Improving Lubricant Formulations Using Machine Learning

Purpose This paper aims to find the best method to predict the friction coefficient of textured 45# steel by comparing different machine learning algorithms and analytical calculations. Design/methodology/approach Five machine learning algorithms, including K-nearest neighbor, random forest, support vector machine (SVM), gradient boosting decision tree (GBDT) and artificial neural network (ANN), are applied to predict friction coefficient of textured 45# steel surface under oil lubrication. The superiority of machine learning is verified by comparing it with analytical calculations and experimental results. Findings The results show that machine learning methods can accurately predict friction coefficient between interfaces compared to analytical calculations, in which SVM, GBDT and ANN methods show close prediction performance. When texture and working parameters both change, sliding speed plays the most important role, indicating that working parameters have more significant influence on friction coefficient than texture parameters. Originality/value This study can reduce the exper... Read More

Source

Purpose The lubrication performance of static and dynamic pressure thrust bearings is improved by introducing texture on the sealing edge. Design/methodology/approach Through model building, meshing and boundary condition setting, the influence of square texture on oil film lubrication performance was simulated and analyzed, and an improved algorithm was applied to perform optimization of lubrication performance. Findings The findings of this study reveal that the optimum lubrication performance is attained when adjusting the parameters of the square texture to 0.12 mm, 0.1 mm, 1 mm and 34 mm. In such circumstances, the thrust bearing with square textures demonstrates an increase in loading capacity of around 19% and a temperature reduction of about 2C compared to a smooth thrust bearing. Originality/value The original Reynolds equation is revised, and the influence of square texture on the physical field of oil film is analyzed, considering the turbulence state and cavitation phenomenon. The multi-objective function under square texture parameters was established using BP neural ne... Read More

Source

This paper investigates the use of neural networks to predict characteristic parameters of the grease application process pressure curve. A combination of two feed-forward neural networks was used to estimate both the value and the standard deviation of selected features. Several neuron configurations were tested and evaluated in their capability to make a probabilistic estimation of the lubricants parameters. The value network was trained with a dataset containing the full set of features and with a dataset containing its average values. As expected, the full network was able to predict noisy features well, while the average network made smoother predictions. This is also represented by the networks R2 values which are 0.781 for the full network and 0.737 for the mean network. Several further neuron configurations were tested to find the smallest possible configuration. The analysis showed that three or more neurons deliver the best fit over all features, while one or two neurons are not sufficient for prediction. The results showed that the grease application process pressure cur... Read More

Source

The chemical properties of oils are vital in the design of microemulsion systems. The hydrophilic-lipophilic difference equation used to predict microemulsions' phase behavior expresses the oils' physiochemical properties as the equivalent alkane carbon number (EACN). The experimental determination of EACN requires knowledge of the temperature dependence of the microemulsion system and the effects of different surfactant concentrations. Thus, the experimental determination is time-intensive and tedious, requiring days to months for proper separations. Furthermore, the experiments require high purity of chemicals because microemulsions are sensitive to impurities. Our work focuses on the quick and reliable predictions of the EACN with machine learning (ML) models. Due to the immaturity of ML chemical predictions, we compare three graph neural networks (GNNs) and a gradient-boosted tree algorithm, known as XGBoost. The GNNs use the molecular structures represented as simplified molecular-input line-entry system (SMILES) codes for the initial input, which allows us to assess whether geo... Read More

Source

Abstract This study explored the capability of semi-empirical and neural network approaches for correlating and predicting some equilibrium and non-equilibrium thermophysical properties of liquid lubricants. The equilibrium properties, including the densities and several thermodynamic coefficients for 12 liquid lubricants, were correlated and predicted through a perturbed hard-chain equation of state (PHC EoS) by an attractive term of Yukawa tail. The molecular parameters of PHC EoS were obtained by correlating them with 935 data points for the densities and isothermal compressibilities of studied systems in the 278353 K range and pressure up to 70 MPa with the average absolute relative deviations (AARDs) of 0.36 % and 5.25 %, respectively. Then, that EoS was employed to predict the densities of other literature sources (with an AARD of 0.81 %) along with several thermodynamic coefficients, including isobaric expansivities (with an AARD of 12.92 %), thermal pressure coefficients (with the AARD of 12.93 %), and internal pressure (with the AARD of 13.67 %), for which the reference val... Read More

Source

Elastohydrodynamic lubrication (EHL) is crucial for the longevity and performance of machine parts subjected to high loads and speeds. Analyzing EHL thickness settings with accuracy is crucial for improving construction and mitigating premature wear. Current approaches for forecasting such variables might be limited in their precision due to computation difficulties. This study presents a novel remora-optimized Gaussian process regression (RO-GPR) approach in calculating the factors related to EHL film dimension, addressing the requirement for a more effective and precise predictive model. The suggested RO-GPR approach is trained and verified by utilizing a dataset that was acquired through methods that are either simulation-based or practical. To show how well the suggested RO-GPR strategy handles the complex structure of EHL mechanisms, its efficacy is compared with that of other approaches. The study's findings reveal that the suggested RO-GPR model can predict EHL film thickness properties with an excellent level of accuracy, indicating its potential as a useful tool for tribolog... Read More

Source

Abstract Liquid formulations are ubiquitous yet have lengthy product development cycles owing to the complex physical interactions between ingredients making it difficult to tune formulations to customer-defined property targets. Interpolative ML models can accelerate liquid formulations design but are typically trained on limited sets of ingredients and without any structural information, which limits their out-of-training predictive capacity. To address this challenge, we selected eighteen formulation ingredients covering a diverse chemical space to prepare an open experimental dataset for training ML models for rinse-off formulations development. The resulting design space has an over 50-fold increase in dimensionality compared to our previous work. Here, we present a dataset of 812 formulations, including 294 stable samples, which cover the entire design space, with phase stability, turbidity, and high-fidelity rheology measurements generated on our semi-automated, ML-driven liquid formulations workflow. Our dataset has the unique attribute of sample-specific uncertainty measurem... Read More

Source

Purpose This study aims to propose an optimization framework using deep neural networks (DNN) coupled with nondominated sorting genetic algorithm II and technique for order preference by similarity to an ideal solution method to improve the tribological properties of camshaft bearing pairs of internal combustion engine. Design/methodology/approach A lubrication model based on the theory of elastohydrodynamic lubrication and flexible multibody dynamics was developed for a V6 diesel engine. Setting DNN model as fitness function, the multi-objective optimization genetic algorithm and decision-making method were used to optimize the bearing pair structure with the goal of minimizing the total friction loss and the difference of the average values of minimum oil film thickness. Findings The results show that the lubrication state corresponding to the optimized bearing pair structure is elastohydrodynamic lubrication. Compared with the original structure, the optimized structure significantly reduces the total friction loss. Originality/value The optimized performance and corresponding str... Read More

Source

Abstract The need to move toward more sustainable lubricant materials has sparked an ever growing interest on the tribological performances of additives based on environmentally friendly molecules, such as carbon-based compounds, and green liquid media as aqueous solutions. The prediction of the solubility of the additives into the liquid and the tribochemistry of decomposition and polymerization of the additive molecules under harsh conditions is essential for understanding the atomistic mechanisms leading to the formation in situ of the carbon-based lubricious tribofilms so effective in reducing friction and wear at solid interfaces. To this extent, the application of tools like ab initio molecular dynamics based on first-principle density functional theory is severely hindered by the size of the systems of interests and the need to simulate their dynamics over relatively long times. To enable tribological simulations with quantum accuracy for a first time, we develop a workflow for smart configuration sampling in active learning, to obtain machine learning interatomic potentials w... Read More

Source

Minimum quantity lubrication has proven to be more effective and efficient than dry and wet lubrication during machining of materials. However, components surface roughness and tool wear are still of the major concerns affecting the implementation of the said technology in industry. Prediction of these parameters helps to optimise cutting conditions and/or MQL specific parameters. Although various methods are used in the prediction of surface roughness and tool wear such as dimensional analysis, ANOVA etc., present work aims to develop the model for surface roughness and tool wear using artificial neural network. The proposed estimator is based on a neural network with cutting conditions and MQL specific parameters as the input parameters and surface roughness and tool wear as the predicted parameters. The feed forward network that provides ten hidden neurons has been selected. The coefficient of correlation shows good fit with the training, testing, focused and confirmation data sets. The results give encouraging results while using neural network model for the specified conditions ... Read More

Source

The intricate development of liquid-crystal lubricants necessitates the timely and accurate prediction of their tribological performance in different environments and an assessment of the importance of relevant parameters. In this study, a classification model using Gaussian noise extreme gradient boosting (GNBoost) to predict tribological performance is proposed. Three additives, polysorbate-85, polysorbate-80, and graphene oxide, were selected to fabricate liquid-crystal lubricants. The coefficients of friction of these lubricants were tested in the rotational mode using a universal mechanical tester. A model was designed to predict the coefficient of friction through data augmentation of the initial data. The model parameters were optimized using particle swarm optimization techniques. This study provides an effective example for lubricant performance evaluation and formulation optimization.

Source

Machine learning-based predictive models allow rapid and reliable prediction of material properties and facilitate innovative materials design. Base oils used in the formulation of lubricant products are complex hydrocarbons of varying sizes and structure. This study developed Gaussian process regression-based models to accurately predict the temperature-dependent density and dynamic viscosity of 305 complex hydrocarbons. In our approach, strongly correlated/collinear predictors were trimmed, important predictors were selected by least absolute shrinkage and selection operator (LASSO) regularization and prior domain knowledge, hyperparameters were systematically optimized by Bayesian optimization, and the models were interpreted. The approach provided versatile and quantitative structure-property relationship (QSPR) models with relatively simple predictors for determining the dynamic viscosity and density of complex hydrocarbons at any temperature. In addition, we developed molecular dynamics simulation-based descriptors and evaluated the feasibility and versatility of dynamic descri... Read More

Source

Abstract Machine learnings prowess in extracting insights from data has significantly advanced fluid rheological behavior prediction. This machine-learning-based approach, adaptable and precise, is effective when the strategy is appropriately selected. However, a comprehensive review of machine learning applications for predicting fluid rheology across various fields is rare. This article aims to identify and overview effective machine learning strategies for analyzing and predicting fluid rheology. Covering flow curve identification, yield stress characterization, and viscosity prediction, it compares machine learning techniques in these areas. The study finds common objectives across fluid models: flow curve correlation, rheological behavior dependency on variables, soft sensor applications, and spatialtemporal analysis. It is noted that models for one type can often adapt to similar behaviors in other fluids, especially in the first two categories. Simpler algorithms, such as feedforward neural networks and support vector regression, are usually sufficient for cases with narrow ... Read More

Source

Choosing an appropriate lubricating oil replacement strategy is crucial for the machine's operation and maintenance. Based on the concept of condition-based maintenance (CBM), this article proposes a method for predicting the remaining useful life (RUL) of lubricating oil using lubrication condition monitoring (LCM) data and machine learning (ML) theory. Firstly, obtain lubricating oil samples through engine bench tests and quantitatively analyze the elemental content of the lubricating oil in use using atomic emission spectroscopy (AES). Then, a method for finding the optimal back propagation (BP) neural network was proposed to construct a lubricating oil RUL prediction model. The content of 12 elements in lubricating oil is used as the input variable, and the three states of lubricating oil are used as the output variable. Finally, by comparing with the lubricating oil RUL prediction model based on support vector machine (SVM), it is shown that the proposed optimal BP neural network model has better accuracy and robustness.

Source

Opportunities and challenges in data-driven chemical engineering thermodynamics, statistical mechanics and molecular simulation are discussed, and new possibilities offered by machine learning in these areas are assessed. Examples suggest how integration of data science and molecular simulation can prove impactful for the future of chemical engineering.

Source

Friction behaviour is an important characteristic of dynamic seals. Surface texturing is an effective method to control the friction level without the need to change materials or lubricants. However, it is difficult to put the manual prediction of optimal friction reducing textures as a function of operating conditions into practice. Therefore, in this paper, we use machine learning techniques for the prediction of optimal texture parameters for friction optimisation. The application of pneumatic piston seals serves as an illustrative example to demonstrate the machine learning method and results. The analyses of this work are based on experimentally determined data of surface texture parameters, defined by the dimple diameter, distance, and depth. Furthermore friction data between the seal and the pneumatic cylinder are measured in different friction regimes from boundary over mixed up to hydrodynamic lubrication. A particular innovation of this work is the definition of a generalised method that guides the entire machine learning process from raw data acquisition to model predictio... Read More

Source

The viability of employing soft computing models for predicting the viscosity of engine lubricants is assessed in this paper. The dataset comprises 555 reports on engine oil analysis, involving two oil types (15W40 and 20W50). The methodology involves the development and evaluation of six distinct models (SVM, ANFIS, GPR, MLR, MLP, and RBF) to predict viscosity based on oil analysis results, incorporating metallic and nonmetallic elements and engine working hours. The primary findings indicate that the radial basis function (RBF) model excels in accuracy, consistency, and generalizability compared with other models. Specifically, a root mean square error (RMSE) of 0.20 and an efficiency (EF) of 0.99 were achieved during training and a RMSE of 0.11 and an EF of 1 during testing, utilizing a 35-network topology and an 80/20 data split. The model demonstrated no significant differences between actual and predicted datasets for average and distribution indices (with P-values of 1.00). Additionally, robust generalizability was exhibited across various training sizes (ranging from 50 to 80... Read More

Source

Abstract Conventional liquid lubricants prove inadequate for effective lubrication in conditions characterized by high temperatures and high vacuum environments. In such extreme scenarios, powder lubricants emerge as a more viable solution. The present study is to conduct a series of experiments using a reciprocating wear test setup and evaluate the capability of four different machine learning models in analysing the tribological attributes of metals when lubricated with three distinct powder types: zirconium dioxide, copper oxide, and molybdenum disulfide, specifically under conditions of elevated contact pressures and dry environments. The experiments were systematically carried out encompassing a range of load and temperature combinations. Four different machine learning models (MLP, KNN, extreme gradient boosting and light gradientboosting machine) were used for predicting the coefficient of friction of metals lubricated with different powders. Extreme gradient boosting machine learning model gives better result than the other models with mean absolute error, root mean squared ... Read More

Source

The tribological performance of lubricant is significantly affected by additive concentration. To realize optimization of additive concentrations, an eXtreme Gradient Boosting machine learning method was proposed to predict the tribological performance of a lubricant, reflected by wear volume, with data collected from four-ball friction experiments. The Shapley Additive exPlanation tool was used to explain the predictions of the model. Particle swarm optimization was used to optimize additive concentration proportions. The tribological properties of grease additives were verified through validation experiment. The chemical components of wear scars include MXenes, carbon films, iron oxides, and iron phosphides, which facilitate direct contact between friction surfaces. This study provides a novel data-driven approach for optimization of additive concentrations with effective and predictable tribological performance.

Source

Tribology is the study on friction, wear and lubrication for contacting and moving two solid materials with fluid lubricants. Friction can be found easily in our life and should be reduced, and lubricants are a powerful answer for this purpose. The mechanism behind friction and lubrication has been primarily analyzed by the Stribeck curve, which should be 1) interpretable, 2) non-homogeneous in variances and 3) a mixture of Stribeck curves further. We propose a machine learning approach, Estribec, to estimate the Stribeck curves from observed data in tribology. Estribec is, considering all above three characteristics, a finite mixture model, with a component of a piecewise function, where each piece is a comprehensible simple (primarily, linear) function with a unique variance. Entirely our method keeps linear time complexity for all processes, including parameter estimation and prediction. Empirical results with synthetic data showed Estribec achieved favorable predictive performance against Gaussian process regression (GPR) and its tree variant (TGPR). Importantly Estribec ran alwa... Read More

Source