The first phase (0,10s) (chart 02) allows verifying that the initial system remains stationary. This phase could be skipped.

Second phase (10, 20): break actuation with full feedwater flow.
The power qcn1 extracted by each sg increases rapidly.
The step 4% subcriticality margin applied at 10 s diminishes rapidly but remains positive.
Nuclear power f2c remains negligible, but the thermal power q2c is positive and represents the fuel latent heat transmitted to the rapidly cooling water (chart 03, chart11)

Phase 3 (20,86 s): feedwater flow is reduced to its auxiliary level: cooling rate is attenuated (chart 03), qcn1 decreases.
This, pure reactivity, transient phase without feedback terminates, just like in any fast reactivity release accident, such as rod ejection or fast rod withdrawal, when the core approaches prompt-criticality (rc= beta=.0076) (chart 02)

Phase 3 (86, 100 s) is typical of f2c and q2c evolution in fast reactivity transients: f2c peak is checked by Doppler feedback (chart 03), thermal power lags by some 20 s) (chart 02). qcn1 increases again to extract nuclear contribution of q2c.
In the course of this period, core power dramatically shifts towards the bottom (chart 04), as highlighted by the rapid jump of core axial offset to -1, and hot channel factor fqc increases accordingly.

Phase 3 (100, 600) this phase extends from max q2c to moment where the sg starts to be voided (chart 07). The min safety margins related to u2c7 (hot pellet center temp) and dnbr (chart 05) are not observed at q2c peak but only 50 s later, as power profile shifting develops.

Phase 4 (600, 2000).
At end of phase 3, both qcn1 and f2c decrease. Power profile shifts back to the top.
Boron injection (chart 06) brings the core again sub-critical as transient terminates but has no effect on the safety related peaking values, because it comes too late.

Pressu (chart 08) is voided very early in the accident and starts refilling at phase 4 only.
Establishment of the late low power, critical phase 4 raises an issue of accident termination management rather than safety, because the min margins are observed much sooner.
In addition, this long process towards transient termination necessitates a long simulation time.