SMALL is beautiful
15 Jan 2000
Scale-up is one of the perennial problems of all branches of the chemical industry - how do you turn a lab experiment into an industrial process producing thousands of tonnes per year? Stuart Nathan looks at BASF's methods to achieve this transformation, using computer simulation and `miniplants'.
A glass tube emerges from the floor of the lab and reaches through the ceiling. It's less than ten centimetres across, but it's cocooned within a plethora of valves and sensors, and connected to a multitude of other devices which look like they were made by a glassblower with a bad attack of hiccups. Clear Iiquids and vapours gurgle around the piping, which is zealously watched by white-coated technicians. And through a set of double doors is what appears to be a control room for a full-scale chemical plant.
This odd-seeming set-up at BASF's Ludwigshafen headquarters is the centrepiece of the company's innovative scale-up methods. Known as `miniplants', the forests of glass tubing and wiring allow the company to translate a lab-top synthesis to a full-scale plant without having to go through the time and expense of building and operating pilot plants.
The traditional method of scaling up an experimental process is for research chemists to work out a commercially-viable synthesis, using readily available and low-cost starting materials and a suitable catalyst. The project then passes into the hands of engineers, whose use the data gleaned by the chemists to design and build a pilot plant. Generally made of steel, these units have a daily throughput of up to a tonne of product. Monitoring these plants gives engineers the data they need to design a full-scale plant.
BASF uses a quite different system. Process synthesis and evolution run simultaneously, involving the use of proprietary software (some of which is part of the CAPE-OPEN project - see PE October, page s3). Many different process schemes can be tested in simulation, which allows a reaction scheme and prospective process to be mapped out very quickly. The results of these simulations allow process engineers to build a `miniplant', which replaces the pilot plant in the usual development process.
Miniplants generally use reactors with a capacity of about a litre, and handle throughputs of 100g-1kg per hour. Their distillation columns are around 2-5cm in diameter. The dimensions are some 100 000 times smaller than a production-scale plant. The miniplants are constructed from a range of standardised glassware, so new miniplants can be built and altered quickly, explains Martin Molzahn, head of corporate engineering R&D.
Miniplants mimic the operation of a full-scale plant exactly. All the instrumentation and the control technology is automatic and state-of-the-art, and the plants are operated and monitored around the clock. `This enables a miniplant to provide all the required information in a relatively short time and at relatively little expense,' says Molzahn.
The data derived from the miniplant is usually sufficient to scale-up directly to a full-scale plant. The development process generally takes 2-7 years, depending on how well known the chemistry of the process and physical properties of the feedstocks and products were known at the outset. In comparison, says Molzahn, traditional process development takes 8-15 years. Molzahn estimates that the company has saved DM30million-50million by using miniplants.
The use of the miniplant does not stop with the design of the plant. Because the glass apparatus and the control system both mimic the operation of the full-size unit, the miniplant can be used to train staff in the operation of the full-scale plant before construction has even begun. The experience they gain speeds up the plant's start-up phase considerably.
With current trends for smaller and more efficient plants, BASF is confident that its work with miniplants gives it a head start.
