Content

Send Us Feedback

Interactive Technology Portal
From Food Tech Innovation Portal

Decision support system for durum wheat products quality

Identification

Key words Artificial intelligence, decision support, expert knowhow, durum wheat chain, quality
Latest version 2012/11/27
Completed by INRA - IATE

How does it work?

Primary objective This tool is a knowledge management system designed to help prediction for the durum wheat processing chain.
Working principle This hybrid information system allows to model a food chain as a set of consecutive unit operations (by example, in the durum wheat chain, extrusion, drying, cooking, …) transforming raw material into a food product. Each unit operation is modelled by a set of process parameters and product characteristics (by example, in the pasta chain, for the cooking in water unit operation, controlled parameters are temperature, % salt, cooking duration, type of water; product characteristics are e.g. vitamin contents, for niacin, thiamin, etc.). The software permits to load experimental quantitative and/or qualitative data for each unit operation in terms of controlled parameters and food characteristics. Those data are used thanks to learning functionalities (based on decision trees) to determine the controlled parameters which have the main influence on food characteristics for a given unit operation. Those decision trees can be also used to simulate the impact of a given set of controlled parameter values on food characteristics.

It allows evaluating both faults and qualities of food products. (1)

alt text

Images
Additional effects The system does not require an a priori model.

It could be used for risk and benefit analysis.

Important process parameters Process parameters, which may influence the product characteristics such as temperature, % salt, cooking duration, type of water…, must be entered. The learning functionalities will automatically determine the actual influence of the process parameters on the product characteristics with respect to the experimental data entered in the database.
Important product parameters

What can it be used for?

Products This technology is generic; it can be used for any kind of food product chain. It requires inputting the data for each specific food chain.
Operations Quality managing
Solutions for short comings Demand for safe, healthy and tasty foods

What can it NOT be used for?

Products none
Operations none
Other limitations This method requires experimental data and/or formalized expert knowledge to be used. For a given unit operation (by example, pasta cooking in water), controlled parameter values, impact on the studied characteristic(s) (value before and after the unit operation) must be entered.
Risks or hazards not known

Implementation

Maturity registered at the Agency for the Protection of Programs.

The software is not commercialised. It is a research prototype.

Modularity /Implementation This system is to be installed alongside the whole durum wheat processing chain.

It could also be adapted to other food chains.

Consumer aspects Consumers will certainly accept food products with improved safety, health and taste
Legal aspects registered at the Agency for the Protection of Programs (07.12.2007, n°IDDN.FR.001.490041.000.R.P.2007.000.30605).
Environmental aspects No information currently available

Further Information

Institutes INRA - IATE, INRA - MISTEA, UMII - LIRMM
Companies (no current commercial product)
References (1) R. Thomopoulos, B. Charnomordic, B. Cuq and J. Abecassis, 2009. Artificial intelligence-based decision support system to manage quality of durum wheat products. Quality Assurance and Safety of Crops & Foods. 1(3): 179-190

Process parameters, which may influence the product characteristics such as temperature, % salt, cooking duration, type of water…, must be entered. The learning functionalities will automatically determine the actual influence of the process parameters on the product characteristics with respect to the experimental data entered in the database.warning.png"Process parameters, which may influence the product characteristics such as temperature, % salt, cooking duration, type of water…, must be entered. The learning functionalities will automatically determine the actual influence of the process parameters on the product characteristics with respect to the experimental data entered in the database." cannot be used as a page name in this wiki.

Software and Models 2.1.3 biological not applicable ICT Interviews with the researchers from INRA-IATE: Rallou Thomopoulos, Joël Abecassis, Patrice Buche WikiSysop :Template:Review document :Template:Review status



Translate this page with Google Translator (automatic translation)
Created by Hte inra on 21 June 2011, at 10:18