Application of Bayesian Networks for Diagnosis Analysis of Modified Sequencing Batch Reactor

The diagnosis analysis of MSBR has remained difficult due to the complexity of biological reaction mechanisms and the involvement of highly non-linearity and uncertainty. In this paper, the Bayesian network was used to modeling and diagnosis analysis of a MSBR. The suggested BN model for MSBR was evaluated using one-year of operation data. Results showed that the BN-based model for MSBR is reasonable and the prediction analysis algorithm is feasible. According to the framework of the diagnostic analysis, an example was given to illustrate the detailed information of diagnostic of a MSBR. Experimental results indicated that the corrective measures based on the diagnostic analysis were reasonable.