Etude du nettoyage d'un stérilisateur de lait U.H.T. Ordre d'utilisation des détergents alcalin et acide et aspects cinétiques

Abstract - Cleaning of milk UHT sterilizer: Kinetics aspects with regard to acid and alkaline detergents
One of the main disadvantages of indirect U.H.T. sterilization of milk is fouling of heat transfer surfaces. Energy consumption generally increases since pressure drops grow and heat transfer coefficients lower. Three times a day sterilizers need a cleaning in order to bring the exchange surface back to its original state. This procedure is expensive (immobilization time, thermal and mechanical energy consumptions, detergent costs...) and optimization of the cleaning process is only possible with a right knowledge of cleaning dynamics. The objective of this work is first to determine the effect of alkaline and acid cleaning according to the chemical composition of formed deposits (protein, fat and minerals) and secondly to evaluate the dynamic of the cleaning process.
Cold raw bulk milk was prepasteurized at 82.5° C, then sterilized in a plate heat exchanger (total exchange area of 10 m²) at a flow rate of 400 l / h. This sterilizer was comprised of a preheating section (E), a heating section (F) and a precooling section (H). Milk was homogenized at 68°C during preheating and sterilized at 138°C. At the end of each trial, the sterilizer was water rinsed and dismantled to weight the amount of deposit on each plate. One plate of each pass was kept in order to take samples for protein, fat and minerals analysis. The E and F sections were studied separately for cleaning. During the cleaning each plate used for sampling was replaced by a clean identical plate in order to keep the same configuration. Each section received two cleaning procedures. The first cleaning procedure (S-A) used first alkaline solution (sodium hydroxide 2 %) then acid solution (nitric acid 2 %), while the second cleaning procedure (A-S) used the same cleaning solutions but in reverse order. Cleaning temperature and flow rate were respectively 90° C and 400 l / h. Cleaning kinetics were studied by sampling automatically cleaning solutions at the exit of the studied section to perform protein and minerals analysis. From these measurements cleaning rate were deduced. A differential pressure sensor was also used to determine on line the colmation level of the exchanger. By the assumption that the deposits of the two plates of each pass had the same composition, the total dry mass removed by the cleaning solutions was deduced and compared to the initial mass to be cleaned. From these results, it can be concluded that for E section, cleaning procedure (S-A) and cleaning procedure (A-S) led to correct cleaning, but cleaning is quicker for procedure (A-S). For F section, only cleaning procedure (A-S) is correct. During cleaning procedure (S-A) protein matter is not totally removed and when the apparatus is dismantled the heat exchange surface appeared still fouled in some parts. In all procedures, fat is supposed to be cleaned with protein. A study of cleaning rates was done. The apparent cleaning rate offered the same evolution vvhatever the cleaning procedure and whatever the matter to be cleaned. It was null at zero time, increased to a maximum, then decreased to zero. This trend is the same obtained by GALLOT-LAVALLÉE (1982) though our trials concerned an exchange surface 100 times more important on which the mass was not uniformly distributed. By fitting the two-steps GALLOT-LAVALLÉE model to our study, the velocity constants of each step were determined. In the preheating section E, during removal of protein by the alkaline solution the removing rate were significantly increased by using the cleaning procedure (A-S). On the other hand, during removal of minerals by the acid solution, each cleaning procedure acted differently on each step of the model. Efficient and specific cleaning times were defined to choose the best procedure for practical applications.

Résumé - Un des principaux inconvénients du procédé de stérilisation U.H.T. indirect du lait dans des échangeurs à plaques est l'encrassement des surfaces d'échange thermique. Trois fois par jour, les stérilisateurs à plaques doivent être nettoyés afin de rendre aux surfaces d'échange leur capacité thermique et leur propreté initiales. Mais compte tenu du coût élevé de ces nettoyages, il est indispensable d'avoir une parfaite connaissance des phénomènes afin de pouvoir les optimiser.
L'objectif de notre travail est premièrement de déterminer les effets des nettoyages alcalins et acides sur les différents constituants chimiques des dépôts formés lors de l'encrassement et deuxièmement d'évaluer les cinétiques d'enlèvement de ces composés. Chaque section du stérilisateur est étudiée séparément selon deux procédés différents. Le premier (S-A) consiste à nettoyer d'abord avec l'hydroxyde de sodium (2 %) puis avec l'acide nitrique (2 %) alors que pour le second (A-S), les mêmes solutions sont utilisées mais dans l'ordre inverse. Les conditions de nettoyage sont : 90° C pour la température et 400 l / h pour le débit. Les cinétiques de nettoyage sont étudiées par échantillonnage automatique des solutions nettoyantes à la sortie de la section étudiée afin d'analyser les quantités de matières protéiques, minérales et grasses éliminées et leur vitesse d'élimination.
Sur la section de préchauffage, les deux procédures de nettoyage sont tout à fait satisfaisantes (élimination totale de la souillure macroscopique) mais le procédé A-S est plus rapide. Pour la section de chauffage seul le procédé A-S est efficace ; dans l'autre cas, la totalité des protéines n'a pas été éliminée et à l'ouverture de l'appareil, nous constatons que la surface d'échange est encrassée aux points de contact des plaques d'échange. Des hypothèses concernant la structure des dépôts et leur mécanisme d'élimination sont discutées.