Mechanistic Model of Coke Development in a Coke Oven Situation

The standard approach to evaluate the potential of coals or blends to create good quality coke is to use laboratory tests (dilatometer, plastometers, sole heated ovens etc). However these tests have limitations. This project combined the various approaches into a 'whole of oven' model with models of post re solidification shrinkage.

• Coal carbonisation into coke is modelled within an industrial-class Computational Fluid Dynamics (CFD) platform
• A significant component of the project has been technology transfer that has implemented cutting-edge coking sub-models, based on those developed previously, and incorporated them into established engineering toolboxes. This increases the portability and flexibility of the model, opening the door to many extensions of direct industrial benefits.
• The model is extended and refined using bench dilatation and temperature measurements. Model testing against sponsor's pilot oven tests demonstrates that it performs reasonably well against the experimental data provided, and has revealed new insights into heat, moisture and volatile movement in gas ovens.
• Coke micro-structure time-space development and the full carbonisation cycle thermodynamics are now accessible from the 3-D simulations. Predictions of internal gas pressure and lump size distribution are available and have been confirmed using experimental results. To our knowledge this is the first time mean coke size and more importantly, coke size distribution, have been able to be predicted from such a simulation.
• The CFD platform is modular, allowing different modelling variations to be included in simulations (material thermo-properties, carbonisation chemistry, condensable phase change, micro-structure, etc.) and expanded as knowledge of the processes increases.

Coking models available in the literature describe the carbonisation, micro-structure formation, and internal gas pressure rise processes using heavily idealised configurations and often with ad-hoc correlations. The CFD approach used here makes a significant step forward because it aims to fully reflect the design, operation, and inherent multi-physics taking place in a coking oven.

The project has provided for assessment a series of coal/blend samples properties with corresponding dilatometer and pilot run data. This CFD model recovers temperature variations as measured inside the pilot coking chamber. Coke structure formation within the load is reasonably predicted by the model and visualised through local representative cavity diameter, porosity. Moreover, using an improved fissuring model, estimates of average coke lump size and even coke size distribution were found to agree with measured values. Internal gas pressure associated with this micro-structure development is consistently recovered as well.

Implementation of coking sub-models into an industrial-class simulation platform facilitates technology transfer to the industry because of its versatility to embrace real designs and its capability to interface with other engineering toolboxes. For instance, it would be possible to extend the solution domain to include the flue channel. By addressing both the coking chamber and heating channel in a single model, coking battery core functionality could be captured.

The simplified approach for changing the industrial designs used in the model opens the door to:

• Cost-effective exploration of coking facility breakthrough new designs;
• Assessment and trouble-shooting of sub-optimal operations;
• Predicting changes in coke size distribution from changes in coal properties and coke oven operation.

The current simulation capability will benefit from further fundamental investigations in future projects on coking coal because many physical and chemical models have room for refinement; given the limited current knowledge on the underlying coal carbonisation processes.