Constant flux reactor

It's one of those fundamental equations drummed into the heads of chemical and process engineers at the beginning of their careers.

The amount of heat (Q) transferred in a heat exchanger is proportional to the area (A) of the exchanger and the temperature difference (?T) between the process fluid and the heating/cooling medium - with the constant of proportionality being the heat transfer coefficient (U).

Or, in case anyone has forgotten:

Q = UAdeltaT

In conventional exchangers, such as a jacketed reactor vessel, the heat load in the process is controlled through the log mean temperature difference deltaT, by adjusting either the flow rate of the heat transfer medium or its delivery temperature.

This convention, however, has now been turned on its head in the constant flux reactor from Ashe Morris. Rather than using deltaT as the controlling mechanism, this patented design varies the heat transfer area A instead, while effectively keeping deltaT constant.

The concept was developed by Robert Ashe, a director of Ashe Morris, not so much as a heat exchanger but more as a reaction calorimeter. His thinking was that if you could monitor the heat into and out of the process, you could infer a great deal about what is actually going on in the process. 'I felt that thermodynamics was the perfect solution for on-line monitoring,' he says. 'In essence, you look at three things - power, enthalpy and heat transfer coefficient'.

Power (in Watts) reveals how fast a change takes place and can be used to monitor rate in such operations as cell growth, crystallisation, synthesis reactions or polymerisation. Enthalpy (in Joules) can be used as a reliable method for detecting reaction endpoints. The heat transfer coefficient can be used to detect fouling or ice formation at the heat transfer surface. Alternatively it can be used to evaluate agitation efficiency or monitor changes in viscosity.

Hitherto, calorimetry on large heat transfer devices has been neither simple to measure nor accurate. 'Once you put a jacket around a vessel to control temperature, your problems start,' says Ashe. The temperature shift in the heat transfer fluid is often too small to detect, let alone measure and the noise to signal ratio is generally extremely large due to fluctuations in the jacket heat.

Ashe's solution was to start from the premise of a fixed inlet and outlet temperature on the heat transfer medium - a constant deltaT in other words - which he thought would give much more sensitive heat balance measurement, whilst drastically reducing the noise to signal ratio.

But to maintain the balance of that fundamental equation, a constant deltaT implies a variable heat transfer area A. So, how has he achieved this?

At first sight his design resembles a conventional half-coil jacketed reactor, in which the heat transfer medium flows through a coil wound around and welded to the outside of the vessel. But this gives a fixed heat transfer area defined by the number of coils. What he has done is literally to take this basic idea apart and arrange individual coils incrementally around the outside of the vessel.

In the prototype version each coil consists of a full tube soldered onto a flat copper conducting band that is wrapped around the vessel. The coils are connected to a vertical manifold containing a piston controlled on a PID loop from the process. By raising or lowering this piston, the number of coils receiving heat transfer fluid from the manifold can be increased or decreased, changing the heat transfer area accordingly.

Each coil band is clamped to the vessel so as to allow for differential expansion of the jacket elements and the vessel itself. Only one control valve is needed for the heat transfer medium, as the PID-controlled piston manifold regulates flow to the coils. Depending on vessel size there could be anywhere from 10 to 200 coils involved.

Tests on a prototype 15 litre model were successfully completed last year and simulation tests, carried out at Imperial College using the gPROMS simulation software package, have demonstrated the much improved reaction control it delivers.

While traditional heat flow calorimetry might only be accurate to around ±20%, Ashe says his constant flux system can measure heat flows to ±0.1%.

Following the winning of a £325,000 Smart Exceptional Award in 2002, Ashe Morris is now working with vessel and reactor manufacturer Pfaudler Balfour to commercially develop the concept into a full-scale offering.

Andrew Wills, MD of Pfaudler Balfour, comments: 'With our experience in terms of both conventional and customised reactor designs, Ashe Morris turned to us to develop the equipment to meet current and future design legislation. At the moment, we are looking to provide a range of reactors from 63 litre to 63m3, and we see no reason for any of the sizes in between not to be accommodated.'

The main development thrust at the moment, he says, is 'to design the equipment to be fully compatible with both existing vessel and building geometries and, where possible, to utilise the existing services, especially where single fluids are used for jacket services. We also have the ability to upgrade older mixed services by utilising the geometric reactor design plus a single thermal fluid package, which can be offered as an upgrade.'

The design of the reactors will be such that customers will be able to use traditional heating and cooling - that is, full jacket zone heating or cooling - or the variable area heating technology associated with calorimetric control. Pfaudler Balfour already has designs that can be used for full commercialisation of equipment and it is currently building a full-scale unit for its pilot plant at its headquarters in Leven, Fife. Because of the extensive use of copper or brass that is required to achieve the necessary heat transfer rates, various new manufacturing techniques have also been developed.

'As yet, our main concentration has been on the pharmaceutical and fine chemicals sectors,' says Wills, 'but we are sure there will be other industries which could benefit from a more reliable and measurable heat control. At present, however, the process lends itself perfectly to the regime of high value products requiring extensive qualitative validation.'