Combining Neural Methods and Knowledge-Based Methods in Accident Management
Figure 4
Modelling example: in this case all these multivariate methods indicate that process variable is in a good calibration. Correlation between model and measured values is even better before than after the calibration. x-axis: time in days. y-axis: interpretative variable 211K122 (reactor pressure). Calibration limits (5%) in this case are MLR: 0.8218, PCR: 0.8164, PLS: 0.821.
