축산폐수처리를 위한 연속회분식반응기의 이상진단 알고리즘 개발

초록

Recently, the importance of detection technique is increasing to prevent from process deterioration and this paper proposed a knowledge-based diagnosis system using Dynamic Time Warping(DTW) algorithm with ORP(Oxidation-Reduction Potential) and DO(Dissolved Oxygen) values which usually applied as control parameters in swine wastewater treatment. A semi-real scale SBR (effective volume, 20m3) for the treatment of swine wastewater was auto-controlled using real wastewater in situ. The react phase in SBR was performed by sub-cycle operation consisting of repeated short cycle of anoxic-aerobic step, where the aeration step for nitrification was optimized with ORP control. Using ORP and DO profiles which were extracted from sub-cycle, the SBR status divided into normal case and abnormal 6 cases such as controller malfunction, influent disturbance and instrumental trouble. Through the warping process between 7 reference profiles and 28 operation profiles, difference values(D) were determined and classified to 7 cases. As results, in spite of mis-classification in high loading rate(F2), ORP profile showed good diagnosis results. However, DO profiles detected 5 mis-classification which indicated quite different status. Consequently, it was founded that ORP profiles can be more efficiently applied than DO as diagnosis parameter and DTW diagnosis algorithm can be applied to stable SBR operation for the swine wastewater treatment.

키워드