abolfazl foorginejad

Gas metal arc welding (GMAW) is a widespread process used for rapid prototyping of metallic parts. In this process, in order to obtain a desired welding geometry, it is very important to predict the weld bead geometry based on the input process parameters, which are voltage, wire feed rate, welding speed and welding nozzle angle. For this purpose, a global model of the welding geometry must be defined based on these parameters. Due to the non-linear and coupled multivariable relationship between the process parameters and the weld bead geometry, it is not possible to define this model in form of an explicit mathematical expression, and therefore application of supervised learning algorithms can be investigated as an efficient alternative in this problem. In this paper, application of the extreme learning machine (ELM) and support vector machine (SVM), as two efficient and powerful machine learning algorithms for predictive modelling of this process has been investigated and error analysis of the proposed models suggest that the output parameters of this process can be predicted by the ELM algorithm with higher precision and generalization capability.

The Electro-discharge coating process is an efficient method for improvement of the surface quality of the parts used in molds. In this process, Material Transfer Rate (MTR), an average Layer Thickness (LT) are important factors, and tuning the input process parameters to obtain the desired value of them is a crucial issue. Due to the wide range of the input parameters and nonlinearity of this system, the establishment of a mathematical model is a complicated mathematical problem. Although many efforts have been made to model this process, research is still ongoing to improve the modeling of this process. To this end, in the present study, three powerful machine learning algorithms, namely, Relevance Vector Machine (RVM), Extreme Learning Machine (ELM) and the Least Squares Support Vector Machine (LS-SVM) that have not been used to model this process, have been used. The values R2 above 0.99 for the training data and above 0.97 for the test data show the high accuracy and generalization capability degree related to the LS-SVM models, which can be applied for the input parameters tuning in order to attain a preferred value of the outputs.

Nowadays, simulation methods are used to obtain a preliminary estimate, in order to avoid many costing experiments. One of the common ways in which multiple commercial software has been developed based on it, is the finite element method. The present study has been tried by using the modeling of rubber in a finite element software, the patterns of propagation of longitudinal ultrasonic waves in rubber compounds are investigated. By definition of rubber as a hyperelastic material and by using of the Arruda-Boyce model for a rubber compound introduced into the finite element software, and the propagation of longitudinal ultrasonic waves was evaluated from two qualitative and quantitative perspectives. In order to quantitatively evaluate the proposed finite element model, a rubber compound was produced and the propagation velocity of the ultrasonic waves was measured. The comparison of the propagation velocity of longitudinal ultrasonic waves obtained from the finite element model and experimental data shows that the Arruda-Boyce model has a high accuracy in velocity estimation. Simulated proposed finite element is a very suitable method for interpreting and evaluating complex problems associated with the propagation of longitudinal ultrasonic waves in rubber materials, due to the accuracy of the results.

Human beings have always made their tools and instruments they need using patterns in nature. Mimicking nature has become the foundation of a new science called Biomimetics. In the present article, multiple forms and levels in nature were utilized to design and create a mouse. The rivers are a good source for choosing the shape of a mouse with lots of stones abraded through the centuries which also have smooth surfaces. In this research, a significant number of stones fitted to hand size were collected and then the best ones were scanned by an optical scanner. The point cloud model obtained was used to design and create the mouse and determine the geometric parameters of the mouse. After extracting the 3D model of the point cloud using a rapid prototyping technique with the Fused Deposition Modeling (FDM) method, some mouse models were designed ambidextrously for left-handed and right-handed people. Considering the results of the mouse evaluation by 30 people who were provided with the mouse, it can be concluded that the created mouse provided a high rate of satisfaction.

Non-destructive tests have capabilities to investigate and identify the defects and properties of the test piece without changing the physical and mechanical properties of the sample. Non-destructive ultrasonic test has been used in many researches to study different material properties such as mechanical and structural properties. The ultrasonic wave propagation velocities have been widely used for metal hardness measurement. In this study, for the first time nondestructive ultrasonic testing has been employed to measure the hardness of rubber compounds using Gaussian Process Regression. Eighty seven samples with different formulations were prepared and vulcanized. After the vulcanization the compound hardness of the samples and the longitudinal ultrasonic wave velocity through them was measured. Result of Gaussian Process Regression model show that this model perform well and able to predict tire hardness. Also investigating repeatability shows that this method can be a good alternative for conventional hardness testing method in measuring hardness of rubbers. According to low time of test in this method and no need to sample preparation, the proposed method can be used in tire production lines.

Additive Manufacturing describes the technologies that can produce a physical model out of a computer model with a layer-by-layer production process. Additive Manufacturing technologies, as compared to traditional manufacturing methods, have the high capability of manufacturing the complex components using minimum energy and minimum consumption. These technologies have brought about the possibility to make small pieces of raw materials in the shortest possible time without the need for a mold or tool. One of the technologies used to make pieces of the layer-by-layer process is the Gas Metal Arc Welding (GMAW). One of the basic steps in this method of making parts is the prediction of bead geometry in each pass of welding. In this study, taking into account the effective parameters on the geometry of weld bead, an empirical study has been done in this field. For this purpose, three parameters of voltage, welding speed and wire feeding rate are considered as effective parameters on the welding geometry of the process. Width and height of the bead are also determined by the parameters of the geometry of the weld according to the type and application of the research as output parameters are considered. In this paper, an adaptive neuro-fuzzy inference system (ANFIS) is used to create an adaptive model between input process data and parameters of weld bead geometry. The least squares mean error is used to evaluate the model. The predicted results by the model have a good correlation with the experimental data.

In tire industry, it is very crucial to evaluate physical and mechanical properties of the rubber which is used for production of the tire, to ensure the quality of the final product. Resilience is an important property of a rubber, which cannot be evaluated through direct measurement in production cycle in this industry. Therefore, non-destructive ultrasonic testing, which has been used in many applications for examination of various material properties, can be used as an alternative approach for this purposes. In this study, the non-destructive ultrasonic testing method has been employed to investigate the resilience of nanoclay reinforced rubber compounds. By changing Physical and mechanical properties of materials, ultrasonic wave velocities are changed. For this purpose, sixteen different samples of nanoclay reinforced rubber compounds with different formulations were prepared and both their resilience and the longitudinal ultrasonic wave velocity through them were measured. At the next step, using the relevance vector machine regression analysis, a mathematical expression for the rubber resilience based on the longitudinal ultrasonic wave velocity was developed, which was proven to be qualified with acceptable accuracy and generalization capability. The results of this research can be used for online evaluation of the rubber resilience in tire production cycle.

Saffron is the most expensive agriculture crop and spice in the world. There is very few information about physical and geometric properties of Saffron flower and its parts in academic resources. Due to the lack of a virtual model of Saffron flower it is difficult to analysis and design a post-harvesting mechanism. By generation of a 3D geometrical model of the flower as a free form, not only obtained geometric parameters of flower, but also a more accurate mechanical behavior of the flower including its aerodynamical behavior becomes possible. In this article, using reverse engineering, 3D data of Saffron flower were extracted by employing laser scanning technology, after data pre-processing and processing, the 3D model of saffron flower as a free form was developed. Because of non-rigidity and flexibility of saffron flower and the need for scanning the hidden components including anther and stigma, there is no possibility for integrated scanning of the flower. Therefore, every component of flower must be scanned individually followed by joining various components of saffron flower, after which the whole free form rigid 3D model of saffron flower was developed. Because of the very low thickness of the petals, the extracted point clouds are interfering, and using the commercial software available in the market does not allow direct modeling of the flower using the raw scanned data. An algorithm was proposed for addressing the problem of interference in point cloud and to separate point cloud of the top surface and the bottom surface of the petal named as “projected homogenous neighbors”. The algorithm is capable of solving some problems in geometric modeling of plants and flowers.

Abrasive water jet cutting process can produce tapered edges on cutting kerf. This problem can limit the applications of abrasive water jet cutting process and in some cases it is necessary another edge preparation process. In this paper, an experimental investigation kerf characteristics of Ti-6Al-4V titanium alloy under abrasive water jet cutting is presented. In this regards, it is shown how to use the hybrid approach of Taguchi method and principal component analysis to optimize abrasive water jet cutting are used in this paper. The abrasive water jet cutting process input parameters effect on material removal rate and the characteristics of the surface. A considerable effort was made in understanding the influence of the system operational process parameters such as water jet pressure, traverse speed, abrasive flow rate, and standoff distance. Due to appropriate selecting abrasive water jet cutting process parameters leads to optimizing of kerf characteristics include top kerf width, kerf tapper and kerf deviation, therefore it is important to select appropriate input parameters. The obtained results from this method show that the hybrid approach of Taguchi method and principal component analysis is a suitable solution for optimizing of abrasive water jet cutting process.

Point based 3D modeling has recently received greater attention، mainly due to its simplicity. One of the most fundamental operations for point set processing is to find the neighbors of each point in point clouds. This paper presents a new method called homogeneous neighborhood for determining neighbors in point clouds. This method of choosing neighbors، in addition to the distance takes into consideration the directional balance by improving the k nearest neighbors. The directional balance describes whether the neighbors are well spread around the point of concern. In this study effects of selecting neighbors on normal vector estimation are investigated. Normal vector is calculated using homogeneous neighborhood. For evaluation of the proposed method in determining neighbors، normal vector are calculated using the k nearest neighbors. The results show that the homogeneous neighborhood method is more accurate in normal vector estimation than the k nearest method. For evaluation of the homogeneous neighborhood method، it was employed in point cloud registration application. The results of registration by using the homogeneous neighborhood show that this method of neighbor selection yields reduced registration errors.

In this paper، it is shown how to use the recently developed Firefly Algorithm to optimize abrasive water-jet cutting as a nonlinear multi-parameter process. Back propagation neural network were developed to predict surface roughness in abrasive water-jet cutting (AWJ) process. In the development of predictive models، machining parameters of traverse speed، water-jet pressure، standoff distance and abrasive flow rate were considered as model variables. Firefly Algorithm by using back propagation neural network optimizes glass surface roughness in abrasive water-jet cutting and proposes appropriate parameters for minimum surface roughness. Testing results demonstrate that the model is suitable for predicting the response parameters. However this algorithm has not be tested for practical problems، the results showed this algorithm applicable for processes with complex nature.