Research Article

A Denoising Based Autoassociative Model for Robust Sensor Monitoring in Nuclear Power Plants

Table 3

Results of DAASM and K-AANN performance metrics.

Performance
metric
Model typeFeed flow 1Feed flow 2STM flow 1STM flow 2STM PSR 1STM PSR 2STM PSR 3PZR PSR 1PZR PSR 2PZR LVL 1PZR LVL 2SG LVL NR 1SG LVL NR 2
AANNSp-DAANN

Accuracy (% span)0.3820.3610.3430.4120.1860.2110.1660.3860.4110.2430.2230.5120.621
0.2810.2530.2460.2930.1210.1320.1220.2430.3150.1730.1560.3940.465

Autosensitivity0.2730.2940.3210.3320.2050.1820.1980.2530.2330.2250.2310.2680.301
0.1530.1670.1870.1630.1120.0790.0820.1310.0710.140.1170.1330.187

Cross-sensitivity0.1280.1240.1130.1160.1120.1080.0920.0790.0830.0980.110.1030.105
0.06320.05270.05630.06140.0390.03110.02870.01860.02030.03440.03360.030.0286

SPRT detectability0.3120.3150.330.350.040.050.040.080.0860.10.090.060.058
0.1840.1650.190.20.0210.0230.020.0480.050.040.040.0320.028