Acoustic chemometrics (a.c.) monitoring of chemical production processes
Systematic methods and tools for managing the complexity
Process Analytical Technology - PAT (T4-6)
Keywords: acoustic chemometrics, process analytical technologies
Acoustic chemometrics (a.c.) is a new general Process Analytical Technologies (PAT) approach for on-line monitoring of industrial processes. Acoustic chemometrics concerns capturing system vibration characteristics, e.g. from two-phase systems (gas-solids/liquid-solids) generated by the manufacturing process or by transportation. The resulting vibrations can easily be measured by non-intrusive, "clamp-on" sensors (e.g. accerometers). Acoustic signatures carry embedded information about physical and chemical parameters, such as composition (oil, fat, ammonia, buttermilk, glycol, ethanol), mixing progress, fiber length, flow, density, temperature - as well as system state. For extraction and quantification of these types of specific analytes and parameters of interest, domain transforms (FFT, WT) and PLS-regression is essential for multivariate calibration (process chemometrics).
Acoustic chemometrics is here applied for monitoring of a industrial production process, wiz. a feasibility study of fluidized bed granulation of a fertilizer product (urea) (in a semi-industrial pilot plant: SIPP), illustrating the main acoustic chemometrics features and benefits. Conventional process monitoring and control is sub-optimized by slow and labour-intensive laboratory analysis: particle size distribution, liquid urea-concentration (correlated with water content), which are typically delayed by up to some 2+ hours of analysis time. It is also critical to be able to detect process transients and upsets. The new acoustic chemometrics approach goes directly into the real-time, on-line domain. In this examplar study we focus on:
1. Optimal sensor localization in the semi-industrial granulator plant
2. Testing new sensor type (high-temperature microphone) in a cooling chamber
3. Assessment of the feasibility to predict (PLS):
• Bed movement (air-flow trough granulator)
• Liquid feed concentration
• Reflux of fine material to granulator
• Product level in fluidized bed
• Average product particle size
4. Monitoring for visualization of critical situations - early operator warnings
With acoustic chemometrics it is possible to monitor both process state and product quality for industrial process control. Relevant early warnings trigger the process operator to change relevant process parameters to control product quality or to prevent critical shut-down situations. Successful validation of this type of PLS-prediction models signify that acoustic chemometrics has matured into a proven on-line technology in the Process Analytical Technologies (PAT) domains. There is a huge carrying over potential for industrial a.c.
See the full pdf manuscript of the abstract.
Presented Wednesday 19, 12:00 to 12:20, in session Process Analytical Technology - PAT (T4-6).
