Data-driven ESP modelling and optimisation

• Published in: Journal of aerosol science 2014-04, Vol.70, p.59-66
• Main Authors: Toimil, Daniel, Gómez, Alberto, Andrés, Sara M.
• Format: Article
• Language: English
• Subjects: Aerosols; Chemistry; Colloidal state and disperse state; Data mining; Electrostatic precipitators; ESP modelling; Exact sciences and technology; General and physical chemistry; Optimisation
• ISSN: 0021-8502; 1879-1964
• DOI: 10.1016/j.jaerosci.2013.12.013

The process that takes place in an electrostatic precipitator (ESP) is complex and is influenced by several phenomena. The numerical models present in the literature are continuously growing, but the complexity inherent to the process and the limits of current computers make it impossible to carry out a complete modelling process, being necessary to carry out important simplifications on the models. Taking into account these limitations of numerical models, the use of data mining techniques is proposed for the analysis and modelling of the ESP׳s performance, the advantages of which are the possibility to include a large amount of complex phenomena and factors and their applicability to any ESP, regardless of its specific configuration or shape. This approach is especially interesting for the analysis of ESPs that have already been implemented, from which we can recover data and take into account the environment׳s characteristics. Furthermore, the resulting models can be used in optimisation processes in which the best ESP configuration is sought at all times, in order to maximise its yield. •New modelling approach: analysis and modelling of ESP using data mining.•Comparison between models: numerical and data mining models.•ESP parameter optimisation: using data mining models and metaheuristics.