Component failure non-stop control method for molecular sieve oxygen generation system

The invention discloses a component failure non-stop control method for a molecular sieve oxygen generation system, which comprises the following steps of: creating a digital twin equivalent to a molecular sieve oxygen generation system entity in a digital space, performing simulation analysis on the molecular sieve oxygen generation system, and performing part failure non-stop control on the molecular sieve oxygen generation system on the basis of a simulation analysis result. A molecular sieve oxygen generation system control model is constructed based on a deep reinforcement learning technology to perform feedback control on a molecular sieve oxygen generation system with component faults, so that the whole molecular sieve oxygen generation system is ensured to run without shutdown under the condition of reducing part of oxygen generation amount, and the oxygen generation concentration can reach the national standard (93% +/-3%); the MBTF of the molecular sieve oxygen generation system is increased, and th