Amadou Ndiaye

Qualitative modelling of a multi-step process: The case of French breadmaking

In this paper, we investigate a problem of qualitative modelling of a multi-step food process, French breadmaking. The French breadmaking process has been represented as a sequence of steps where each step is defined through its control variables, the state variables of its output, and the causal relations between the control and state variables. In addition, the output of a step is the input of the followed step, and then, the state variables of one step depend on its control variable and the state variables of its input. A qualitative model of mixing, certainly the most complex operation of breadmaking process, has been built up. Human experts reasoning has been represented through seven cognitive operations and a qualitative algebra (Q,~,@?,@?) has been defined to model the calculation of state variables of mixed dough from the mixing control variables and the ingredients (the mixing input) condition. The relations well known by the human experts and other relevant relations between state variables of the mixed dough have been found out through the qualitative equations established. The mixing model has been implemented using the QualiS(C) expert-system shell. The score established according to French standard of the dough after the mixing step was compared to the one computed. In most of the 81 cases simulated, satisfactory results were obtained since the only unfavourable cases had never been experimented by the experts before, which finally validated this approach, thus worth to be extended to the following steps of breadmaking process.

Meta-heuristic Methods Applied to the Design of Wood—Plastic Composites, with Some Attention to Environmental Aspects

This article addresses the optimization of the quality of wood—plastic composites (WPC) designed for outdoor uses such as decking, taking into account the environmental impact during the life cycle of the product, from production to end of life (from ground to ground). In a context where several conflicting objectives must be satisfied simultaneously in the design process, meta-heuristic approaches provide efficient methods for optimization. Multi-objective particle swarm optimization (MO-PSO) has been chosen here to solve a complex problem in which physical properties such as creep and duration of load, water absorption and swelling, need to be improved with a limited impact on environment. This requires to get reliable information on material properties as related to its composition, environmental impacts through life cycle analysis (LCA), and to implement relevant information such as exhaustion of fossils resources through analytical or probabilistic models in the PSO algorithm in order to obtain a set of optimal solutions for the composite. This article shows the feasibility of this approach, which can be generalized in the design of any type of composite structures, provided objective functions can be specified.

Formulating preliminary design optimization problems using expert knowledge: Application to wood-based insulating materials

Abstract The design of wood-fiber based thermal insulating material with optimized properties and characteristics requires a good scientific knowledge of the latter. Currently, the technical, economic and ecological characteristics of a wood-fiber based composite mat are not well known. This article presents a methodology for the acquisition of expert knowledge so that the properties and characteristics can be modeled. The knowledge domain is multidisciplinary and it was delimited and decomposed into disciplines and domains of expertise. A panel of seven experts was constituted to cover the various disciplines and domains of expertise of the knowledge domain. Knowledge acquisition sessions, guided by the estimated importance and the availability of knowledge, were conducted using semi-structured interviews and the mapping of the existing causal relations between variables. A causal map was established to represent the causal knowledge of each of the experts and then, the established causal maps were assembled into a unique global causal map, which was subsequently validated by the experts. It contains the information necessary for formulating the properties and characteristics to be optimized, which were: thermal conductivity; thickness recovery of the material; the manufacturing cost and the products environmental impact. Properties and characteristics are function of raw material type, their morphological properties and the manufacturing process variables. This methodology makes it possible to establish which objectives to optimize and which variables influence each objective. Consequently, the objective functions of the optimization problem can be clarified, specified and modeled.

Predicting the quality of wheat flour dough at mixing using an expert system

Modelling the links between the mixing process conditions of wheat flour dough and the properties of the dough is a challenge. This paper presents a systematic modelling approach based on qualitative algebra to represent human expertise in this domain. Qualitative models of wheat dough mixing have been implemented as an expert system, called Ascopain. The relations between the process conditions - flour specifications and kneading conditions - and the dough sensory properties, have been formalised by means of qualitative functions. An extensive evaluation of Ascopain is provided by comparing the simulation results, first to experts predictions, and second, to experimental results of sensory evaluation of mixed dough properties. The good matching level proves the accuracy and the robustness of the expert-system and, overall, its ability to implement a reasoning on the influences of process conditions to predict actual dough properties, starting from ingredient characteristics.

Phenomenological model of maize starches expansion by extrusion

During extrusion of starchy products, the molten material is forced through a die so that the sudden abrupt pressure drop causes part of the water to vaporize giving an expanded, cellular structure. The objective of this work was to elaborate a phenomenological model of expansion and couple it with Ludovic® mechanistic model of twin screw extrusion process. From experimental results that cover a wide range of thermomechanical conditions, a concept map of influence relationships between input and output variables was built. It took into account the phenomena of bubbles nucleation, growth, coalescence, shrinkage and setting, in a viscoelastic medium. The input variables were the moisture content MC , melt temperature T, specific mechanical energy SME , shear viscosity η at the die exit, computed by Ludovic®, and the melt storage moduli E’(at T > T g ). The outputs of the model were the macrostructure (volumetric expansion index VEI, anisotropy) and cellular structure (fineness F) of solid foams. Then a general model was established: VEI = α (η/η 0) n in which α and n depend on T, MC , SME and E’ and the link between anisotropy and fineness was established.

Structuring and representation of scientific knowledge for a transfer through an electronic knowledge-book

The use of scientific knowledge on waste wood management is the main factor that facilitates the responsible use of recovered wood respecting the environment. Nowadays, numerous scientific results are generated and published by universities and research institutions; but small quantities are exploited by nonscientific people. This problem is due to the difficulty of transfer and assimilation of this knowledge in its current representation. In this paper, we propose a new structure and representation of both type of scientific knowledge (declarative and procedural) to facilitate its transfer to non-scientific people through a hypertext electronic knowledge-book. The declarative scientific knowledge has been represented by concept maps and hierarchically structured through semantic relationships and the procedural scientific knowledge has been represented by technical sheets.

A Knowledge Base on Cereal Food Foams Processing and Behaviour

In a recent project on the processing and behaviour of cereal food foams, namely Salve, works have been carried out focusing on “cellular structure” as a common concern. The effects of product composition on expansion and its sensory properties, its creation and influence on mechanical properties, heat and mass transfers during baking and numerical modelling of these various steps have been taken into account. These works were also concerned with the technological know-how. A knowledge management approach has been developed in order (i) to collect the knowledge in this well defined domain (cereal food foams); (ii) to represent it using concept maps by locating the tasks, the concepts and their relations and identifying the lack of knowledge; and finally (iii) to verify experimentally the accuracy of the representation using an implementation through a knowledge book. Data collection was performed by adapting to scientific knowledge the models available from a method used for the know-how capitalisation. A knowledge book was engineered for the Web using the concepts of navigation, ontologies and keywords, using the MySQL web oriented database management system. Once the accuracy of the representation is checked, it will be used as the basis of knowledge transfer to the main people involved in manufacturing.

Knowledge book on lignocellulose deconstruction: an INRA project to identify key actions in research on biorefineries

The existence of pilot and industrial scale biorefineries worldwide demonstrates the technical andneconomic feasibility of fractionating lignocellulose (LC) for chemistry and energy. This raisesnnew questions about the biomass supply, management of its quality and about the elementary stepncombination in processes, to choose which compound will be main- or co-products from plantnbiomass.nIntegrated and systemic approaches are requested to invent and/or to improve the biotechnicalnfractionation of LCs and there is a need to collect and correlate the existing knowledge in anstructured way, to gain a better insight of the overall process. Building such a knowledgenrepresentation is important for scientists, research institutes, universities and industries, as it willngive a shared description of the knowledge in that field, that will further facilitate its diffusion,nre-use, review, reassessment and updating with new findings.nPractically, an extensive literature has been published in the past five years on the biorefinery ofnLC, focusing mostly on the saccharification of polysaccharides (cellulose, hemicellulose). As anconsequence, most of the data available results from biochemical and physicochemical analyzesnfrom several processing chains, which combine different modes of physical pretreatments and/ornchemical typologies of variables biomass and/or various hydrolytic enzyme cocktails.nThat is why a project for development of a hypertext electronic Knowledge Book onnLignoCellulose DeConstruction (KB-LCDC) was initiated by the French National Institute fornAgricultural Research (INRA) with two main goals: i/ to elicit the available knowledge fromnvarious sources, more specifically related to the enzymatic hydrolysis of wheat straw intonglucose, ii/ to represent the knowledge and implement it into a web-based format of KnowledgenBook (KB) taking into account the overall saccharification process.nThe knowledge was first elicited by means of semi-structured interviews with a group of sixnexperts working in several INRA research Units, and involved in the Institute’s biomass transformation network. Concomitantly, the collating of data and knowledge from “grey-” andnpeer reviewed- literature was also done.nThen, our approach consists in building a knowledge book (KB) whose pages are formattednconcept maps (Cmap) and technical sheets that are connected by hypertext links. A Cmap is ansemantic graph where nodes represent concepts that are connected by arcs expressingnrelationships between them. A formatted Cmap answers a specific question about one centralnconcept (for instance: what is the impact of the pretreatment on the reactivity of biomass ? Howndoes enzyme diffuse into the lignocellulose ?). Hyperlinks existing between Cmaps and technicalnsheets form a network of knowledge, into which the user can navigate, to find relevant answers,nbut also associated concepts. Hyperlinks can also link Cmaps or technical sheet to an Internetnpage, scientific article and any document selected to illustrate the reality of a concept.nUp to nine knowledge areas have been identified so far, among them: biomass pretreatment;nseparation methods; enzyme cocktails; substrate reactivity, hydrolysis mechanisms. A globalnrepresentation of the overall process from wheat straw to glucose, based on the individual Cmaps,nhas been built. It includes a static structural view (environment and reactivity of the media,nencompassing the cell wall); a dynamic view (hierarchy of the different sub-processes at work)nand a functional view (description of the elementary steps and how they are organized in time).nIn the frame of the LBT III congress, the practical structuration of the knowledge and the originalnversion of the KB will be disclosed. The potential development and use of this new approach fornthe representation of biotechnology processes applied to LC will be discussed (process workflow;nunlocking cell wall recalcitrance; strategic roadmaps).

Acquisition, représentation et transfert de connaissances scientifiques via un livre de connaissances

Une methodologie a ete developpee ici pour l’acquisition, la representation et le transfert (transmission, assimilation et utilisation) de connaissances scientifiques vers les industriels. L’acquisition integre la collecte, la modelisation et la structuration des connaissances selon un format scientifique. La representation est basee sur une carte conceptuelle canonique et un format de graphe dinfluences. Les connaissances acquises et representees sont implementees dans un livre de connaissances electronique. Deux livres sur le biscuit satiete et les produits recuperes a base de bois ont ete implementes et testes suivant trois criteres : la desorientation, la charge cognitive et la comprehension des connaissances de lutilisateur au cours de l’utilisation.

INTEGRATING EXPERT KNOWLEDGE IN CEREAL FOOD MANUFACTURING PROCESSES

We present a procedure of knowledge representation based on a qualitative algebra, to predict the wheat flour dough behaviour from mixing settings. The procedure guarantees the consistency of the knowledge base and provides a concise and explicit representation of the knowledge. The qualitative model is implemented as a knowledge-based system (KBS) accessible and understandable by scientists and technologists in breadmaking. The KBS is a record of the domain knowledge, mainly know-how, and a tool to confront predictions of the dough condition with real observations. An example of such a confrontation about the wheat flour dough mixing process is shown; the results gives insight into ill-known relations between the process settings and the dough condition.Copyright