Comparison of Machine Learning techniques for atmospheric pollutant monitoring in a Kraft pulp mill

Scientific Conference (2011, November)

In this paper, machine learning techniques are compared to predict nitrogen oxide (NOx) pollutant emission from the recovery boiler of a Kraft pulp mill. Starting from a large database of raw process data ... [more ▼]

In this paper, machine learning techniques are compared to predict nitrogen oxide (NOx) pollutant emission from the recovery boiler of a Kraft pulp mill. Starting from a large database of raw process data related to a Kraft recovery boiler, we consider a regression problem in which we are trying to predict the value of a continuous variable. Generalization is done on the worst case configuration possible to make sure the model is adequate: the training period concerns stationary operations while test periods mainly focus on NOx emissions during transient operations. This comparison involves neural network techniques (i.e., static multilayer perceptron and dynamic NARX network), tree-based methods and multiple linear regression. We illustrate the potential of a dynamic neural approach compared to the others in this prediction task. [less ▲]

Supervised learning for a Kraft recovery boiler: a data mining approach with Random Forests.

in Favrat, Daniel; Maréchal, François (Eds.) ECOS 2010 Volume IV (Power plants and Industrial processes) (2011, January 11)

A data mining methodology, the random forests, is applied to predict high pressure steam production from the recovery boiler of a Kraft pulping process. Starting from a large database of raw process data ... [more ▼]

A data mining methodology, the random forests, is applied to predict high pressure steam production from the recovery boiler of a Kraft pulping process. Starting from a large database of raw process data, the goal is to identify the input variables that explain the most significant output variations and to predict the high pressure steam flow. [less ▲]

Performance monitoring of an industrial boiler: classification of relevant variables with Random Forests

in Pierucci, S.; Ferraris, G. Buzzi (Eds.) 20th European Symposium on Computer Aided Process Engineering (2010)