I have seen many changes in pilot plants over the course of my career, but I predict that we are on the verge of an unprecedented evolution of these units. My crystal ball sees 10 key factors influencing next-generation pilot plants:
- fugitive emissions;
- multiple trains;
- online analytical capabilities;
- safety and control system interaction;
- wireless technology;
- instrument availability;
- instrument multi-functionality; and
- unit size.
Let’s look at each of these and what they may spur.
Contractors will play an expanding role in supplementing or replacing in-house resources in the conventional design, construction, start up and operation sequence — prompted by companies’ desires to be more efficient and responsive while minimizing commitments to longer term in-house resources. This will range from contract design, construction and maintenance to increasing use of outsourced analytical services, programming and even operations. The greater flexibility contracted services offer to gear up for a sudden short-term need or scale back during an industry downturn will prove irresistible to many organizations. However, those firms apt to be the most successful will maintain some in-house expertise — at a reduced level overall but concentrated in more depth and considered more of a strategic resource. Companies will continue to value in-house design skills but will be more willing to bet that an outside firm can design a pilot-plant vessel right or built it just the way they want. The most successful will recognize the need to maintain some fairly high level of expertise to find, evaluate, review and select the best contractor — and probably to do some or all of the unit process design — as well as to make use of the resultant pilot plant and its data.
Manual operation has already almost completely given way for operation of all but the simplest pilot plants. Automation currently is moving along the path of reducing operating staff attendance from essentially full-time down to progressively less-and-less part-time. The next generation of units will require even lower operator presence and make much greater use of recipe-driven menus that allow the operator to select the operations sequence from a master list and then depart, secure in the knowledge that the pilot plant will properly execute each step (well, at least most of the time). Different pilot plants will employ the same sequences, as organizations strive to develop a more-standardized approach to common operations like charging, pretreatment and sampling. Efforts to develop the “best” approach will make these operating sequences more uniform and sharable. Examples include more-complex charging, filling and preparation arrangements, automated sampling protocols and even operational sequences like planned experimentation based on the latest test results.
Increased toxicity of materials, reduced exposure limits and growing concerns for the long-term health effects of any exposure will push efforts to design and construct units that are leak-free under all circumstances. Decreasing operator attendance, which reduces the time available for identifying and locating leaks, also will promote this trend. The combined health and operational concerns will spur companies to install more equipment that is less leak-prone. Sealless pumps and mixers, bellows-seal valves, and high-integrity fittings typify the leak-free components rapidly becoming common on pilot plants. Automatic tube welders, which make welding easier and a more viable alternative to conventional joining methods, will proliferate, while specialty closures and assemblies will increasingly replace conventional flanges and piping. More and more instrumentation will come as sealed units or with higher-integrity seals. Routine automatic online leak detection, currently rare and intermittent, will become more popular to address the reduced operator presence and ensure safety when no one is around.
The reduced staffing that automation makes possible, coupled with the enormous expansion in data work-up and mining capabilities offered by today’s computers will promote the increased use of multiple trains. This will increase the complexity of pilot plants as well as their support and maintenance requirements — but the added productivity and effectiveness will outweigh the higher costs. Such setups may consist of multiple trains on the same unit or multiple copies of a single unit, depending upon the organization’s requirements. They will provide not only traditional data but also more-in-depth analytical and operational results for use in evaluation and design.
Online analytical capabilities
Over the last 20 years, the number of online analytical tools has dramatically grown. This trend will accelerate as analytical data become more integrated into process operation and not just data analysis. Process control based on real-time analytical data, already increasingly popular, will widely spread. Process optimization, only just beginning to grow in pilot plants, will proliferate. More importantly, integrating these data into the pilot plant’s control system will become more uniform and hence easier and less expensive. Third-party programs, integrated systems and common bus structures will allow the data to be fed to the process control system in a more-straightforward, less-proprietary manner. The complexity of the pilot-plant control system will grow as these inputs are integrated to the maximum feasible extent. More-difficult analyses such as particle-size distribution and complex product compositions will gain a greater role.
Safety and control system interaction
The age-old separation of control and safety systems has largely blurred into being almost unrecognizable on many pilot plants. Growing concern over how well the safety system will respond should the control system be unavailable or non-functional will force pilot-plant control systems in new directions. In some cases, simplified layer-of-protection analysis will lead to an overall safe design. In others, the prevalence of separate microprocessors in a single control system will allow operation safely in both modes, given proper initial configuration. Continued use of separate control and safety systems will remain common for the foreseeable future, but they both will be microprocessor-based, smaller, cheaper and more failsafe, as well as easier to integrate and program. Integrated systems that have separate microprocessors on each board or rack will become more common and provide the redundancy a safety system requires.
We have only begun to scratch the surface of using wireless technology for pilot-plant operations. While the long distances between sensors and control, which are driving this technology in plants, usually do not exist in pilot plants, its lower cost, greater flexibility and reduced construction time make the technology too attractive to ignore for much longer. As wireless devices become cheaper and more common, thanks to their use in plants, they will gain greater acceptance for pilot plants. As time goes on, they will replace the usual hardwired systems from small tank farms and remote operations. We will see greater use of wireless highways not just to gather data but also to transmit data to end users and storage.
Small-scale magnetic flow meters, vortex meters, corrosion probes and numerous other devices were but a dream for most pilot-plant designers 20 years ago. Now many are becoming increasingly common and low cost. The growth in this area will continue. The availability of these devices will allow pilot-plant designers to solve some issues that have plagued them for years (much as the advent of thermal mass-flow meters in the 1970s finally put to rest the search for an ultra-small-size control valve to use with differential pressure devices). The resultant boost in accuracy and reliability will, in turn, enable pilot plants to produce valid useful data with every run — obviating statistical analysis of several runs to address inaccuracy and non-repeatability.
Multi-functional units will proliferate. Pressure transmitters will simultaneously measure temperatures; flow meters also will provide pressure or density in a single unit. Calibration of most new transmitters will occur while the unit is in place and online. While the individual transmitter will be more expensive, it will be smaller, more accurate and more reliable. The decrease in installation costs will more than offset the higher purchase price. These multi-functional units also will interface more easily with control and data-acquisition systems, generating additional savings in programming and set-up.
The days of the size of pilot plants shrinking every generation are probably approaching a realistic end. However, the use of very small high-throughput “pilot plants” (which actually are more akin to very complex experimental equipment) will increase. These high-throughput units will handle much of the screening currently performed more slowly and expensively in standard small pilot plants. Highly automated pilot plants then will run the promising leads at a more realistic and scalable range, to evaluate synergistic effects and operations at transient conditions as well as process conditions more realistic of a plant environment. The combination, when properly applied , will produce a greater number of high quality leads faster, and provide a means to screen these for the next generation of process or product improvements. Modeling will continue to augment and validate pilot-plant operations and, in the always symbiotic relationship, pilot plants will continue to augment and validate modeling.
The combination of all of the trends described will translate into increased cost to design, construct, start up and operate next-generation pilot plants. It also will raise the expense and effort to keep these pilot plants effectively running. The days of maintenance support being a few craftspeople on loan from the plant or hired when needed through a local contractor are over — although many maintenance functions will be routinely outsourced for cost or to gain access to specialists. Just as the “tooth to tail” ratio in the modern military keeps getting smaller as the lethality of weaponry and their associated complexity increase, so the “unit to support” costs of pilot plants will shrink; the data will become better, more useful and more focused — but keeping units working properly will incur higher costs and effort. The traditional process and mechanical engineering support requirements will continue, matched now by computing, automation, safety and electronics support requirements.
Will all these predictions come to pass? Probably not, although I think most will, in some form or another. Beyond these, I forecast that an even-more-novel trend, not mentioned nor even imagined by most pilot-plant personnel, will arise and significantly change the way we all design, construct and operate our pilot plants. After all, that’s what research is all about, change, both planned and predicted, as well as new and unexpected!
These predictions, of course, represent my personal view, not necessarily that of ExxonMobil or any of its affiliates.
Richard Palluzi is a senior engineering associate at ExxonMobil Research and Engineering Co., Annandale, N.J. E-mail him at firstname.lastname@example.org.