The Challenges of Pilot Plants

A good friend now consulting in Egypt asked me the seemingly innocuous question, “How do you define a pilot plant?” The question arose because a government agency wanted him to examine several facilities in Lebanon and help determine if they should be certified, whatever that means there.

My first response was this: “A pilot plant is a flexible facility housing adequate utilities and a variety of relatively small-scale equipment that can be configured to study or simulate new or existing processes and unit operations. It is meant for research, not small-scale production, though they are often used for that purpose. Small-scale production facilities used for plant trials and market tests or scale-up are more  properly called semi-works, and are built and staffed differently. Compared with pilot plants, they are usually less flexible and have production-oriented staff rather than research technicians.”

What Is a Pilot Plant?
He responded that “a pilot plant is a facility housing small-scale and flexible reproductions of various unit operations common to the commodity groups they serve.” A pilot plant can serve several functions, including the following:

• To generate samples of a new product, line extension, reformulation or packaging for sensory or consumer testing

• To generate samples for more-clinical types of testing such as analytical (chemical) testing, shelf life, and resistance to adverse storage conditions

• To predict how new products, line extensions, or reformulations would run on existing equipment

• To develop manufacturing specifications

• To develop equipment and operating specifications for state-of-the- art processes.

He said that the question of the day, given the above definition, is this: “Is there anyone who certifies or accredits pilot plants?” 

According to my friend, HACCP or ISO9001 might apply for food and/or ingredient plants, but he could not think of any accreditation or certification (is there a difference?) or a research facility.

“I also believe that any certification of a pilot plant for consumer testing may limit its flexibility for research,” he continued. “Will it?”

He contended that “non-technical people tend to be overly impressed with certifications. Hence, this question.”

My friend noted as an aside that at his former company, there was a widely held belief that the pilot plant did not duplicate actual production. He believed this was true for many reasons, including the following:

• Nothing on a small scale duplicates anything on a larger scale, so the argument is semantic.

• The pilot plant was conceived and designed (unconsciously) as a sample-generating facility.

• The company would never pay what was required or wait the amount of time required for effective scale-up studies.

• Most people (food technologists) didn’t know what scale-up was.

He said that as a result of perceptions such as the preceding, the company most often used its production facilities for pilot plants.

--- PAGE BREAK ---

Scale-up
At this point in our exchange there then ensued a long dialogue on the issues of scale-up, which has been discussed in earlier Processing columns. As Leon Levine says in our co-authored book (Food  processing Operations and Scale-Up, K. J. Valentas, L. Levine, and J. P. Clark, Marcel Dekker, 1991, p. 247), geometric similarity is very important and often imposed as a requirement in scale-up. My correspondent then commented as follows.

At this point in our exchange there then ensued a long dialogue on the issues of scale-up, which has been discussed in earlier Processing columns. As Leon Levine says in our co-authored book (, K. J. Valentas, L. Levine, and J. P. Clark, Marcel Dekker, 1991, p. 247), geometric similarity is very important and often imposed as a requirement in scale-up. My correspondent then commented as follows.

“Regarding scale-up, … I think geometric similarity is a good place to start (and maybe even to stay there if possible), but I do not believe that it is always necessary or even desirable. Take the example of heating a liquid in an agitated kettle. A geometrically similar kettle in the pilot plant would heat the liquid much more quickly with the same steam temperature and agitation because of a larger ratio of the heat transfer area to the volume of the kettle. You could turn down the steam pressure in the pilot plant to achieve the same come-up time, but the product might be sensitive to high temperatures, and, thus, the pilot plant kettle will be more benign to the product than its larger cousin.

“So,” he continued, “you could fabricate a kettle with a small heat transfer surface to make the come-up time and the steam temperature the same as the larger version. It seems that one question remains, where to put the heat transfer surface in the smaller kettle, since the heat transfer coefficient will vary with location. Actually, like so many problems of scale-up, that question has already been answered, since you have already specified the come-up time and the steam temperature. Or has it? The heat transfer area will vary according to its heat transfer coefficient (and therefore location). So where to put it?

“I don’t have an answer for that,” he said, “but the point is, by now we have come a long way from geometric similarity. Problems of scale-up frequently end with an imponderable that can only be solved by experimentation.”

Levine agrees that geometric similarity is not always feasible, but he emphasizes that experimentation at more than one scale is required for reliable scale-up. That requirement helps dictate what should be found in a good pilot plant—geometrically similar equipment of at least two sizes, both less than full production size.

Why Have a Pilot Plant?
Sometimes a pilot plant is built specifically to study, demonstrate, and prove a new process or technology. When its purpose is achieved, it might be dismantled or kept as a simulator of a full-scale plant. Such a pilot plant might be focused on just one critical element of a new process, identified by engineers as essential but difficult to simulate theoretically. Only those ancillary elements necessary to operate the critical step might be provided in such a case, meaning that the unit is not a complete process. 

Other pilot plants, common in the food industry, are a collection of flexible unit operations and equipment that may be configured in various ways as needed. Examples would be bakeries with a variety of batch mixers, formers, and batch ovens. These bear little resemblance to the continuous processes found in commercial operations, but translation of batch baking to continuous baking is relatively easy. Matching heat transfer coefficients in the two situations is more problematic. 

It is common to use small-scale facilities or pilot plants to produce test-market quantities of new products. This requires that the facility be a USDA establishment if the products contain more than 2% meat or poultry. Satisfying  the USDA can be difficult but not impossible. More significantly, one should ask whether it is a good idea to make test-market products this way. It may not be. 

--- PAGE BREAK ---

Research technicians are typically trained in the scientific method and are, or should be, accustomed to making “bad” product in the course of experiments. Production personnel are trained, and accustomed, to making only good product. Production typically requires long shifts of repetitious physical work. Research work is more sporadic. Typical pilot plants are not usually designed for routine and frequent sanitation. They also may lack packaging equipment and sufficient storage to support manufacturing. 

On the other hand, using full-scale plants for trial runs is usually very expensive. The savings from avoiding many plant trials can easily justify a properly designed, equipped, and staffed semi-works or small-scale manufacturing facility specifically intended for short test runs. Sadly, relatively few food companies have such facilities. More probably should.

Academic pilot plants rarely attempt to simulate an integrated process, but rather consist of independent unit operations meant for education and research. Which unit operations to include is a matter of choice and preference, but certainly mixing, a variety of heat exchangers, a variety of pumps, drying, evaporation, membrane separation, retorting, freezing, and baking should be  represented. There are commercially available small-scale units specifically designed for low-volume production, demonstration of principles, and research. Some have sophisticated, computer-aided data collection and controls. A few simulate integrated processes for pasteurization of liquids, aseptic processing using several different forms of heat exchange, and vegetable oil refining. 

Regardless of exactly how a facility is equipped, whether by engineered units or a collection of small individual units, a good pilot plant must be designed to be easily cleaned, must have a profusion of utilities (steam, several voltage levels of electricity, gas, vacuum, compressed air, hot and cold water, and circulating chilled  glycol, for instance) and must have adequate storage for unused equipment and supplies. It could have a separate area for packaging, with humidity control for sensitive products.   

Both academic and industrial pilot plants often have  viewing galleries or picture windows to permit observation by visitors without interfering with users. A multi-product pilot plant might have separate areas for wet and dry processing, with dust collection in the dry area. Floor, wall, and ceiling finishes should be similar to those that would be used in a manufacturing facility, for sanitation and durability, as well as to demonstrate in an academic setting good commercial practice.  It is almost always useful to have one or more conference rooms and bench laboratories associated with the pilot plant for small meetings and to perform analytical work. A good pilot plant is expensive, but usually a good investment.

by J. Peter Clark
Contributing Editor,
Consultant to the Process Industries, Oak Park, Ill.
[email protected]