The MULTIPLE project will strengthen EU photonics manufacturing base. Cost-efficiency, flexibility, high productivity and quality attained thanks to MULTIPLE solutions will allow laser, optical measurement and image processing equipment, companies in Europe to compete glob ally through a breakthrough technology in manufacturing, thus benefiting from business opportunities in fast growing markets.
CONTACT US

Latest News

Share this

info@multipleproject.eu

Spectroscopic analysers

MULTIPLE is developing cost-effective and versatile spectrometers specifically tailored to the current target applications (steel manufacturing, wood working and the food industry). The developments will provide cost effective spectrometers for accurate chemometric analysis covering a wide spectral range form (0.4-5 µm). In addition, MULTIPLE is developing innovative monolithic detectors based on organic electronics with customised resolution and spectral bands in the 0.4- 1.7 µm range.

The spectrometer solutions will allow measurement of gas, liquid and solid substances in transmission mode, diffuse reflection and interactance. MULTIPLE will combine a recent breakthrough in tunable organic electronics based sensors with highly specific laser-based spectroscopic schemes to a variety of chemometric analysis in the VIS/MWIR range (0.4 µm – 5 µm), from gas concentration measurement to the detection and concentration measurement of constituents, initially specified to those in food and wood products.

 

Chemometric models for spectroscopy  

Spectrometric technology for chemometric analysis has gained popularity thanks to its non-destructive nature and the capability to provide highly specific and accurate measure MWIR range are still synonyms of either sensitivity issues or expensive equipment. MULTIPLE will develop neural network models and analytical models for robust estimation of gas concentration ratios of different compounds, including CO/CO2/O2, NOx, and formaldehyde using laser-based IR spectroscopy. Different types and structures of neural networks will be studied and developed in order to reach the optimum structure for the application.  

 

The final choice of structure will be based on number of calibrations data sets used for learning, accuracy and precision of results, immunity to cross interferences form other gas constituents in the sample, immunity to temperature and pressure variations and complexity for introduction into serial production. The neural networks will be simulated in MATLAB environment and finally deployed on the embedded spectrometer hardware platform. In parallel to development of algorithms for chemometrics, a work will be carried out towards implementation of analyzer auto calibration features which is crucial for long term autonomous operation. The auto calibration schemes will also be implemented in the embedded spectrometer platform. It is planned that a dedicated PIC microcontroller will be responsible for auto calibration and self-diagnostics functionality.