A New Transfer Learning Ensemble Model with New Training Methods for Gear Wear Particle Recognition

Aiming at solving the acquisition problems of wear particle data of large-modulus gear teeth and few training datasets, an integrated model of LCNNE based on transfer learning is proposed in this paper. Firstly, the wear particles are diagnosed and classified by connecting a new joint loss function and two pretrained models VGG19 and GoogLeNet. Subsequently, the wear particles in gearbox lubricating oil are chosen as the experimental object to make a comparison. Compared with the other four models’ experimental results, the model superiority in wear particle identification and classification is verified. Taking five models as feature extractors and support vector machines as classifiers, the experimental results and comparative analysis reveal that the LCNNE model is better than the other four models because its feature expression ability is stronger than that of the other four models.

STATE OF THE ART OF SURFACE TEXTURING FOR BIOTRIBOLOGY APPLICATIONS

Surface texturing can be defined as a technique that consists of creating micro cavities in the surface of a material. There are different texturing techniques such as chemical etching, electromechanical micromachining, diamond embossing, electric discharge, pellet-pressing and laser surface, the last being the most common. For years, these surface texturing techniques have been used in tribological applications because microcavities can store oils or fluids and constantly lubricate the surfaces that are subjected to wear; they can also be used to trap wear particles (debris) that would otherwise act as abrasive particles (three body abrasive wear). This literature review seeks to analyse and compare the advantages that the use of surface texturing techniques can offer in reducing the wear of prosthetic components and therefore lengthening their useful life, to provide a better quality of life to patients. The results of this review showed a growing interest in the scientific community in the use of surface texturing for biotribogical applications, using to a greater extent the Laser Surface Texturing (LST) technique and the surface pattern composed by dimples.

Revisiting mechanics of ice–skate friction : from experiments at a skating rink to a unified hypothesis

The mechanics underlying ice–skate friction remain uncertain despite over a century of study. In the 1930s, the theory of self-lubrication from frictional heat supplanted an earlier hypothesis that pressure melting governed skate friction. More recently, researchers have suggested that a layer of abraded wear particles or the presence of quasi-liquid molecular layers on the surface of ice could account for its slipperiness. Here, we assess the dominant hypotheses proposed to govern ice– skate friction and describe experiments conducted in an indoor skating rink aimed to provide observations to test these hypotheses. Our results indicate that the brittle failure of ice under rapid compression plays a strong role. Our observations did not confirm the presence of full contact water films and are more consistent with the presence of lubricating ice-rich slurries at discontinuous high-pressure zones (HPZs). The presence of ice-rich slurries supporting skates through HPZs merges pressure-melting, abrasion and lubricating films as a unified hypothesis for why skates are so slippery across broad ranges of speeds, temperatures and normal loads. We suggest tribometer experiments to overcome the difficulties of investigating these processes during actual skating trials.

Evidence that abrasion can govern snow kinetic friction

The long-accepted theory to explain why snow is slippery postulates self-lubrication: frictional heat from sliding melts and thereby lubricates the contacting snow grains. We recently published micro-scale interface observations that contradicted this explanation: contacting snow grains abraded and did not melt under a polyethylene slider, despite low friction values. Here we provide additional observational and theoretical evidence that abrasion can govern snow kinetic friction. We obtained coordinated infrared, visible-light and scanning-electron micrographs that confirm that the evolving shapes observed during our tribometer tests are contacting snow grains polished by abrasion, and that the wear particles can sinter together and fill the adjacent pore spaces. Furthermore, dry-contact abrasive wear reasonably predicts the evolution of snow-slider contact area and sliding-heat-source theory confirms that contact temperatures would not reach 0°C during our tribometer tests. Importantly, published measurements of interface temperatures also indicate that melting did not occur during field tests on sleds and skis. Although prevailing theory anticipates a transition from dry to lubricated contact along a slider, we suggest that dry-contact abrasion and heat flow can prevent this transition from occurring for snow-friction scenarios of practical interest.

Analysis of wear particles in used marine engine oils

В данной статье предпринята удачная попытка комплексирования магнитооптического метода и метода фотонной корреляционной спектроскопии для регистрации ферромагнитных частиц продуктов износа дизельного двигателя и определения их дисперсных характеристик. Такой подход позволяет фиксировать низкое содержание магнитных наноразмерных частиц и их дисперсность в отработанном судовом моторном масле, является новым и решает актуальную задачу. Выполнены эксперименты по наблюдению продольного магнитооптического эффекта Фарадея в модельной магнитной жидкости и в отработанном моторном масле. Показано, что низкая концентрация магнитных частиц в отработанном масле не позволяет наблюдать в нем эффект вращения плоскости поляризации в магнитном поле. Для измерения крупных частиц износа, смытых с фильтрующего элемента масляного фильтра, использовались закономерности процесса осаждения частиц в растворе и регистрация поглощения света, изменяющегося со временем. Полученные результаты позволяют получать важную информацию о состоянии и неисправностях деталей и механизмов двигателя по содержанию частиц износа в отработанном масле. In this article, a successful attempt has been made to combine the magneto-optical method and the method of photon correlation spectroscopy to register ferromagnetic particles of diesel engine wear products and determine their disperse characteristics. This approach makes it possible to record a low content of magnetic nanosized particles and their dispersion in used marine engine oil, is new and solves an urgent problem. Experiments on observation of the longitudinal magneto-optical Faraday effect in a model magnetic fluid and in used engine oil have been carried out. It is shown that the low concentration of magnetic particles in the waste oil does not allow observing the effect of rotation of the plane of polarization in a magnetic field. To measure large wear particles washed away from the filter element of the oil filter, we used the regularities of the process of particle sedimentation in solution and the registration of light absorption that changes over time. The results obtained make it possible to obtain important information on the condition and malfunctions of engine parts and mechanisms by the content of wear particles in the used oil.

Study and analysis of the morphology and composition of wear particle in the used engine oil of marine diesel engines

В статье представлены результаты исследования состава и морфологии частиц продуктов износа в работающем моторном масле судового двигателя внутреннего сгорания. Полученные данные базируется на результатах физико-химического и спектрального анализа работающего масла, идентификации дисперсного состава частиц загрязнения в нем и углубленном анализе элементного состава и формы продуктов износа пар трения судовых двигателей. Из представленных результатов следует, что основное количество частиц износа не превышает 3 мкм и представляет собой сложные сплавы, полученные в результате изнашивания трущихся сопряжений двигателя. Анализ формы частиц износа, выполненный с помощью сканирующего электронного микроскопа Lyra3 Tescan, показал что у более чем 90 % всех частиц величина Wr превышает значение 0,8. Показано на основе выполненного исследования, что спектральные методы анализа обеспечивают высокую достоверность трибомониторинга и трибодиагностики судовых дизелей. The article presents the results of a study of the composition and morphology of particles of wear products in used engine oil of a marine diesel engine. The obtained data are based on the results of physicochemical and spectral analysis of operating oil, identification of the dispersed composition of pollution particles in it, and in-depth analysis of the elemental composition and shape of wear products of friction pairs of ship engines. From results follow that the main amount of wear particles does not exceed 3 microns and they are complex alloys obtained as a result of engine wear of friction. The analysis of the shape of the wear particles, carried out using a Lyra3 Tescan scanning electron microscope, showed that more than 90% of all particles had a Wr value greater than 0.8. It is shown on the basis of the performed research that spectral methods of analysis provide high reliability of tribomonitoring and tribodiagnostics of marine diesel engines.