MOPP. Made To Order Process Plants (TSB/Perspective)

Project: Research

Description

"New reactor technologies are set to change the operation of batch manufacture in the process industries into the new
wave of semi-continuous Make to Order Processing Plants (MOPP). These have the potential to transform these sectors
by reducing the environmental burden, inventories and cost of manufacture and distribution. This project develops an
adaptive 'Dial a Product' control system to deliver the precise control required for these unique high value low volume
manufacturing systems. Bringing together control design and analytical techniques to complement these reactors will
enable the system to reach optimum performance and have commercial impact. The solution will offer the chance to
change the very way the industry operates. Instead of investing in a number of product specific batch reactors, a single
reactor will be used for a number of applications, allowing companies to reduce CAPEX or even work on a rental basis,
bringing in reactors as required."

Key findings

"The Make to Order Processing Plants (MOPPs) project was conceived to deliver the full benefits of advanced process control to the new generation of continuous reactor/crystalliser systems that have recently been developed for use in High-Value/Low-Volume manufacturing. This report describes the technical delivery of the project, broken into Work Packages (WP). It describes both the work performed to develop transferable control systems on both continuous reactor and continuous crystalliser systems. It then goes on to describe the application of these control methodologies on a second reaction in the same reactor system and on a second crystalliser for the same product for the crystallisation system.

The main technical challenges faced in this project were:
Integrating automation systems onto the reactor rigs, allowing computer operated control to be implemented on such systems
Gaining an understanding of the processes involved in transferring a reaction from the batch to continuous domain and defining a suitable, repeatable work procedure for the efficient implementation of this work
The transfer of control concepts from the time profile basis used in batch reactions to the spatial profile present in continuous reactors.
The integration of on-line PAT into these reactor systems and the challenges of moving from their use in batch where their output is available through all stages of a reaction to continuous, where only point measurements are available.

The major outcomes from the project include:
Model Predictive Control system and work flow on a continuous flow reactor that has been shown to be transferable between reaction types.
Control system on Model Predictive Control system and work flow on continuous oscillatory baffled crystallisers that has been shown to be transferrable across different crystallisers that are producing the same product.
A work procedure for the efficient transfer of crystallisation systems from the batch to continuous domains.
A software tool that combines and automates two of the most time consuming activities performed during the development of a new reactor and its associated control system. These are the performing of a DoE trial and the controller development step test.
A methodology to convert process from batch to continuous or produce new on continuous.
Continuous, automated systems to measure MSZW and solubility curves
Automated control of continuous oscillatory baffled crystallisers
Automatically adjust processes to react to changes in both conditions and feedstock quality in real-time.
Utilisation of assay information and in-line PAT measurements to assess process performance.
Reduced need for staff to perform low value-adding work (manually adjusting process, reacting to basic events).
Can automatically move process from product one grade to another without the need to stop the process and re-configure reducing waste and energy use.
Speeds up development time and Scale-up.
Perform automated DoE runs by feeding in a DoE plan and moving the process automatically between the respective factors to deliver significant development savings.
Information outputted by one control system can be used as input to next, allowing the process to be adjusted in a pro-active rather than reactive manner.
Proven ability to run for 5 days and 26% higher yield (vs batch) in continuous crystallisation in COBC improved yield
2 fold reduction in span of PSD vs stirred tank reactor higher quality
Reduction in Lactose crystallisation time from 16 hour to 5 hours intensified process
A successful predictive model for concentration control in COBC system
Successful transfer of concentration control strategy from one crystalliser to the other"
StatusFinished
Effective start/end date29/04/1330/04/15

Funding

  • EPSRC (Engineering and Physical Sciences Research Council): £79,217.00

Fingerprint

Processing
Control systems
Crystallization
Predictive control systems
Model predictive control
Instrument dials
Industry
Batch reactors
Feedstocks
Process control
Assays
Automation
Solubility
Controllers
Costs