(79d) Model Enhanced Prediction of Self-Heating in Detergent Spray Dryer Wall Build-up

In the spray drying of detergent powders, it is common for layers of powder to accumulate on the inner walls of the spray drying tower. At the high temperatures at which these towers operate these powder layers have the propensity to self-heat, whereby exothermic reactions occurring within the build-up cause an increase in temperature at the core of these layers. Significant self-heating in these systems can cause the powder to char, leaving burnt particles which compromise the quality of the finished product, while in extreme cases thermal runaway of these layers can occur. Limiting this threat requires a greater understanding of these systems, methods to characteristics these powder, and a means of modelling this behaviour. This work explores methods of characterising the self-heating reaction kinetics and thermal properties of these detergent powders, including a novel parameter estimation approach, and develops numerical models to predict the behaviour of self-heating detergent powder systems, such that a workflow can be developed to limit these problems for both current and new detergent powder formulations.

In characterising these detergent powders, the self-heating reaction kinetics were measured using basket heating methods, namely the steady state approach and cross-point temperature method. In these methods the temperature profiles, as measured by a number of thermocouples embedded in baskets of detergent powder, are used to determine zero-order self-heating reaction kinetics. These methods are well established and have been previously used to measure self-heating reaction kinetics for a range of powdered materials. As part of this investigation these basket approaches have been applied to a number of detergent powder formulations and the measured kinetics compared to those measured using a novel parameter estimation approach.

This parameter estimation approach is performed using a 2D transient model of an oven heated basket of powder developed in gPROMS ModelBuilder. This model captures the heat and mass transfer in an equi-cylindrical basket of self-heating detergent powder. The model predictions are compared to a series of experimental basket temperature profiles, measured by an array of thermocouples embedded in the basket of powder. The prediction error is minimised to determine the best-fit self-heating reaction kinetics, thermal conductivity, and specific heat capacity. The correlation of these parameters is explored, and to improve this approach, Differential Scanning Calorimetry (DSC) and the Ozawa-Flynn-Wall approach has been applied to determine the reaction activation energy of the powder prior to fitting. The advantage of this approach is that fewer experiments are required compared to the steady-state method, and this approach has been found to be less sensitive to thermocouple errors than the cross-point temperature method.

The reaction kinetics found using this novel parameter estimation approach compare well with those from the other basket heating techniques. Using the best-fit parameters the 2D model was successfully validated against critical ambient temperatures (temperature above which thermal runaway occurs) of oven heated baskets of powder, and the influence of different conditions on the self-heating behaviour of these baskets explored. This model and the parameter estimation approach provide a viable and more efficient means of characterising current and new detergent powder formulations. Adapting this model to represent self-heating in tower wall accumulations, the influence of tower temperatures, airflow patterns, layer thickness, and other process conditions on the critical ambient temperature and optimal tower operating temperature of different formulations has been explored.