As growing numbers of manufacturers turn to on-line particle size analysis, the issue of how best to ensure a successful project becomes increasingly relevant. This white paper examines different strategies for the design and implementation of on-line particle sizing, highlighting the economic and practical benefits of choosing the one best suited to the specific project.
The proven ability of on-line particle size analysis to transform process control and deliver substantial economic benefit in the form of higher throughput, reduced waste, lower energy consumption and enhanced product quality has seen its widespread application across the processing industries. Fundamental to the development of a project proposal are discussions as to how best to implement the work, with a spectrum of possible options. At one end lies the turn-key solution where every aspect of the project is handled by an external supplier, right through to commissioning and productive operation. The opposite extreme involves buying in just the hardware that can't be manufactured in-house and completing the rest of the project with existing resource.
The white paper examines the choices around project implementation, how to make them, and their impact on the success of the overall project. Typically the installation of on-line particle size analysis involves a number of steps, including: hardware selection; process interface design; automation and integration within an existing control platform; validation; and commissioning and control optimization. Examining these elements in turn we conclude with an assessment of the pros and cons of different implementation strategies, through the presentation of a number of example case studies.
When considering the adoption of on-line analysis it is vital to be realistic about the likely benefits, but it is equally important to weigh these against a pragmatic and properly costed delivery schedule. Deciding on an appropriate implementation strategy upon which to base this estimate requires careful assessment of available resources, and a thorough evaluation of the true costs involved in completing the project in different ways.
Although many on-line techniques are still in their infancy, on-line laser diffraction particle size analysis is relatively mature and is proven across a number of industries. Early adopters have, over the preceding decade or so, led the way in tackling implementation issues. Consequently there is now a secure knowledge base covering almost every type of application. New users therefore have choices in terms of implementation strategy that range from simply purchasing a sensor to handing the whole project over to the instrument supplier - providing, of course, that service is on offer.
Taking the decision to reduce external input as far as is possible has the advantage of reducing CAPEX to a bare minimum, and in some cases may be necessary either to justify the project or to get approval. A turn-key solution in contrast will carry a higher price tag, but mitigates any perceived risk and alleviates the difficulties of resource allocation.
Every project is unique with the risk-reward balance varying from case-to-case, along with levels of relevant in-house knowledge, however the central elements of an on-line sizing project are shared. Looking at each of these elements in turn helps to focus attention on which areas may be feasibly handled in-house and, equally, where external support may be especially beneficial.
A first step in any particle sizing project is identifying hardware that meets the necessary specification. Operating temperature and pressure may influence choice, as may any zoning requirements. The nature of the process stream is clearly critical: Is the sensor going to operate in a moist or wet environment measuring a slurry or emulsion, for example? Are the particles in the stream abrasive?
To select the best solution it is also important to consider how frequently the process needs to be measured. Real-time laser diffraction particle size analyzers are capable of measuring complete particle size distributions in less than a second but in some instances this is more than is strictly necessary. If the process has slow dynamics, for example, or if instant upset detection is not essential, then a lower measurement frequency may be adequate. Where this is the case a single sensor may be able to cover multiple points in the process, increasing the potential benefit. This can have an important impact on the economic case.
With this element of the project, the split between what is handled by the supplier and what comes from the customer is straightforward. The specification relies on plant information and the supplier must provide hardware to meet it. Assessing the quality of different products will though form part of the decision-making process. This is particularly important for process equipment since reliability is crucial and the operating environment is normally quite challenging: noisy, dusty or subject to vibration, for example. Proven reliability and low maintenance requirements are important selling features.
Analytical instrumentation designed for continuous use in the plant environment has to successfully integrate and interface with the process. For particle size measurement a key issue is reliable, representative sampling since failures in this area are often the major source of error in reported data. In-line laser diffraction measurement is the simplest option and used wherever practicable. However in some instances, for high tonnage flows, for example, or highly concentrated streams, it is necessary to install the analyzer on-line, on a discrete sampling loop. A representative stream is continuously extracted from the process, circulated through the loop, measured and then recycled to the process or stored separately
The importance of this aspect of the project leads many people to seek expert support at least in the design of the interface, if not its manufacture. However, there are many proven, off-the shelf solutions for sampling, ranging from two-stage screw auger systems to simple flutes and eductors for more modest flow rates. The customer might know of a proven, commercially available system, have an existing installation to copy, or sufficient expertise to engineer a secure solution. The most important point is that enough attention is paid to this aspect of the installation and its validation, as a sub-standard design can compromise the entire project.
Certain companies, such as Malvern Instruments, provide a consultancy service to tackle this aspect of the project and identify an optimal solution via an on-site trial. Such consultancies, as well as confirming the most appropriate interface design, provide reassurance that the technology will deliver to expectations and they enable a realistic assessment of the potential benefits.
In the simplest case, a continuous particle size analyzer may have a bare minimum of automation, operating via a stand-alone PC, with clean air and power the only additional requirements. Increasingly though, the trend is towards automated control, often on the basis of multiple variables. Here, the automation requirements are far more involved, and may be very specific. Interfacing productively with the supplying company is especially important to ensure an optimal solution that meets all site protocols.
The trend towards multivariate control has stimulated significant advances in supporting hard and software. A notable development has been the introduction of the OPC architecture [1], a standardization designed to make it easier to integrate an array of instruments from different suppliers in a single control system. The development of multivariate control systems is therefore becoming easier and there are companies that specialize in this provision. However, the legacy of these same rapid advances is that there are many different systems in active use across the processing industries.
Choices in the area of automation hardware, and associated software, are therefore numerous. It may be that the peculiarities of the site's unique control system are best handled by those who know it best, for example, or that the instrument is being purchased as part of a much wider multivariate control capital project, in which case expert control engineers may well be involved, or conversely it may just be that requirements are extremely simple.
In this last case it is worth noting that even if the initial scope of the project is simple QC or support for manual control, it is becoming increasingly clear that real-time analysis often delivers maximum return when used to automate control. Including the necessary equipment to upgrade may therefore be sensible future-proofing.
Validation needs vary from industry to industry but are a critical issue for some, the pharmaceutical industry being a prime but not exclusive example. Typical requirements may include: 21 CFR Part 11; IQ/OQ testing and documentation; Site Acceptance Testing; and Factory Acceptance Testing. An audit of the supplying company may also be necessary.
Companies with extensive validation requirements may have departments dedicated to the activity but many find that where a supplier offers validation support it is less burdensome and more cost-effective to outsource as much as this activity as is feasible, while recognising that the ultimate responsibility for validation rests with the end-user. Such a strategy tends to go hand-in-hand with allowing a single supplier to provide all of the hardware for the project and most of the automation equipment. Where the decision is taken to request validation support then the full extent of any requirements must be clearly communicated, to allow reliable assessment of the associated costs.
The early days for a new piece of equipment are critical, and may ultimately define the value it goes on to deliver. Rapid, effective commissioning is vital but it is equally important to ensure optimal use of the resulting data. Where the goal is automated control then a primary aim is to close the control loop quickly and efficiently.
Laser diffraction continuous particle size analyzers deliver complete particle size distributions. This mean that control - whether manual or automated - can be based on any number of parameters: % below 5 microns, for example, between 3 and 30 microns, or over 200 microns, or some combination of parameters. Choices here depend on which specific elements of the particle size distribution best define product performance. These may be known or may quickly become apparent once real-time sizing is in place.
In these final stages of the project, having the support of an instrumentation engineer familiar with the system and its use can be invaluable. Indeed, some suppliers have industry-specific expertise, knowledge of how best to use the real-time data that can be pivotal in maximizing return over the long term.
Clearly there are a number of further choices to be made to tailor the project implementation strategy to truly meet user needs. The following examples help illustrate how these choices may be exercised in practice, by users with different requirements, to maximize the value of on-line particle sizing in each case.
Company A mills batches of powder coatings to order, using plant that is manually operated. Particle size is controlled by the operator by changing mill conditions on the basis of lab results measured using a laser diffraction analyzer. Continuous particle size analysis is seen as a potential route to better product quality, higher throughput and lower manual input but there is minimal engineering support, and CAPEX is very limited.
Although this company has little in-house expertise for project implementation, its requirements are relatively simple. The plant has very limited control architecture with which to integrate and the main aim is simply to improve manual control by providing a continuous information flow.
The pragmatic option here is a sensor only purchase, to bring down CAPEX to an accessible level. Drawing on freely available literature and in-house knowledge the company can design and manufacture a simple process interface, and subsequent automation requirements are minimal. Some experimentation may be required to optimize data use but existing company experience with the analytical technique will help.
Company B is a global cement producer, with multiple sites. The company is investing heavily in expert control systems with the aim of using sophisticated multivariate process control to continuously optimize production 24/7. Real-time particle size analysis is essential to control the grinding circuits that produce the finished cement.
This application is relatively complex with automation demands a particular issue. The scope of the project suggests that purchase of a proven process interface may be the best route, and alongside it a relatively sophisticated automation package to integrate the instrument within the planned control structure. For companies with multiple sites the goal is to identify a reliable solution that can be duplicated easily. Proven packaged offerings may therefore provide the solution.
In this instance it may also be advantageous to draw on external applications support. A technology trial, for example, would help reliably estimate the potential benefits of automaton, as well as confirming the suitability of the process interface. During commissioning, expert support could help to fully exploit the capabilities of the expert control system. Given the multi-site nature of the company, such outgoings may only be required for the first one or two sites with subsequent projects then drawing on in-house expertise for implementation to reduce the roll out costs.
Company C is a pharmaceutical manufacturer looking to upgrade from manual to automated mill control. Starting with a pilot scale installation the company is seeking to develop a secure automated PAT solution suitable for excipients and actives that can, eventually, be transferred into manufacture.
Over the long term this company stands to make significant return from an effective, widely applicable automated solution and in part this project is about investing for the future in line with the FDA's PAT initiative. The pharmaceutical industry is still relatively inexperienced in the area of real-time process analysis and, of course, has a stringent regulatory requirement. All these factors suggest that a turn-key project, including all necessary validation and support during commissioning, may be the best way to go.
This approach would mean a high initial outlay but would also provide the most secure of foundations for a proper assessment of the full potential of the technology. By buying in an optimized process interface, a fully functional integrated automation package and the application support needed to ensure comprehensive use of the resulting data, the company should be able to maximize return over the short and long term. A further advantage is that the experience gained during the project will help to develop expertise that will support the wider implementation of process analysis with techniques that are less well-established.
A successful project delivers the required return, in the planned timeframe. Central to achieving this goal is being realistic about how the project will change operation and what it will cost in terms of time, money and expertise to deliver these benefits. An important part of these considerations is how best to implement the project.
The maturation of on-line laser diffraction particle sizing has brought with it a range of choices as to how best to work with a supplier, when adopting the technology. In the past these systems were usually purchased in the form of turn-key solutions because of the associated technical risk and the expertise required to successful engineer a solution. Now though it is possible to take certain elements of a project in-house. This isn't the best solution for everyone, but it may be beneficial for some. Being realistic about in-house strengths and limitations is the key to choosing the best approach, to accessing the technology in the best possible way, and to reaping the fullest reward.
[1] News item from the OPC foundation website 'OPC unified architecture companion specification for analyzer devices released' Nov 2009. http://opcfoundation.org/Default.aspx/02_news/02_news_display.asp?id=740&MID=News
