Crystallization

Simulation and modeling methodology

"PSE's expert ModelCare consulting support allowed us to develop a state-of-the-art model for a complex crystallisation process with minimum time and effort"

Steve Pietsch, Senior Research Associate, BP Chemicals

Deliverables

The deliverable from the work described here can take one of several forms:

A customised and validated executable model of one of PSE's "off-the-shelf" solutions, such as a lactose crystalliser
A custom-built executable model for your particular process, based on PSE's AML models
A 'configurable' model of either of the above, for which you can – for example – modify the growth kinetics.
A report describing the optimal crystalliser design, the optimal batch recipe, or suggested operational improvements or any other objective.
ModelCare assistance to your internal modelling efforts
Transfer of crystallisation modelling know-how to your internal teams, to help develop an effective modelling culture within your organisation.

ModelCare Objectives

ModelCare is PSE's unique way of providing consulting. Key objectives are:

Rapid project delivery
Fit-for-purpose models
Effective use of data
Integration of R&D and Engineering work
Transfer of modelling know-how to customer personnel

The modelling of crystallisation processes can result in significant benefits.

However, it presents several challenges for companies investing in building, validating and executing crystallisation models:

the complexity of typical crystallisation processes requires a sound understanding of physical and chemical phenomena and their mathematical representation
Advanced Process Modelling expertise is required to fully customise first-principles models and validate these against experimental data.
engineering judgement and experience is required to determine the validity of data and identify any additional experimentation requirements
the time constraints and workloads under which customer plant, engineering and R&D personnel typically operate limit the time for these activities.

For these reasons, typically the most cost-effective and time-effective approach is for PSE to work collaboratively with customer personnel to deliver a project under a ModelCare agreement.

PSE's crystallisation modelling service

PSE provides a comprehensive crystallisation modelling process. A typical 'complete' project that includes six steps:

1. Consultation. We sit down with the 'owners' of the business, process and technology (as appropriate) and discuss business and technical requirements and how (and whether) these can be achieved.

For a FREE initial consultation to scope out whether modelling can help with your requirements, please contact PSE Consulting.

2. Proposal and work plan. Once the requirements are clear, we formulate a proposal and work plan.

This answers questions such as: How much will this cost? How long will it take? How should the workload be split between PSE and your personnel to minimize cost, maximize speed of delivery and technology transfer?

3. Model build. This typically involves customisation of models from PSE's AML:SC, to include appropriate nucleation, growth, agglomeration and attrition kinetic submodels as well as suitable properties and thermodynamics.

We will also construct any upstream and downstream flowsheet elements that need to be taken into consideration.

4. Model validation and analysis of data fits. Typically we work with your personnel to validate the model and analyse the data fit.

With the correct data, this results in accurate values for kinetic and other constants, as well as confidence information.

Where necessary, we will identify any further data requirements and help design the necessary experiments for maximum effectiveness.

5. Execution, case studies, optimization and implementation. Once the model is fully validated, we will either provide it to your personnel to execute cases, or do this for you if preferred and provide a report.

A typical application is the use of optimisation techniques to calculate objectives such as optimal batch policy directly.

If necessary, we will install the model at your site for further execution.

6. Transfer of know-how. Depending on how steps 1 to 5 are executed, we will be transferring modelling know-how to your personnel over the duration of the project.

If required, we can augment this with on-site training in further use and extension of the model.

Typical project times

Projects can often be executed in a much shorter time than is commonly believed.

A typical time for developing a validated model of an uncomplicated system is:

Initial data analysis 2 – 5 days
Model construction 5 – 10 days
Validation 10 – 25 days
Execution and reporting 5 – 15 days

Note that project times can differ considerably depending on the quality of available data.