Bayesian estimation theory was used in this study to establish a state-space model of the ceramic filter tube degradation process. The model continuously integrates the latest residual pressure drop to update its own parameters, then the change rate of the degradation state, remaining life, and failure probability density distribution of the tube in real time. The residual pressure drop of the Shell Coal Gas Process was analyzed to find that the model results converge to real values as the residual pressure drop increases. The change rate of the ceramic filter tube degradation state calculated by the model gradually decreases over time, which is consistent with the initial rapid increase in residual pressure drop followed by a slower increase in later stages of operation. The amount of particle deposition in the ceramic filter tube wall under different operating times was measured and predicted; the predictions are consistent with the state-space model results. The state-space model also reflects variations in filter tube performance degradation caused by emergent conditions such as leakage or fractures, as it does not make stationarity assumptions for the degradation process.

本研究采用贝叶斯估计理论建立陶瓷滤管退化过程的状态空间模型。该模型不断集成最新的残余压降更新自身参数，然后实时更新管子退化状态、剩余寿命和失效概率密度分布的变化率。对壳牌煤气过程的残余压降进行了分析，发现模型结果随着残余压降的增加而收敛到实际值。模型计算出的陶瓷滤管退化状态的变化率随着时间的推移逐渐减小，这与残余压降的初始快速增加随后在运行后期缓慢增加是一致的。测量和预测不同操作时间下陶瓷过滤管壁的颗粒沉积量；预测与状态空间模型结果一致。状态空间模型还反映了由泄漏或破裂等紧急情况引起的过滤管性能退化的变化，因为它没有对退化过程做出平稳假设。