A measurement-based closed-loop control system using in-process ATR-FTIR spectroscopy coupled with a multivariate chemometric PLS calibration model is developed, validated, and applied to the monitoring and control of supersaturation in a 250- L industrial pilot-plant crystalliser. Supersaturation control experiments are carried out on seeded batch cooling crystallisation of ß-l-glutamic acid from aqueous solutions using two methods of seeding involving addition of seeds to the solution and generation of seeds within the solution. The generic applicability of the approach is demonstrated through this challenging system reflecting this molecule's weak chromophore for infrared and relatively low solubility compared with previous solute-solvent systems. Based on the laboratory experiments, the system was fully tested and optimised prior to a series of trials carried out in an industrial pilot plant at Syngenta, Münchwillen, Switzerland. Good control of the supersaturation is achieved at three levels, 1.1, 1.2, and 1.3, within a prescribed range of ±0.025. The average product crystal size is found to decrease with increasing supersaturation. Comparison between product crystals produced at the 20- and 250-L scales indicates that secondary nucleation is more prevalent at the smaller-scale size. For the same level of supersaturation, the rate of depletion of solute is faster at the 20-L scale size than at 250-L scale, and hence a higher cooling rate is required to maintain the desired supersaturation. However, for a given crystalliser scale size, as expected, the mean cooling rate required to maintain a constant supersaturation is found to increase with increasing supersaturation level. © 2011 American Chemical Society.