A Data-Driven Fault Prediction Method for Nuclear Power Systems Based on End-to-End Deep Learning Framework

Table 2

Simulation industry conditions.

ID of IC	Class of IC	Sampling period/tensor size
IC1∼IC10	(1) Steady-state	0.1 s/
IC11∼IC40	(2) Power changes within normal range from steady-state	0.1 s/
IC41∼IC70	(3) Two main pump rotors are stuck	0.1 s/
IC71∼IC100	(4) One main pump rotor is stuck	0.1 s/
IC101∼IC130	(5) Two pumps of the secondary circuit are stuck	0.1 s/
IC131∼IC160	(6) One pump of the secondary circuit is stuck	0.1 s/
IC161∼IC190	(7) Condenser completely loses heat removal capability	0.1 s/
IC191∼IC220	(8) Control rod withdraws by mistake	0.1 s/