L’encrassement des équipements a lieu dans de nombreux secteurs industriels. Parexemple, la contamination des surfaces de contact avec les aliments, provoquée par un encrassement, entrainent des pertes économiques considérables et augmentent les risques de santé publique. Le nettoyage de l’encrassement est généralement réalisé à l’aide de produits chimiques très polluants. Par conséquent, il est important de développer des dispositifs permettant de surveiller la formation/l’élimination des encrassements sur ces surfaces (sans perturber la production) afin de réduire les risques microbiologiques, les impacts environnementaux et économiques liés aux processus de nettoyage. Dans ce travail, on s’intéresse à la détection du facteur d’encrassement à l’aide d’une méthode ultrasonore non invasive, dite “interférométrie d’ondes de coda”, abrégée en anglais (CWI). Cette technique a été testée pour différents types d’applications (Nettoyage de cire, détection de la formation de biofilm et nettoyage de dépôts protéiques). Les résultats obtenus sont très prometteurs et montrent que la CWI est capable de déceler même un léger changement du facteur d’encrassement. En particulier, l’évolution du coefficient de décorrélation pour chaque application montre une bonne concordance avec l’état d’encrassement réel de la surface. Dans l’ensemble, ces travaux fournissent un ensemble de preuves montrant que la méthode CWI, est applicable au suivi du facteur d’encrassement de dépôts sur des surfaces solides. / Fouling of equipment occurs in many industrial sectors. For example, contamination of surfaces in contact with foodstuff, caused by fouling, causes considerable economic losses and increases public health risks. The cleaning of the fouled surface is generally carried out using highly polluting chemicals. Therefore, it is important to develop devices to monitor the formation / removal of fouling on these surfaces (without disrupting production) in order to reduce the microbiological risks and environmental/economic impacts associated with the cleaning processes. In this work, the detection of fouling factor using a noninvasive ultrasonic method, called "coda wave interferometry", abbreviated in English (CWI), was investigated. This technique has been tested for various types of applications (wax cleaning, biofilm formation detection and protein deposit cleaning). The results obtained are very promising and show that the CWI is able to detect even a slight change in the fouling factor. In particular, the evolution of the decorrelation coefficient for each application shows good agreement with the actual fouling factor. Overall, this work has provided evidence that the CWI method is applicable to the monitoring of fouling factor of solid surfaces.

Membrane Distillation (MD) is an emerging sustainable desalination technique. While MD has many advantages and can be powered by solar thermal energy, its main drawback is the low water production rate. However, the MD process has not been fully optimized in terms of its manipulated and controlled variables. This is largely due to the lack of adequate dynamic models to study and simulate the process. In addition, MD is prone to membrane fouling, which is a fault that degrades the performance of the MD process. This work has three contributions to address these challenges. First, we derive a mathematical model of Direct Contact Membrane Distillation (DCMD), which is the building block for the next parts. Then, the proposed model is extended to account for membrane fouling and an observer-based fouling detection method is developed. Finally, various control strategies are implemented to optimize the performance of the DCMD solar-powered process. In part one, a reduced-order dynamic model of DCMD is developed based on lumped capacitance method and electrical analogy to thermal systems. The result is an electrical equivalent thermal network to the DCMD process, which is modeled by a system of nonlinear differential algebraic equations (DAEs). This model predicts the water-vapor flux and the temperature distribution along the module length. Experimental data is collected to validate the steady-state and dynamic responses of the proposed model, with great agreement demonstrated in both. The second part proposes an extension of the model to account for membrane fouling. An adaptive observer for DAE systems is developed and convergence proof is presented. A method for membrane fouling detection is then proposed based on adaptive observers. Simulation results demonstrate the performance of the membrane fouling detection method. Finally, an optimization problem is formulated to maximize the process efficiency of a solar-powered DCMD. The adapted method is known as Extremum Seeking (ES). A Newton-based ES is designed and the proposed model is used to accurately forecast the distilled water flux. Although good results are obtained with this method, a practical modification to the ES scheme is proposed to enhance the practical stability.

<p dir="ltr">Pea protein is one of the most used plant proteins in food products, acting as an alternative to conventional animal protein sources due to its abundant, nutritious, and ease in supply chain characteristics. The objective of this study was to investigate the impact of homogenization and UHT processing parameters on the properties of protein emulsion. Protein emulsions (8% w/w pea protein isolate and 1% w/w sunflower oil) were freshly prepared prior to processing, and the untreated sample was considered as the control (NT). The pilot-scale aseptic processing system (APS) used in this study consisted of two coil-in-shell heaters and two coolers. Samples flowed through each section of the APS system following this order: balance tank, pre-heater, final heater, hold tube, pre-cooler, and final cooler. The homogenizer was located either after the pre-cooler (AC) or the pre-heater (AH) with a controlled temperature of 165F. A third setup was utilized by bypassing the homogenizer in the UHT system. An additional 8-hour continuous run was conducted to mimic a commercial manufacturing operation by recirculating the protein emulsion in the UHT system, and fouling detections were made using a non-intrusive sensor (NICS). 5% w/w soy protein, 1% w/w sunflower oil oil-in-water emulsion was also used for fouling tests. Protein concentration, pH and zeta potential, Cryo-SEM microscopic image, particle size distribution, flocculation index (FI), coalescence index (CI), viscosity and color data were collected and analyzed. The protein concentration had a 23.20 ± 4.00 %, 28.35 ± 5.02 %, 27.98 ± 5.05% and 21.38 ± 5.75% reduction for AC, AH, UHT and NT samples, respectively, when compared with the initial concentration in the formula. AC, AH, UHT and NT samples had pH values of 7.24 ± 0.01, 7.27 ± 0.01, 7.28 ± 0.02, 7.41 ± 0.01, and zeta potential values of -42.91 ± 0.89, -47.30 ± 0.91, -46.91 ± 1.40 and -50.11 ± 1.47 mV. AC sample had a smaller and NT sample had a bigger, respectively, mean weighted size D 4,3 value than AH and UHT samples, which could also be seen in Cryo-SEM images where only AC images contained more visually observable smaller particles. FI and CI for AC, AH and UHT indicated the formation of flocs but no irreversible aggregations were found. Shear-thinning AC, AH, UHT and NT samples had viscosity decreases from 4.00 to 3.56, 3.88 to 3.75, 4.02 to 3.79 and 10.42 to 9.56 mPa*s in 1 1/s to 100 1/s shear rate range. NT sample had a very noticeable color difference from the other three treated samples. Overall, AC samples had similar or better emulsion stability in all aspects than AH and UHT samples, suggesting that AC processing could potentially be used in the protein beverage industry for manufacturing products with improved shelf stability. Severe foulants buildups were neither observed nor detected by a non-intrusive continuous sensor (NICS) in the UHT system within 8 hours of process for both pea protein and soy protein emulsion, indicating that this UHT-homogenization processing can potentially be adapted to current industrial practices for higher-quality protein beverages.</p>

Le travail de recherche présenté dans ce mémoire est une contribution à l'estimation des modèles Linéaires à Paramètres Variants (LPV) à temps continu. Dans un premier temps, il traite de façon originale la modélisation quasi-LPV des échangeurs de chaleur pour la détection d'encrassement durant les régimes transitoires. La méthodologie définie dans cette thèse est introduite pas à pas pour caractériser un échangeur à courants croisés puis pour celle d'un échangeur à contre-courants montrant ainsi la généricité de l'approche pour la modélisation des échangeurs thermiques. Une méthode locale basée sur une estimation initiale de modèles LTI en différents points du domaine de fonctionnement est utilisée pour construire le modèle LPV en agrégeant les paramètres des différents modèles locaux grâce à une méthode d'interpolation polynomiale liée aux débits massiques. Le modèle quasi-LPV de l'échangeur élaboré en fonctionnement sain permet ainsi d'envisager la détection de l'encrassement dans les échangeurs de chaleur en comparant ses sorties à ceux provenant du procédé. Dans un second temps, ce travail porte sur l'identification des systèmes LPV à représentation entrée-sortie à temps continu par une approche globale. Une solution pratique pour l'identification directe de tels systèmes est proposée. Elle est basée sur l'utilisation d'un algorithme à erreur de sortie initialisé par une approche à erreur d'équation, la méthode des Moments Partiels Réinitialisés. Des simulations illustrent les performances de l'approche proposée. / The research work presented in this thesis is a contribution to the estimation of Linear Parameter Varying (LPV) models in continuous-time. First, a quasi-LPV modeling of heat exchangers is tackled in an original way. This quasi-LPV model is meant to be used for fouling detection (during transient phases). A step-by-step description of the methodology is given on how to model a cross flow heat exchanger. Applying the same approach on a counter flow shows its genericity in the heat exchanger field. A local method, based on an estimation of LTI models for different operating points, is used to build the LPV model by interpolation of the various LTI model parameters. The resulting quasi-LPV model for a clean heat exchanger can thus be used for fouling detection by comparison of real and model outputs for the same input signals. Secondly, this work concerns the identification of continuous-time input-output LPV systems through a global approach. A practical solution for the direct identification of such systems is proposed. It is based on the use of an output-error algorithm initialized by an equation error based approach, the reinitialized partial moment's method. Simulations illustrate the performance of the proposed approach.