Case studies for selective agglomeration detection in fluidized beds: Application of a new screening methodology

We have recently presented a new methodology for screening different signal analysis methods in combination with signal pre-treatment methods with the goal to effectively identify those combinations that are highly selective towards a specific process change (Bartels et al., Ind. Chem. Eng. Res. 48 (2009) 3158-3166). The main outcome of the methodology is visually represented in an overall result matrix with coloured tiles illustrating a measure for the suitability of each combination of analysis method and signal pre-treatment. Suitable methods can be visually identified very quickly. For the early detection of agglomeration in fluidized beds we illustrate this methodology by four different cases: two cases from a pilot-scale bubbling bed, one from an industrial scale bubbling bed and one case from a lab-scale circulating bed. With the result matrix for each case several suitable methods are identified. The data are also evaluated to identify methods that are more generally applicable for a range of different cases. The suitability of a positively identified method is subsequently analyzed for its temporal response to both agglomeration and other effects. The influence of the different data pre-treatment methods is also addressed. (C) 2010 Elsevier B.V. All rights reserved.