Identification of Industrial Furnace Temperature for Sintering Process in Nuclear Fuel Fabrication Using NARX Neural Networks

Table 1

Loss function (LF) and AIC, BIC, and MDL criterion obtained by trial and error with difference NARX neural network structure and number of hidden layers for constant heating rate step.

Hidden node number	Model order 𝑛𝑎𝑛𝑏𝑛𝑘	LF	Criterion
			AIC	BIC	MDL
5	219	2.6𝑒−05	−5.2402	−5.2103	−5.1653
	222	6.5𝑒−06	−6.6331	−6.5931	−6.5332
	235	2.0𝑒−05	−5.5014	−5.4515	−5.3766
	311	6.2𝑒−06	−6.6774	−6.6375	−6.5776
	328	8.6𝑒−06	−6.3369	−6.2870	−6.2121
10	219	3.3𝑒−05	−5.0209	−4.9910	−4.9460
	221	1.2𝑒−04	−3.7181	−3.6781	−3.6182
	234	5.3𝑒−06	−6.8339	−6.7840	−6.7091
	316	3.3𝑒−05	−5.0067	−4.9668	−4.9069
	327	1.5𝑒−05	−5.7774	−5.7276	−5.6527
15	217	3.4𝑒−06	−7.2667	−7.2367	−7.1918
	226	7.9𝑒−06	−6.4328	−6.3929	−6.3330
	239	8.5𝑒−02	2.8593	2.9092	2.9841
	317	3.3𝑒−03	−0.3955	−0.3556	−0.2956
	337	3.3𝑒−05	−4.9836	−4.9337	−4.8588