Modelling powder flow
Accelerating powder process design using discrete element modelling
Modelling powder flow within bespoke engineered parts
CPI have developed a discrete element model to describe powder flow within a bespoke-engineered interface, which collects granulated products as they are manufactured by a twin-screw granulator.
This model was used to predict the flow around redesigned powder diverters that could be inserted into the interface to prevent powder from aggregating in unwanted areas and ensure even distribution of sample across the PAT sensors. These diverters were successfully manufactured and inserted into the interface, alleviating the powder flow problems. This modelling capability can be applied to de-risk new designs by evaluating their performance prior to fabrication.
An input is a unique capability, service or method of support that was provided, such as equipment or expertise.
- Technical expertise.
- Access to state-of-the art facilities and bespoke engineered equipment.
- Investment fully funded via an Innovate UK grant as part of CPI’s Formulation Strategic Projects Programme.
An output is the result of the work, such as an experimental finding, an actual product or a pilot demonstrator.
- A discrete element model that identified the cause of poor flow to the bottom sensor
- Optimised geometry of powder diverters to improve powder flow through the PAT interface
- Modelling the system including the diverters gave confidence that this change would solve the problem
- Validation of the model by comparing the prediction to real-world images showed a high level of predictive accuracy
An outcome arises from implementing the outputs, for example, a profit, an investment, providing jobs or delivering societal benefits. Outcomes continue once an innovation has been implemented and deliver benefit every year.
- The optimised diverters were manufactured and inserted into the device, solving the problem and allowing sufficient powder flow to all sensors
- Facilitated further development of the PAT interface due to improved function and access to the third sensor
- Developed a modelling capability which can be applied to de-risk new designs by testing and optimising their function prior to fabrication
Designing powder processing plants and operations within formulated goods poses significant challenges for companies of all sizes. While large companies can rely on historic data, coupled with the ability to do extensive pilot, and scale-up trials, small and medium sized enterprises (SMEs) often do not have the necessary time and resources to de-risk their designs. Hence, companies must often proceed with designs of equipment and operations with a high degree of technical risk, or rely on tried and trusted designs, which may work but are sub-optimal from a quality, safety, or operator perspective.
The ability to make predictions about how a piece of powder processing equipment will perform, before committing to purchasing this equipment or fabricating the design, is therefore highly pertinent to SMEs. Over the past 20 years, computational fluid dynamics has seen widespread adoption by industry in predicting liquid and gaseous flows and is now more accessible to SMEs via commercial software packages. However, approaches for modelling the flow behaviour of powder and granular systems has not yet reached this level of industrial readiness.
CPI possess a bespoke-engineered powder flow interface that collects granulated products as they are being manufactured by a twin-screw granulator. The quality of this granulated product is measured via various PAT sensors that feedback into the manufacturing process. Hence, the flow of powder down the device is vital in ensuring that the granules are presented to the sensors in an optimal way. While the device was designed based on many years of experience and best practices, certain elements of the design did not allow it to function optimally, and powder was not reaching the final sensor.
How CPI helped
A particulate modelling approach was performed on the PAT interface which showed that while an abundant amount of powder was reaching the top two sensors ports, the geometry of the device was preventing powder from reaching the bottom sensor. It was speculated that flow diverters may improve powder flow to the bottom sensor, hence a suitable design for these was produced and their result in relation to powder flow was modelled again. Within the model, these diverters significantly improved the powder flow behaviour in the device, and this enhanced geometry now provided a good flow of powder to all three sensors.
Following the confidence gained from this modelling activity, the powder diverters were manufactured and inserted into the interface. As in the model, these changes significantly improved powder flow and supplied each sensor with an abundant quantity of granules. Comparison of the model predictions and real-world flow shows how accurately the prediction is able to match the performance of the actual device.
While this example shows how existing equipment can be modified to improve performance, the same approach can be used to ensure that new designs are tested and evaluated prior to fabrication. Utilising process modelling of operations from an early stage can ensure that sub-optimal designs are not delivered, which can improve process performance and reduce the likelihood of blockages or downtime.