Currently, finite element analysis (FEA) serves as the only analytical tool used in the evaluation of seals. The FEA does not allow the dynamic analysis of the seal, which must be performed experimentally. As a result, the designing of a seal can be a costly and extensive procedure. The aim of this project has been to develop a numerical model and computer program that will have the ability to predict key seal performance characteristics, such as leakage and friction. This numerical model provides a means for evaluating potential seal designs, which can be performed without having to endure the costs of creating and evaluating the performance of the seal. Thus, the numerical model reduces the time and cost involved in evaluating seal designs. The numerical model developed differs from previous models as the effects of mixed lubrication and surface roughness are investigated. This model consists of three coupled analyses of fluid mechanics, deformation, and contact mechanics. After computational procedure has converged coupling the three analyses, auxiliary calculations are performed to obtain the quantities of leakage rate and friction force. These obtained results then allow the evaluation of the seal design, which will lead to better seal designs with lower friction and less (or no) leakage. The results obtained for a typical hydraulic seal show that the leakage characteristics depend strongly on the seal roughness.

Reciprocating rod seals are widely used in hydraulic systems to prevent the hydraulic fluid from leaking into and polluting the environment. In this research an elastohydrodynamic model of a generalized reciprocating hydraulic rod seal, including mixed lubrication and surface roughness, has been successfully developed. This model consists of coupled fluid mechanics, contact mechanics, thermal analysis and deformation analyses. Such model is capable of predicting the key seal performance characteristics, especially net leakage and friction force. This allows evaluation of potential seal designs and serves as design tools. Also as this model has been developed, the basic physics of seal operation has been clarified, which stimulates the development of innovative seal concepts, such as seals with engineered sealing surfaces. The results of this study indicate that in general, hydraulic rod seals operate in the mixed lubrication regime, although under certain conditions full film lubrication may occur over a portion of the sealing zone. The roughness of the seal surface and the rod speeds play important roles in determining whether or not a seal will leak. Cavitation during the outstroke and partial full film lubrication during the instroke tend to prevent net leakage. The behavior of a reciprocating hydraulic rod seal with a double lip or two seals in tandem arrangement can be very different from that of a similar seal with a single lip. For the double lip seal, the secondary lip can strongly affect the behavior of the primary lip by producing an elevated pressure in the interlip region. The same seal characteristics that promote effective sealing in a single lip seal and, in addition structural decoupling of multiple lips, are found to promote effective sealing in a multiple lip seal. The model is validated through comparisons of model predictions with experimental measurements and observations by industry partners. The results have shown the predicted leakage and friction force for various seal and operation conditions are consistent with the measurements. A seal with micro-pattern on the sealing surface also has been investigated. The results indicate that an elaborately designed pattern can improve the sealing characteristics of the seal, without significantly affecting the friction force. In the end, the selection of the rod seal for a specific application using this analytical model is demonstrated. The best design can be picked up before a prototype being built.

Ce travail est une contribution à l'étude du comportement des joints en translation. Un modèle numérique, basé sur la théorie inverse, est développé pour en prédire les performances. La principale contribution consiste dans le traitement des effets hydrodynamiques liés à la région d'entrée du contact. Les résultats numériques sont ensuite validés par comparaison avec lesmesures expérimentales. Ces dernières sont obtenues sur un dispositif original, conçu et réalisé au laboratoire INOE 2000 IHP de Bucarest. Ce banc reproduit les conditions réelles de fonctionnement d'un joint hydraulique et permet de mesurer la force de frottement à pression et vitesse constantes. La comparaison des résultats numériques et expérimentaux a été effectuéepour un joint de tige en forme de "U" à des pressions de service comprises entre 4 MPa et 20 MPa et pour deux vitesses de translation. Dans un premier temps, on a idéalisé le problème en considérant que les surfaces en contact tige-joint sont lisses. Si l'épaisseur du film est supérieure à la rugosité de la tige, les résultats numériques qu'on a obtenus sont en bonne corrélation avecles résultats expérimentaux. Dans le cas contraire, le modèle doit être amélioré. Cette amélioration a été entreprise de manière originale. En effet, la théorie inverse est appliquée à la distribution de pression statique du contact obtenue à partir de la simulation par éléments finis de l'assemblage d'un joint rugueux et d'un arbre lisse. La rugosité moyenne (Ra) de la surface dujoint en contact avec la tige a été choisie égale à la rugosité moyenne mesurée du joint étudié. Les résultats ainsi obtenus améliorent sensiblement la corrélation avec les mesures e / This work is a contribution to the study of reciprocating seals behavior. A numerical model, based on the inverse hydrodynamic lubrication theory, is developed to predict their performances. The main contribution consists in the treatment of the hydrodynamic effects in the entry region of contact. The numerical results are validated by comparison with experimentalones, obtained on an original experimental device, design and conceived in the laboratory INOE 2000 IHP of Bucharest. This bench reproduces actual operating conditions of a hydraulic seal and measures the friction force at constant pressure and speed. The comparison of experimental and numerical results was carried out for a "U" type rod seal at different operating pressuresvarying from 4 MPa up to 20 MPa and for two reciprocating speeds. Initially, we idealized the problem by assuming that the surfaces in contact are smooth. The numerical results obtained are in good correlation with experimental ones if the film thickness is greater than the rod roughness. Otherwise, the model must be improved. This improvement was undertaken in an original way. The inverse lubrication theory is applied to the dry frictionless contact pressure distribution, obtained from a FEM simulation of a rough seal and a smooth rod assembly. The average roughness (Ra) of the seal surface in contact with the rod is chosen equal to the measured average roughness of the studied seal. The results obtained significantly improve the correlationwith experimental measurements. The roughness distribution on the entry region of contact appears to have an important influence on the numerical results.