just might) constitute a proof that it should be possible to achieve what I am looking for, so let me explain it, and you tell me if it is correct.

Suppose that I hook up not only a generator on the coil, but also a separate motor. (Which I will have to do anyway in order to be able to wind the coil when my windmill has excess energy.) Now, suppose that the coil has so much mechanical energy stored in it that it wants to exert so much torque to my generator that it would be capable of producing 10KW while I only need 110W. So, suppose that I redirect the excess 9890W to the motor, which counters the torque of the coil. Wouldn't then this entire system reach an equilibrium? Wouldn't the coil end up unwinding at a rate which would correspond to the number of watts required by the lamp, plus about 10% of that for losses of the generator, plus about 10% of the 10KW for losses of the motor? So, wouldn't I then be losing about 1.1KW of mechanical energy stored in the coil instead of the whole 10KW?

And another question: if the answer to this is yes, then is it possible to have a single device which acts both as a generator and as a motor and achieves this behavior without having to have the generator and the motor as separate devices?