TY - GEN
T1 - Classification of UO 2 green pellet quality using intelligent techniques
AU - Sutarya, Dede
AU - Putro, Benyamin Kusumo
PY - 2012/10/15
Y1 - 2012/10/15
N2 - Modern production facilities are large and highly complex, and they operate with numerous variables under closed loop control. In the production of green uranium pellets, pellet quality control involves many variables. Therefore, the classification of the quality of pellets is important for improving the efficiency of the production process. Classification of pellet quality using the conventional graphical method has some drawbacks; for example, the scale of the graph affects the accuracy and ease of use. In this paper, intelligent techniques are used to classify the quality of the pressurized water reactors(PWRs) green pellets into three categories according to the guidelines in the quality control manual of the experimental fuel elements laboratory of BATAN. Four features are used for classification, namely, height, volume, weight, density and theoretical density of the pellets. A dataset (150 observations) was collected from one lot of compacted UO 2 pellets and was used for training and testing of an ANFIS model. Up to 86.27% of the data can be classified correctly using the ANFIS model. Such performance is comparable to that of artificial neural networks. Thus, this model can be applied effectively for the evaluation and classification of pellet quality.
AB - Modern production facilities are large and highly complex, and they operate with numerous variables under closed loop control. In the production of green uranium pellets, pellet quality control involves many variables. Therefore, the classification of the quality of pellets is important for improving the efficiency of the production process. Classification of pellet quality using the conventional graphical method has some drawbacks; for example, the scale of the graph affects the accuracy and ease of use. In this paper, intelligent techniques are used to classify the quality of the pressurized water reactors(PWRs) green pellets into three categories according to the guidelines in the quality control manual of the experimental fuel elements laboratory of BATAN. Four features are used for classification, namely, height, volume, weight, density and theoretical density of the pellets. A dataset (150 observations) was collected from one lot of compacted UO 2 pellets and was used for training and testing of an ANFIS model. Up to 86.27% of the data can be classified correctly using the ANFIS model. Such performance is comparable to that of artificial neural networks. Thus, this model can be applied effectively for the evaluation and classification of pellet quality.
KW - Adaptive neuro-fuzzy inference system
KW - Artificial neural networks
KW - Quality classification
KW - Uranium dioxide pellets
UR - http://www.scopus.com/inward/record.url?scp=84867273077&partnerID=8YFLogxK
U2 - 10.4028/www.scientific.net/AMR.557-559.2054
DO - 10.4028/www.scientific.net/AMR.557-559.2054
M3 - Conference contribution
AN - SCOPUS:84867273077
SN - 9783037854570
T3 - Advanced Materials Research
SP - 2054
EP - 2064
BT - Advanced Materials and Processes II
T2 - 2nd International Conference on Chemical Engineering and Advanced Materials, CEAM 2012
Y2 - 13 July 2012 through 15 July 2012
ER -