Dear Edwin,
There are many claims in your paper that need to be discussed, but perhaps it is best to start with the main one. You claim to have produced a local theory that nonetheless predicts violations of Bell's inequality. But the theory simply does not appear to be local in Bell's sense. Of course, if one enforces certain global conservation restrictions on a system, that will have consequences for what is observed. The perfect anti-correlation between results for spin measurements in the same direction on particles prepared in the singlet state, for example, is predicted by enforcing 0 net spin for the system. But if each particle is not in a state which predetermines the outcome of the experiment, and is completely unaffected by whatever distant experiment is carried out, then enforcing the global conservation means that theory is not local in Bell's sense.
Let's put is more directly. Suppose that a system has two, widely separated parts and I carry out experiments on the parts at space-like separation. And suppose that as a consequence of carrying out an experiment on one part, the energy of that part changes. If I now enforce global energy conservation, so the energy that disappears from one side must appear on the other space-like-separated side, then the theory is not local in Bell's or Einstein's sense. Calling this non-local interaction between the sides "energy exchange" does not change this: energy exchange between space-like separated subsystems is a violation of locality. So it is not a surprise that your model can generate violations of Bell's inequality: it is not a local model.
The claim that your model is local is critical part of the paper: if it were true, then you wold have shown that there is some flaw in Bell's reasoning. As for your remarks on Bell's argument, you seem to have mistaken an illustrative example that he gives for part of the theorem itself, which it is not. The theorem is about any theory--whether the theory uses quantum-mechanical formalism or a completely different formalism in which there is no talk at all of eigenstates or eigenvalues--that makes certain predictions about correlations between outcomes of distant experiments. Because of this complete generality, it is not even correct to us the term "hidden variables" to describe the theorem, since that term itself is used only in connection with quantum theory. Peres, who you cite, has it right here. Since Bell's theorem only refers to the results of experiments and their correlations, he makes no assumptions at all of any kind about the theory predicting those results, save that it is local. His theorem does not apply to "energy exchange physics" because in this setting the energy exchange would not be a local process, and the theory would not be local. It is therefore not a counterexample to Bell's theorem.