For the modeling and evaluation of fluid systems, continuous Petri nets offer a suitable formalism. Continuous places, modeling fluid storage, can be connected using continuous transitions, which model fluid transportation with their specified nominal flow rates. During the evolution of such systems, the fluid levels in continuous places evolve continuously, and places with less in- than outflow can get empty. To remain physically meaningful in such conflicted cases, the semantics reduces the outflow rates to match the inflow, in a process called rate adaption. For the computation of the rate adaption results during simulation or analysis, the state-of-the-art approach is to reduce the outflow rates for single conflicted empty places in a certain fixed order. However, the termination and physical meaningfulness of this greedy rate adaption method could be guaranteed only for a strongly restricted model class. In this paper we (i) develop a formal framework for the underlying concepts of rate adaption, (ii) present a generalized iterative rate adaption algorithm, (iii) prove that under some mild conditions, the computations always converge to unique results for the whole considered model class, and (iv) provide an implementation and some experimental results to demonstrate applicability.