Banadda E N, Smets I Y, Jenné R, Van Impe J F
BioTeC - Bioprocess Technology and Control, Chemical Engineering Department, Katholieke Universiteit Leuven, 3001 Leuven, Belgium.
Bioprocess Biosyst Eng. 2005 Aug;27(5):339-48. doi: 10.1007/s00449-005-0412-6. Epub 2005 Jul 15.
The performance of the activated sludge process is limited by the ability of the sedimentation tank (1) to separate the activated sludge from the treated effluent and (2) to concentrate it. Apart from bad operating strategies or poorly designed clarifiers, settling failures can mainly be attributed to filamentous bulking. Image analysis is a promising technique that can be used for early detection of filamentous bulking. The aim of this paper is therefore twofold. Foremost, correlations are sought between image analysis information (i.e., the total filament length per image, the mean form factor, the mean equivalent floc diameter, the mean floc roundness and the mean floc reduced radius of gyration) and classical measurements (i.e., the Sludge Volume Index (SVI)). Secondly, this information is both explored and exploited in order to identify dynamic ARX and state space-type models. Their performance is compared based on two criteria.
（1）将活性污泥与处理后的出水分离；（2）浓缩活性污泥。除了不良的运行策略或设计不佳的澄清池外，沉降失败主要可归因于丝状膨胀。图像分析是一种很有前途的技术，可用于丝状膨胀的早期检测。因此，本文的目的有两个。首先，寻求图像分析信息（即每张图像的总丝状长度、平均形状因子、平均等效絮体直径、平均絮体圆度和平均絮体约化回转半径）与经典测量值（即污泥体积指数（SVI））之间的相关性。其次，对这些信息进行探索和利用，以识别动态自回归外生（ARX）模型和状态空间型模型。根据两个标准对它们的性能进行比较。
