Batch process optimisation

Modelling software specifically designed for batch

The application of model-based techniques to batch processes can result in millions of dollars of per year in enhanced revenues with little or no capital expenditure.

"Process modelling using gPROMS identified a 30% saving in batch time." Senior Polymer Technology Manager BASF Polymers

Modelling can be used to minimise batch time, maximise product recovery, determine the optimal feed profiles for a preferred process trajectory (the “golden batch”, and many similar applications.

Even in a relatively 'simple' batch process such as the one shown below there are many decisions that can significantly affect profitability.

However batch processes are inherently dynamic systems, and cannot be adequately modelled by traditional process simulation tools.

PSE's gPROMS Advanced Process Modelling package was specifically designed to handle the onerous modelling requirements of batch and semi-batch systems.

The specialised requirements of batch process modelling

In the design and operation of a “simple” batch process, there are many questions to answer – and the only way to answer them easily is by modelling.

gPROMS provides the following unique combination of capabilities:

Powerful and robust dynamic modelling, an essential requirement for operations in which there is no steady state.
A comprehensive task language, used to model complex operational sequences
Dynamic optimisation, a powerful technology capable of optimising transient processes over time subject to process and equipment constraints.
High-fidelity model libraries of typical equipment used in batch processes – reaction, crystallisation, separation and reactive distillation.
A powerful custom modelling capability, to model non-standard unit operations.
The ability to execute online as part of a Model-based Predictive Control (MPC) or real-time optimisation (RTO) application.

Example 1: Optimisation of a Di-Octal Phthalate (DOP) process

Optimisation using gPROMS doubled the profit margin on this plant.

Batch flowrate

The flowsheet shows a Mitsubishi Chemical batch process for manufacture of di-octyl phthalate (DOP), a plasticiser widely used in the automotive industry.

Because of the competitive market the profit margin for the operation is relatively low, and even a small improvement in process efficiency can significantly increase profitability.

As shown in the diagram, there are many possible optimisation variables. Some are constants – for example, reactor size – and others time-varying – for example, steam supply rate.

These were all defined within a gPROMS dynamic optimisation, along with the process constraints. The objective function was to minimise the batch time required to reach the DOP specification.

The plots on the right show the input trajectories (upper) for the fresh reactants, recovered 2EH and steam required to achieve the optimal temperature and product composition profiles (lower). logo

The gains in efficiency from the optimisation resulted in a doubling of the profit on the process.

Example 2: EPS Batch polymerisation

By building a high-fidelity detailed kinetic model of its batch Expanded Polystyrene (EPS) process and applying dynamic optimisation techniques, BASF was able to identify a 30% reduction in batch time. Batch process model

The first-principles gPROMS batch process model (left) included detailed reaction kinetics, with parameters estimated from experimental data.

In addition it modelled heat and material balances, geometry details, transport and thermodynamic properties (calculated using the PC-SAFT equation of state) and plant operating procedures.

Dynamic optimisation was then used to minimise batch time taking into account process constraints.