The paper focuses on the problem of energy-saving control of a large low-pressure air separation plant with a turbo-expander. The plant is designed for simultaneous production of nitrogen, oxygen and argon and operates at variable performance over a time interval when the need for air separation products is repeatedly changed. The purpose of the research is to improve the efficiency of the air separation plant, in terms of minimizing energy consumption for air separation over a time interval, using destabilizing optimization algorithms for the operating modes of its rectification subsystem, the latter being the core of the proposed plant. The study introduces the author's "concept of destabilization", which consists in expanding the range of feasible solutions to the problem by replacing in its statement some constraints like equalities with inequality constraints. This is technologically equivalent to purposefully changing some parameters within acceptable limits instead of stabilizing them, which provides an additional effect over a time interval. The authors developed methods and algorithms for controlling an object over a time interval, which provide for all possible changes in the performance of various target air separation products with a linear dependence of the optimality criterion on additional control actions resulting from destabilization. A specific example demonstrates the use of destabilization control algorithms for solving the problem of optimizing the modes of the rectification subsystem of the air separation plant over a certain time interval with an estimate of the effect obtained. Using the proposed approach and making structural changes to the existing equipment of the rectification subsystem of the air separation plant, we can significantly increase the resulting effect. Destabilization can be fully applied to other complex technological objects with liquid capacities and when a change in the liquid levels is acceptable within certain limits. Further research is needed to find out how plant performance and the time of the plant constancy influence the degree of the resulting effect, as during the plant operation the need for separation products, the moments of their switching and the time of constant performance intervals repeatedly change.
