Study on Predicting and Warning of the Wet Flue Gas Desulfurization System Slurry Poisoning Based on Improved K-means Clustering

To address the problems of increased load of the wet flue gas desulfurization (WFGD) system, decreased WFGD efficiency, and shortened equipment service life caused by slurry poisoning, this paper proposes the optimal clustering center selection strategy based on k-means clustering and the differential threshold method. This paper constructs a normal operation space of the WFGD system, and to achieve real-time early warning of slurry poisoning faults in the WFGD system. With the combination if mathematical abstraction and mathematical analysis of business experience, our experimental results show that the warning duration of the model is 88.75% within 4 hours before the slurry poisoning fault occurs, and 98.32% when the slurry poisoning fault occurs. No alarm was generated for a stable time after the slurry poisoning fault was repaired. The constructed health state space of the WFGD system can effectively cover the health state of the system and has a good prediction ability for slurry poisoning failure.