Process analysers - pushing for the standardisation of devices

Paul Boughton

Oliver Schmitt, founder member of the OPC Foundation's Analyser Device Integration working group and Business Line Manager with Malvern Instruments, looks at what lies behind this initiative.

In February this year, the OPC Foundation released the draft Analyser Device Integration (ADI) specification intended ultimately to provide a common method for data exchange for analyser data models applied to process and laboratory analysers1.

Even before the current economic challenges became so acute, manufacturing industry was under pressure to both drive down production costs and improve efficiency. Operational excellence, in its broadest sense, is the goal: consistent operation of the plant at an economically optimal operating point.

Under these ideal circumstances energy consumption and waste are minimised, throughput is increased and storage costs are lowered (produce to order).

Fully automated process control is increasingly being recognised as critical to realising this goal.

In-house studies by various major manufacturers have shown that investing in the right process analytical and control technology can reduce production costs by as much as 10 per cent. The estimated payback periods for many of the projects considered are in the order of two to 12 months, underlining the potential for economic gain.

Success, however, requires the integration of process analysers from a variety of different vendors, in ways that deliver effective data transfer and control.

It was with this goal in mind that a variety of interested parties came together to address issues of standardisation, the roots of the initiative lying in the pharmaceutical industry.

Opportunity

Spiralling R&D costs and an aggressive generics industry are squeezing pharmaceutical revenues and providing an imperative for more efficient, cost-effective, development and manufacture. The US Food and Drug Administration's Process Analytical Technology (PAT) initiative and the Quality by Design agenda (ICH Q8/9 and 10) are catalysts for change, encouraging the industry to focus on processing in a way that it has not in the past.

Globally, pharmaceutical manufacturers are working to transform production practice, adopt automated control and move from batch to (semi)- continuous processing and real-time release.

The scale of change makes it economic to invest in solutions that will ease this transition.

Concern about the difficulties involved in the adoption and integration of new PAT solutions was a significant driver for the establishment of the OPC Analytical Device Integration (ADI) working group, an initiative launched by ABB. The standards that this group develops will drive down the cost of installing validated systems.

For other industrial sectors the work is equally relevant, and valuable, since it potentially swings the cost-benefit argument towards automation.

Towards integration and automation

Off-line analysis, which in some cases remains the only method adopted or even available, is useful for both QC and development but is less suitable for process control. The routine is to extract a sample from a stream or unit, analyse it in a laboratory and store the results in a database that is usually not connected with the control system, or indeed the operational environment. Laboratory Information Management Systems (LIMS) are helping to connect lab and process, but this is not a standard.

Analysis every one to two hours is not unusual and it takes time to make a measurement and report results. The received data therefore reveal how the plant was operating some time ago (perhaps 30 minutes or more), not what is happening as any change is made. This lag makes it difficult to take action in a timely way, hampering process control. Quite large operational changes are made relatively infrequently so the process runs within wide operating margins.

The past 10 to 20 years have seen a shift towards closer integration of the analytical and operational functions, primarily as a result of the increasing availability of analytical instruments for the process arena. At-line systems are rugged enough to sit within the operating environment and offer streamlined, often automated, measurement that de-skills analysis. Results may be fed directly to the plant control platform. Such systems bring an analytical regime under the direct control of the operational team making it more responsive to process needs. However, when the aim is optimal control, frequency of analysis and time lag remain issues.

In-process analysers deliver continuous real-time measurement, which affords valuable insight into process dynamics. Demanding requirements are placed on these instruments, since they must be fully automated, highly reliable and capable of delivering relevant data in a timeframe that matches the demands of the process.

Although this is technically challenging, analytical solutions are now commercially available for many process variables. Particle size, for example, is a critical product performance parameter for many processes and products, and on-line laser diffraction analysis is a well-established tool.

Real-time measurement in isolation is not sufficient to achieve demanding process control targets. Installing in- or on-line analysers provides continuous monitoring of a single variable, but effective plant control demands the instantaneous evaluation of all critical parameters. Failure to properly integrate analysers makes it easy for 'data islands' to develop, whereby individual parameters are closely monitored but information is not linked for maximum usefulness, so optimal control is not achieved.

The availability of continuous analysis for many parameters is prompting manufacturers to critically evaluate which variables to measure, and, equally importantly, how to use the resulting data to drive the plant towards more efficient production.

Multivariate process models are increasingly becoming the norm. Data from sensors of different types are fed into a mathematical description of the process which produces outputs to drive controlled parameters (see box story 'Granulation').

Such an approach requires linking sensors from a range of instrument suppliers in ways that allow effective data transfer. Different analysers produce different types of data, including complex arrays and structures.

Currently, integrating such analysers in a single control system requires manufacturers to develop a dedicated solution capable of taking data from various sources. A number of companies have released control software to do this but these packages have to use the proprietary drivers of the analyser. There is no standard data model. Standardisation will make it much easier to maintain and expand control software, reducing the cost of project implementation.

The OPC Foundation is "dedicated to ensuring interoperability in automation by creating and maintaining open specifications that standardise the communication of acquired process data"2. The OPC ADI working group currently has 21 members including representatives of leading manufacturers (users of process instrumentation), instrument vendors and automation companies. Its goal is to develop a model that will allow equipment vendors to produce a standard software interface that will expose the capabilities of both process and laboratory instruments.

The group has adopted OPC-UA (Unified Architecture) as the basis for the model. OPC-UA is universally accepted, platform neutral and achieves high-speed data transfer. It has broad industry support beyond process automation and is being used in the development of many other standards. The ADI standard is suitable for a very wide range of devices including laser diffraction particle size analysers. It has recently been released in draft form for industry feedback.

In conclusion

While real-time analysis is extremely valuable for continuous monitoring, anyone installing single unconnected process sensors runs the risk of creating data islands of limited value. Combining the data from different analysers in a single control system holds out the prospect of much greater reward. In practice, though, this requires analysers that have truly effective data management options and capabilities.

At present there are no standards for analyser device interfaces and as a result each integrated control solution is custom built. The OPC-ADI working group has developed a model that will allow analyser vendors to produce a standard software interface. This makes it much simpler to develop control architectures that can acquire and use data from a variety of sources, and strengthens the ability of manufacturers to really exploit on-line analysis as a route to efficient processing.

Oliver Schmitt is now the key interface between the Board of Malvern Instruments and the process sales, product development and process operations team. He is responsible for defining and executing the strategy for application development initiatives, for strategic marketing of process solutions and key OEM relationships. Oliver holds a degree in Process Engineering (University of Mannheim, Germany) and is member of the German Society of Engineers. Malvern Instruments Ltd is based Malvern, Worcestershire, UK. www.malvern.com

Case Study; Granulation

Granulation is a very common unit operation, particularly in the pharmaceutical industry where a principal application is the control of tablet blend properties. The particle size of a granulate affects compressibility and the quality of the finished tablet and is therefore closely specified. In a recent study3 the behaviour of a fluidised bed granulator was studied using real-time particle size analysis.

The analyser successfully tracked granulate size in real-time. It enabled systematic study of the impact on granulate size of control variables such as inlet air temperature, inlet air flow and binder spray rate and the construction of a process model, a tool for optimal control.

Using this model for the automation of control relies on integration of the particle size analyser, flow meters for binder fluid and air, and the inlet air temperature sensor, all of which are very different devices. While this is clearly possible on a one-off basis, standardisation facilitates this process, particularly for companies with fewer resources and less well-developed expertise.

Regerences:1. OPC Foundation Releases Specification for Analyzer Devices Integration (ADI), 10 February 2009; www.opcfoundation.org/Default.aspx/02_news/02_news_display.asp?id=653&MID=News; 2. www.opcfoundation.org;3. Schmidt-Lehr S, Moritz H-U and Jurgens KC 'Online control of particle size during fluidised bed granulation' Pharm. Ind. 69, Nr. 4, 478 - 484 (2007).

Recent Issues