Search

Automated inspection of surface defects is beneficial for casting product manufacturers in terms of inspection cost and time, which ultimately affect overall business performance. Intelligent systems that are capable of image classification are widely applied in visual inspection as a major component of modern smart manufacturing. Image classification tasks performed by Convolutional Neural Networks (CNNs) have recently shown significant performance over the conventional machine learning techniques. Particularly, AlexNet CNN architecture, which was proposed at the early stages of the development of CNN architectures, shows outstanding performance. In this paper, we investigate the application of AlexN et CNN architecture-based transfer learning for the classification of casting surface defects. We used a dataset containing casting surface defect images of a pump impeller for testing the performance. We examined four experimental schemes where the degree of the knowledge obtained from the pre-trained model is varied in each experiment. Furthermore, using a simple grid search method we explored the best overall setting for two crucial hyperparameters. Our results show that despite the simple architecture, AlexN et with transfer learning can be successfully applied for the recognition of casting surface defects of the pump impeller.

The paper is devoted to a widespread metallurgical process of filling intermediate ladle (tundish) by melt from falling jet with argon. The tundish is used to stabilize melt-feeding of a continuous casting machine. Additionally in tundish argon removes remains of non-metallic inclusions. There is necessity of low melt speed near outflow holes of tundish to increase homogeneity of steel casting product. As experimental feature it is added cored wire injection from vertical guiding tube near falling jet. The feature gives last chance for modification of chemical contents of melt before casting. Mathematical modeling can significantly decrease cost of conducting experiments as well as avoid complexities of plant trials such as opacity of melt and its temperature. The mathematical model is based on the Navier-Stokes equations, which are used for melt motion, and – on the Euler-Lagrange equations, which are used to predict cored wire motion. It is given parameters of numerical experiment. Presented figures show calculated velocity profiles at horizontal and vertical planes. Adequacy checking of computed melt hydrodynamics using data in scientific literature melt shows satisfactory error level. That leads to recommendation to use proposed mathematical model for further numerical investigations of the process.

This paper discusses the problem of avoiding threats during the cruising flight of hypersonic vehicles (HV). Considering the constraints on kinematics of HV and changing environments, this paper proposes two methods of trajectory planning that taking the overload or the rotational angular velocity of the ballistic deflection angle as actions of agent. Meanwhile, the agent’s policy is optimized with policy gradient method, and Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC) are used for comparison. Experimental results show that PPO and SAC have similar performance in penetration missions. Moreover, in the complicated flight environment, the method of taking overload and exploration distance as actions has a higher penetration success ratio.

Austempered ductile iron (ADI) has recently gained popularity over forged and cast steels for manufacturing of components used in heavy machinery, such as, automobiles, railways and farm equipment due to its high strength to weight ratio, casting capabilities and other attractive properties. The current industry practice for producing ADI is through the use of ADI standards. These standards specify grades of ADI with strength and hardness specified with a range for each grade. This grading system leads to precious alloying elements being wasted and also leads to increase in cost. In this study, a multilayer perceptron (MLP)-based forward modeling scheme is proposed to simulate the ADI process and accurately predict Vickers hardness number (VHN) of ADI based on the austempering process parameters and alloying element proportions. Then, using the trained MLP-based forward model and the particle swarm optimization (PSO) algorithm, a novel PSO-MLP-based inverse model has been proposed, which can predict the austempering process parameters and the alloying element proportions for a given VHN. With extensive simulation results, we have shown that MLP-based forward model can predict VHN of ADI with a mean absolute percent error of 0.21, which is much lower than a recently reported paper and the inverse model can accurately predict the ADI process parameters for a given VHN. This model can reduce wastage and cost by producing ADI with a specific VHN rather than within a range.

Samples of aluminum-silicon eutectic alloys were cast via stir casting method to investigate their wear behavior. Three types of samples were prepared. The first solidified under the effect of mechanical vibration (using an amplitude of 0.5 mm and a frequency of 25 Hz); the second has an addition of 0.13 wt% titanium as alloying element; while the third cast with the combined effect of added titanium and mold vibration. Chemical composition test, scanning electron microscope (SEM) analysis, macro-hardness and wear tests were done for the prepared samples. The results showed that at 27 oC for 25 minute loading with a load of 20 N, the lowest wear rate of sample prepared with the combined effect of mold vibration and addition of titanium, the biggest wear rate for sample with the effect of vibration, and the wear rate of sample prepared with addition of titanium between them. The wear rate at 100 oC is lesser than that at 27 oC for all samples at the same conditions.

Traditional casting character recognition algorithms need to select appropriate position features for different scenes in the character location step, so it is difficult to realize the recognition task of casting embossed concave and convex characters in the different distribution in complex scenes. In this letter, a recognition method of casting embossed characters based on YOLO v5 is proposed. The fast and reliable depth learning algorithm YOLO v5 is used to automatically extract the image features and realize the recognition of casting embossed characters (including numbers and letters) Recognition. The experimental results show that the accuracy of the network model for steel seal character recognition is higher than traditional computer vision algorithms, the average processing time of the algorithm is quickly, and the weight file volume is small, which meets the accuracy and efficiency requirements of engineering application.

Rapid prototyping of Polydimethylsiloxane (PDMS) chips is essential in iterative design processes, but many of the fabrication steps require handcrafting and haven't been adapted to automated manufacturing. This research proposes a novel method of manufacturing high-precision PDMS chips, named “soft casting”. A key component is a multifunctional removable cover that is non-adhesive, gas permeable, and semi-rigid. Using this method, PDMS chips with precise shape and thickness can be manufactured. In addition, the “soft casting” method can produce a large number of through holes and eliminate the need for chip cutting and drilling. This method greatly improved the quality of the PDMS chips, increased batch-to-batch reproducibility, reduced fabrication time, lowered PDMS consumption, and is expected to be useful for automated production.

Casting roller is equipped with high maintenance frequency and interchangeability in continuous casting machine, viewed as key part during its operation and maintenance. To improve the level of casting roller operation & maintenance management and prolong its service life, a novel framework for remaining useful life prediction of casting roller in continuous casting machine based on digital twin is proposed. First, systematic overview of the proposed framework is described in detail. Second, monitoring and traceability for casting roller are designed based on 3D twin model. Then, multiple failure modes of casting roller are analyzed and its performance degradation is evaluated. After that, the process of remaining useful life prediction based on deep learning together with twin data is given. Finally, a prototype system is developed and implemented, and the results verify the effectiveness and feasibility of proposed framework.

In the process of steel smelting, there are restrictions such as different refining processes and variable equipment transportation time. The existing steelmaking-continuous casting production scheduling methods are mainly for the simple production modeling and analysis, which are difficult to effectively solve the practical problems. In this paper, an optimal scheduling mathematical model is constructed for the problems of different refining stage paths and equipment transportation time between adjacent charges in a large domestic steelmaking plant. On the basis of this model, a composite method is proposed for the problems of different refining stages, charge transportation waiting time and the continuous casting process without waiting time. Comparing with simulated annealing (SA), particle swarm optimization (PSO) and the proposed composite method, experiments show that the composite Forward-Reverse method can avoid waiting time, ensure no overlap in the process, and obtain the minimized legal scheduling results in the case of stage skipping.

Dynamic control of secondary cooling in continuous casting is the key technology to improve the quality and efficiency of the slab. In order to improve the quality and efficiency of the slab, a new control strategy of secondary cooling water distribution based on an improved artificial bee colony algorithm is proposed. Based on the traditional artificial bee colony algorithm, adding the learning factor to the optimal individual learning factor and random individual of the current population and integrating the output overshoot mechanism into the objective function is effective to avoid the great overshoot phenomenon caused by the abruptly change of casting speed and the large fluctuation of casting surface temperature. The transfer function was constructed based on the coupling relationship between the distribution of secondary cooling water and slab surface temperature. The Simulink simulation model was established, which was compared with the PSO-DSWPI Method and conventional PID controller. The results showed that the improved artificial bee colony algorithm has a more minor error, higher control accuracy and more smooth step response, which provided a new method for realizing more accurate control of secondary cooling dynamic water distribution and had practical significance for improving slab quality.

This paper analyzes a system for more efficient automatic feed of the mold flux to the crystallizer in a continuous casting machine (an automatic feeder). It describes why a mold flux is necessary for steel smelting. It further presents a unit that feeds the mold and describes its algorithm as developed by the authors. Additionally, the paper summarizes the functions of the developed SCADA system that provides automatic control over the operations of the system. The research team developed a simulation model for the electric drive of the auger dosing unit. The model contains an electromechanical block that is responsible for the functioning of the dosing unit. The asynchronous motors driving the auger are controlled by a frequency converter that uses scalar control based on the active stator-current component. The model output is compared to the experimental data. The automated mold feeder will reduce the defect occurrence rates by 3% while also reducing the MF feed rate by 6%.

This paper proposes relevant intelligent computing solutions for the specific business of a large steel plant in Tangshan. The company adopts the “one iron ladle to the end” production mode, which improves the production efficiency and puts forward higher requirements for the transportation and scheduling of hot metal at the same time. However, the traditional transportation scheduling mode is obviously unable to meet the needs of production, which is not only inefficient, but also easy to affect the smooth progress of production. At this time, a new scheduling method is needed to meet the needs of the company’s hot metal scheduling and management, to minimize the waiting time of hot metal in the way, and to ensure the hot metal temperature and continuous casting requirements for steel-making. In this paper, a positioning technology based on RFID is proposed. By using sensor related technology, the location information of the iron ladle in the way is collected, and the information of the path, position, direction and speed of the iron ladle in the transportation process is reflected in real time, and the real-time data is provided for the dispatcher.

This paper describes a simple method for seamless integration of a CMOS chip with typical centimeter-scale microfluidics. A simple planar mold is used to cast PDMS around the perimeter of a single die-cut CMOS chip (5 mm x 6 mm) mounted on a standard chip package (DIL-48, 1 cm x 4 cm). The casting results in a contiguous intersection between the exposed chip surface and the co-planar PDMS that extends across the package. A microfluidic channel added to the package extends across the chip boundary, while fluid in the channel is in direct contact with the chip surface. The large platform of this chip package allows fluid connections to be located far from the chip to prevent interference with optical detection, and thin-layer PDMS microfluidics (about 350µm) created by casting are compatible with short-working-distance microscope objectives. This method requires no sophisticated chip post-processing, accommodates individual CMOS chips, and expands the footprint available for microfluidics. Temporary bonding using PDMS as adhesive allows removal of the microfluidics and recovery of the chip and planarized package. This is a simple and reversible method that is well-suited for prototyping devices using large footprint disposable fluidics and small high-value CMOS chips.

Mold level stability is a very important factor in the continuous casting process. To effectively control the liquid level at the set value, solve the problems of disturbance, nonlinearity, and time-varying in the system, as well as modeling difficulties caused by uncertain factors, and aiming at the shortcomings of conventional PID control, the gray wolf algorithm is introduced. The gray wolf optimization algorithm is used to self-tune the PID parameters, and the objective function is constructed based on the adjustment time and steady-state error of the mold liquid level system. Finally, the simulation and experimental results show that the algorithm optimization proposed in this paper is compared with the particle swarm algorithm, and the step response of the system is smoother. Gray Wolf optimized PID control has a fast response speed, with an adjustment time of 3.87s, which is lower than 13.46s for conventional PID control and 5.58s for the PSO algorithm. It can significantly improve the control accuracy and other performance indicators of the system. It provides a reference for improving the level control of the continuous casting mold.

Continuous casting mold (CCM) is molding equipment of the casting blank and widely used in iron and steel metallurgy. This paper focuses on the tracking control of vibration displacement of CCM is driven by a servo motor. To address the non-unique inverse solution of the nonlinear periodic function, a new nonlinear processing algorithm is developed. Based on the algorithm, an active disturbance rejection controller (ADRC) with self-adjusting parameters to estimate and compensate the disturbances is proposed for suppressing the disturbances and improving the tracking performance. The simulation results demonstrate the effectiveness of the proposed control scheme.

The occurrence of convex part on metal surface of the product during casting is unpreventable. The purpose of this research is to propose a robot manipulation algorithm with a feedback control system for automatic removal of convex parts, which consists of a tool tip reaction force and a position feedback control, instead of manual removal as done by craftsmen. Removal experiments are carried out to identify the appropriate value of each of the parameters used in the proposed feedback control system. It is shown that better removal time and high-quality surface can be achieved by applying the proposed feedback control in the removal process of convex parts. Besides, sufficient machining performance obtained from the removal processing experiment of convex part of different shape proves that the proposed method is effective on convex parts of unknown shape. In addition, the generality of this system is proven by applying desired reaction force as reference input to the proposed feedback controller s not only to S50C but also to several metal materials such as SUS303, ANP79 and C2801P.

Aiming at the problems of low operation reliability and maintainability of the original vacuum die-casting workers, various complex parameters control and monitoring requirements, this paper proposes a control method for vacuum pressure regulating casting control system based on PLC. We make a detailed description of the design idea of the control system and the design and implementation of the specific software and hardware. During the control process, the absolute pressure sensor of acquisition layer is used to acquire the pressure value of differential pressure casting equipment, and the pressure data is transmitted to the main control layer through CAN bus. In equipment operation, fuzzy self-tuning PID algorithm is adopted to control the production process. The experimental results show that the intelligent control system of vacuum die casting has high control precision and sensitive response, which shortens the time of intelligent control and exception handling in vacuum casting production.

This work aims to produce MMC composites of sustainable matrix. The Scrap Aluminum Alloy Wheel (SAAW) was used as a matrix and alumina as a reinforcement. The process parameters such as squeeze time and pressure, stirrer velocity, and die preheating temperature were designed for the development of alumina reinforced aluminum matrix composites (AMCs) using the Taguchi method. Mechanical analysis such as hardness, tensile, compression and wear properties were conducted on these stir-casted composites. The findings showed that adding of alumina to aluminum matrix has improved the mechanical and tribological efficiency. The optimized 9 approaches (L1 to L9) were obtained from the Taguchi study, and the L5 and L6 methods showed optimum mechanical properties with process parameters of 100 MPa squeeze strength, 30 to 40 sec squeeze time, 250 to 350 ° C die preheating temperature and 450 to 525 rpm stirrer pace. The Grey Relational Analysis (GRA) of Taguchi configured the method parameters on the basis of these results. It was inferred from the study that properties are process-dependent parameters. There are many possible industrial applications for these recycled MMC’s that include automotive and aviation industries.

Microneedle arrays (MNA) are used to enhance the efficiency of transdermal drug delivery. Within the framework of this work, various techniques of obtaining MNA were investigated: inkjet printing, stereolithographic 3D printing, and injection molding. Experimental MNA samples were obtained using a PiXDRO LP50 inkjet printer. The disadvantage of this technology was the limitation of the maximum height of the resulting structures (about 300 nm). In turn, the method of stereolithographic 3D printing showed insufficient resolution for the manufacture of MNA, so this method was used only for the manufacture of injection half-molds. To cast the MNA prototype, a manual casting machine was created, consisting of a bowl and a Teflon piston. A prototype of a membrane valveless microfluidic pump (MFP) based on an actuator made of an ionic polymer-metal composite was used to supply liquid through a manufactured tin-lead alloy MNA. A model of the transdermal system is assembled from the manufactured MFP and the applicator with the obtained MNA.

Due to the increase of the wheelbase and the change of the steering trapezoid angle of the front axle of a certain type of mining truck, the cross arm needs to be redesigned. To determine the shape and material of the new cross arm, which was analyzed by finite element software, and the structure of the cross arm is further optimized under the condition of satisfying its structural strength. The results show that the weight of the cross-drawing arm can be reduced and the stress and strain can be reduced significantly. The new structure can use the method of plate cutting instead of casting and provide an effective design method for reducing the cost.

The paper addresses the prototypical development and implementation of an innovative and functionally integrated gear stage for next-generation electric vehicles, produced by using Laser Beam Melting (LBM) and Fused Deposition Modeling (FDM). Two new design targets are addressed by our approach: Oil-free operation is achieved by innovative coating of gear teeth which enables almost maintenance-free operation. Furthermore, a higher damping of sound transmission from tooth mesh to bearings is implemented by an optimal wheel body geometry that is realized by a combination of printable carbon-containing polymers (flange) and steel (gear rim). The innovative approach is to extend components by areas that cannot be produced by milling and casting, e.g. complex structures, contour-following fluid-carrying cooling elements by means of LBM, thus creating a significant benefit compared to state-of-the-art gearboxes. Function integration made possible by additive manufacturing results in a very high component complexity. It is combined with gear grinding and final ta-C coatings ensuring maximum efficiency in dry operation, to achieve economically viable production processes. The increase in tooth mesh efficiency is achieved by optimization approaches in the design to minimize heat losses while at the same time maintaining highest possible strength and lowest noise emission.

Continuous casting process is a very important process in iron and steel production, mold level control is a very important link in the process of continuous casting. In order to control the mold liquid level more effectively, the artificial fish swarm PID control algorithm is introduced to adjust the mold liquid level. The incremental PID control algorithm is used to obtain better control effect through weighted processing, and the artificial fish swarm optimization algorithm is used to adjust the incremental PID parameters. The simulation results show that the artificial fish swarm algorithm is more effective in the mold level control.

In allusion to the complex nonlinear vibration characteristics of the casting flash cutting machine, this paper was to analyze the vibration mode of the whole machine in the working process with the finite element software—ANSYS Workbench. It was found that the machine tool was severely deformed at the first, second and fifth frequencies. The excitation frequency of the machine tool had a certain difference from the 6th-order natural frequency value obtained from the modal analysis. It provided an effective theoretical basis for the vibration reduction control of the whole machine. At the same time, the deficiencies of the machine tool in the testing process were improved, and the machining accuracy, running stability and production efficiency of the casting flash cutting machine were achieved.

In the process of squeeze casting, the clearance between the shot sleeve and the punch is constantly changing, which results in the friction between the punch and the shot sleeve. The friction has a great influence on the quality of the casting and the service life of the equipment, so this study aims to find out the mechanism of friction. A friction coupling numerical analysis model in squeeze casting was established and verified by the actual experiment. Simulation and experiment results indicated that the friction coupling model is effective, which reflects the actual squeeze casting process more accurately.

The casting material of modified hub in a certain assembling unit was made from 20SiMn Furthermore, it used the puddle welding technology to generate 390L cladding as the surface of hub to improve its anti-cavitation ability. Nevertheless, quality issues were still spotted on the overlay of hubs in ex-factory acceptance. On the one hand, the detecting defects existed in the penetration test; on the other hand, the hardness test was ineligible. After turning off the defective surface, the inner puddle welding layer remained similar defects. Therefore, the surface overlay of hubs was completely turned off and then processed by puddle welding method so as to ensure the overall quality of the hubs.

This article is devoted to the problem of reducing the number of defects and improving the quality of manufacturing gas turbine engine blades. The process of pressing a casting rod designed to form the inner cavity of the blades during their manufacture by pouring on the smelted models is considered. Information about 400 examples of pressing foundry rods has been collected. Each of the examples contained a set of parameters that characterize the process of obtaining the workpieces and the result of manufacturing. Based on the collected statistical information, the neural network was designed. Using virtual computer experiments, the process of pressing casting rods was studied. The parameters that have the greatest impact on the quality of the resulting products are identified. In the course of computer experiments, it was observed that changes in a number of pressing parameters which lead to a decrease in the probability of defects on one billet do not lead to a similar decrease in the probability of defects on another billet. The method of neural network modeling was able to identify parameters the same change in which leads to a decrease in the probability of any type of defect of all workpieces. For example, the probability of getting defects in all workpieces decreases with increasing the holding time of the rod in the mold without pressure. Thus, it is shown that the developed neural network model allows to control the quality of the obtained blades, select the optimal parameters of the technological process that provide the maximum reduction of defects.

There has been a direct relationship between the temperature of the laser point and the quality of the casting in the process of the 3D printing. In the paper, a method based on convolutional neural network (CNN) was proposed to estimate the temperature of the laser point. The collected temperature data used were trained by the deep-learning method. A new structure of the model was proposed on the part of the CNN model, which improved from original LeNet. The process of the prediction for the testing set was carried out through the new model. The unknown temperature in the testing set can be estimated. The experimental result illustrates that the proposed method is satisfactory.

Interactive techniques in Virtual Reality (VR) is rapidly developing in the recent years. However, the consistency of the interaction activities in VR and real-world remains a subject to be further studied. In this work, we study the consistency of casting action between VR and real-world. In our experiments, we let participants cast a ball to a walking human virtual agent, which is driven by a motion capture system in the virtual environment. By comparing to the casting accuracy in real-world, we found that the participants' performance under various of the experimental configurations is consistently reduced in VR. We also found that the moving ‘speed’ factor of the casting target barely affected the accuracy in VR, which was different in the real-world environment. This study will provide an important experimental reference to develop future VR systems based on casting actions.

Soft gripper has excellent flexible grasping ability due to the soft and transformable body. Inspired by the drive principle of ring and longitudinal muscle in jellyfish, a enveloped soft gripper robot is proposed, which is combined of four well-designed bionic soft actuators. The FEA results presents that the bionic soft actuator can bend similar as the ring and longitudinal muscle in jellyfish. Then the molds of casting forming are designed and fabricated by 3D-printing. The casting forming process is presented in detail. Finally, the bionic soft gripper is integrated on the KUKA robot to grasp target objects. The experiment results prove that the proposed bionic soft gripper can stably grasp smooth-faced seafood and eggs.

The accuracy of molten metal level measurement is the base of controlling the molten metal level, which greatly affects the production quality. Due to the oscillation of the caster mold, hot temperature, and noises in the continuous casting sys-tem, the accurate level is hard to obtain. Currently used molten metal level measurement technique includes radioisotope trans-mission, electromagnetic, ultrasound reflection, microwave re-flection, computer vision, and capacitive level measurement method. The radioisotope transmission method has potentially hazardous to exposed workers, while the electromagnetic method needs to maintain the coil frequently for the large excitation source needed, the ultrasound reflection method, micro-wave reflection method and computer vision method can be affected by the casting powder used for preventing molten metal splash. In this work, a new structure of the capacitance sensor designed for minimizing the influence of caster mold’s oscillation and improving heat dissipation was proposed, the variation of capacitance between the sensor electrode and the caster mold with molten metal level change and the oscillation of the caster mold was calculated in simulation and validate through a grounded simulation tank with capacitance to digital converter FDC1004. The result shows that the proposed capacitance sen-sor can achieve an accuracy of 1 mm when the distance between the molten metal surface to the sensor within 80 mm with 2 sam-ples per second report speed.

Clubfoot scientifically known as CTEV (Congenital Talipes Equino Varus), is a congenital disease condition where the child presents with either one foot or both the feet rotated internally at the ankle. The treatment is the ponseti method, including a casting phase and bracing phase. The casting phase is the initial treatment phase and is a crude method by itself followed by a tenotomy and the successive step is the bracing phase. The major drawbacks of the casting phase were that there was no specific gradual angle change and secondly the removal of the plaster using water leading to the baby falling ill. The underlying purpose is to replace the casting with an easier and more technical technique using a brace. To render an alternative foot brace in order to replace the series of casts and at the same time ensures a gradual change in the foot orientation angle. The principle behind the design of a brace is 3-point pressure principle and so that the deformity would be corrected by gradually changing the foot angle. Fabrication is a welcome move which involves 3D printing technology. The brace ensures the gradual change in the foot orientation angle and nullifies the relapse of the foot. Simulation is performed using Autodesk Fusion 360 to understand the contact forces and resultants. The various types of stress forces and principal forces are worked with to understand the maximum and minimum force experiencing regions. Materials taken for the study is Thermoplastic Polyurethane and Carbon fiber. Further, the material properties were studied and material testing is also involved.

Along with the Industry 4.0, Smart Factory is receiving great attention worldwide. In particular, quality control is the most important element of the production system. Of the many processes in manufacturing, the early process, casting is the biggest role in modern root industry. Casting is a manufacturing process in which a liquid material is usually poured into a mold to harden for solidify. The quality inspection process of casting products is largely divided into four stages. The last inspection phase, surface inspection, is inspected directly by the person, or by a vision system. The main key to quality inspection is accuracy and speed. Quality inspection using the vision system improves the competitiveness of future industries through fast and accurate inspection. Deep learning techniques have been widely used and studied for quality inspection problems. This paper analyzes the image of the cast product extracted through the vision sensor and proposes a Convolutional Neural Network (CNN) for quality inspection of the cast product and a Convolutional Autoencoder (CAE) to improve the learning quality of machine learning using a small amount of data.

Under heavy load operating conditions, the fail rate of GIL spacers will be greatly increased. Flashover fault is the most frequently occurring failures in GIL, which seriously affects the safe operation of HVDC transmission systems. This paper introduced the problem of local electric field distortion of GIL spacers under working conditions, proposes a method to uniform surface electric field by constructing a functional gradient layer on the surface of spacers. Then this paper introduced the fabrication method of FGM insulation materials, including centrifugal method, 3D printing and flexible mixed casting. Some examples are given to introduce the current research status of the construction method of the surface functional layer of the novel spacers, including multilayer impregnation, magnetron sputtering and surface fluorination. Finally, the problems that need to be solved urgently and the direction of future research and development are pointed out.

a computerized control system for applying pressure to liquid and crystallizing metal based on industrial controllers was developed. A dedicated instrumentation system to monitor the key process variables is also developed. Appropriate PLCs to control the process were selected. A sequence that compares the current process variables with the target values and makes adjustments as required was also developed. Obtained results apply to the design of computerized control systems for casting with crystallization under pressure processes.

Low Pressure Die Casting (LPDC) is widely used in crankcase production. Owing to the unstable production situation and the complex shape of crankcase, porosity defect usually occurs in LPDC process which leads to loss in quality and productivity. In this paper, we apply data mining methods to predict the porosity defect in advance, so that we can take actions to prevent it from reoccurring in the next production part, thereby increasing the quality and productivity. To do it, firstly, we collect production data from a real-world casting line. Secondly, we prepare the data for prediction by feature extraction and feature selection. Thirdly, we apply an ensemble algorithm named Forest of Local Trees (FLT) for defect prediction. Finally, we present a thorough experimental study of the proposed method. The results show that our method outperforms other five algorithms on five real-world datasets in terms of three indicators, recall, precision and F-measure.

The quality of aluminum alloy ingots is very important to the performance of products. But the casting process is complicated. Chemical composition and other factors also have a great impact on the quality of the ingots. In this paper, based on the actual production of an aluminum casting plant in China, we propose a data-driven ingot quality prediction method using XGBoost. This method includes four parts: data collection, database construction, data cleaning and relation model establishment. Among them, data collection and database construction are based on process knowledge. Make sure the data is realistic. At the stage of data cleaning, the features which have great influence on the quality of ingot casting but are neglected in production are found. In dealing with the missing value, the fitting method is adopted to solve the serious problem of missing industrial data. The ensemble learning method is used to build the relational model. It can completely reflect the aluminum alloy casting process and make accurate prediction of ingot quality quantitatively. As can be seen from a large number of comparative experiment, the XGBoost model works very well and gives the key factors that affect the quality of ingots.

Casting defects detection is necessary for manufacturing industry. In this paper, we propose a casting defects detection method based on the Deep Learning algorithm. We aims at solving problems of insufficient computing resources in most manufacturing enterprises, and we introduce the Efficientnet network and depth separable convolution to saving computing resources while maintaining detection accuracy. The experiment results verifies our method could perform well even if in the case of insufficient computing resources. It provides a automated casting defects detection method and give a proposal to the industrial defect detection researchers who faced the problem about computing resources.

Specific to the field of iron and steel metallurgy, this study analyzes the problem of increased steel loss caused by the intermediate billet head and end cutting in the hot continuous rolling process, and proposes to optimize the shape of the continuous casting billet through the shape compensation method, thereby to reduce the loss of head and end cutting caused by the intermediate billet, and determine the cutting data of the intermediate billet under the representative process conditions through the statistical analysis of the field data. Based on this, the simulation method is adopted to carry out simulation calculation of the hot continuous rolling process, attempting to determine a reasonable optimization cutting plan of the continuous casting billet. The study shows that, using the CAD modeling and Deform software simulation method to simulate the rolling process of the continuous casting billet can effectively simulate the shape change of the casting billet in the actual production process.

The crankshaft of a marine engine breaks after less than 2 years operation. Failure analysis was carried out by means of mechanical property test, chemical composition analysis, metallographic analysis and fracture observation. Analysis shows that the reason for the fracture of crankshaft is that the inclusion of large-scale protective slag in the crankshaft caused by the inclusion of protective slag in the process of casting, which destroys the continuity of crankshaft material and reduces the bearing capacity of crankshaft and the crankshaft produces fatigue crack when working. The heat treatment and nitriding of the crankshaft did not meet the requirements of the drawings, which further reduced the fatigue strength of the crankshaft and accelerated the fatigue crack growth. With the crack growth, the fracture finally occurred.

Many different applications of product inspections have found a significant advantage by the use of 3D-scanners, especially when working with complex surfaces, where traditional inspection tools have significant limitations. This work deals with the structured light 3D-scanning based on fringe-pattern and obtaining polygonal mesh models (stl-models) of several end- and half-products, manufactured by different operations of blank production. In the study, the preparing of point clouds are discussed as well. 3D-scanning is possible to apply for getting the high resolution and accurate results for both metallic and non-metallic materials. Careful 3D-model reconstruction provides the size tolerances up to ±0.05 mm for the rough surfaces after stamping and casting operations and up to ±0.60 mm for wavy surfaces after 3D-printing by FFF-method. The automatic repairing and healing of the 3D-models can cause artefact building and trespassing the accepted thresholds for tolerances.

The combination of reverse engineering and 3D printing technology makes 3D printing technology more and more widely used in the manufacturing field. This article starts with the collection of point cloud data in reverse engineering, completes reverse modeling, 3D printing, cast molding simulation and casting molding and compares actual prototypes with experiments. The application of reverse engineering and 3D printing technology in product innovation design is promoted.

Feeding is an important and also a heavy work in river crab farming, uniform feeding is necessary for raising river crab on the whole pond. The overall scheme of automatic bait casting system carried by paddlewheel catamaran is proposed based on RTK-GPS navigation positioning device. The feeding device equipped with the controllable flow velocity of bait feeder, distance adjustable thrower and bait weight measurable device which could perform variable feeding is designed to solve the problem of nonuniform distribution of the feed. The hardware and software system of automatic feeding device are designed. Then the calculation and simulation for the modeling of the feeding device are pefromed. The test results show that the system linear track positioning accuracy is less than 30 cm, and automatic continuous feeding speed is up to 240 m2/min. The results of this study provide technology reference for further study of automatic uniform feeding on the whole pond in river crab aquaculture.

Aiming at the problem of missing data on the safety status of the mixing furnace in the aluminum foundry, a safety status evaluation model for the mixing furnace based on DS evidence theory was established. By analyzing the status data of the operation process of the mixing furnace, the discretized data was fitted to extract the distribution characteristics. And other information evidence and the index similarity is used to determine the index weight. Finally, the safety status of 1 # mixing furnace in the casting aluminum workshop of Taiyuan Dongalumi Aluminum Co., Ltd. in Shanxi Province was taken as the research object, and the validity of the model was verified.

The Surface-Volume-Surface Electric Field Integral Equation (SVS-EFIE) has been recently generalized to solution of general scattering and radiation problems for 3D composite objects. These objects can be formed by multiple piece-wise homogeneous dielectric regions which do or do not share common boundaries. Generalization to the composite objects formed by metal and piece-wise homogeneous dielectric regions which share common boundaries has also been demonstrated. Since the SVS-EFIE formulation utilizes only the electric field dyadic Green’ s functions, it can be extended to the case of composite objects situated in non-magnetic planar multilayered media by casting its operators into the mixed-potential form using classical Michalski-Zheng’ s approach. Examples of the above SVS-EFIE formulations are summarized in the paper.

Induction melting furnace is the heating chamber induct electricity to melt metal. The Induction melting furnace is widely used for modern foundries, which manufacture their products by casting or ironmaking. In the operating process, heat from melted metal makes a crack on the furnace wall. After the period, maintenance must be proceeded to prevent an explosion caused by the cracked wall. In this paper, we proposed a method to predict the maintenance point of Induction melting furnace based on the neural network. The methodology is configured by three-steps. The first step is detecting the manufacturing cycle from furnace data. The second step is capturing the value from each cycle. The third step is to predict the maintenance point by using a neural network model. By this method, we are able to predict the maintenance point with high accuracy

Frequent occurrence of Orographic precipitation in hills during most of the time of a year causes severe landslide which has adverse effect on the habitat around it. Therefore now casting of rainfall in terms of rain rate sufficient time ahead is very urgent need to save society from severe disaster. Therefore using statistical method like Auto regressive integrated moving average model (ARIMA) or using artificial neural network (ANN), now casting of rain rate had been done in the literature and the two methods are compared in this work in terms of error performance. The retrieval error indicates the performance of the ANN model is best, however both the method shows acceptable prediction.

This paper's main objective is to study the flexural behavior of concrete beams reinforced with basalt fiber-reinforced polymer (BFRP) bars that were exposed to harsh environmental conditions before casting. The bars were placed in a tank outdoors for a year where they were exposed to direct heat from the sun, humidity and occasional rain. A four-point test was conducted on four concrete beams reinforced longitudinally with unexposed and exposed BFRP bars, respectively. A 30 MPa concrete compressive strength was used to cast all beams, which were designed to be over-reinforced, and thus were expected to fail by concrete crushing. The effects of exposure and reinforcement ratio were studied by investigating the cracking moments, load-deflection curves and failure modes. Results indicated that exposure had no significant effect on ductility and did not affect the failure mode. Also, increasing the reinforcement ratio increased the capacity of the beams as per the theoretical prediction by the ACI440-1R.15.

From naval to aerospace and industry to household, everywhere the necessity of aluminum is felt. But oxide film can be formed on this metal and its alloys during casting or welding due to high affinity of aluminum for oxygen. This phenomenon can further lead to porosity which is a vital defect. Visual inspection is not sufficient due to different human constraints. So, many image processing techniques have been developed. For better welding, we propose here two useful and efficient techniques. Unlike most of the previous techniques, our proposed techniques are color based processing. In this manuscript, our main goal is to detect oxide film and we show that the task is successfully done by our methods.

We studied the physical and technological features of packaging high integrated micro-assemblies, as well as multi-chip modules. Technological operations of packaging process test chips at the plate level by casting compound were carried out and the optimal conditions for their implementation were established. Sealing of chips at the wafer level was carried out using filling compounds Silterm-1B and Silterm-1K. As a result, using Silterm-1B compound with additives up to 10% of the total mass of finely dispersed silicon particles with a size not exceeding 5 μm, it was possible to reduce the thermal expansion coefficient of the entire multilayer structure to 50 * 10-6 1 / K, which subsequently reduced the deformation of the structure at high-temperature processes, and also increased the thermal conductivity of the structure.

This paper reports a silicon-chip based electromagnetic vibration energy harvester that is fabricated with a wafer level micro casting technique to form on-chip integrated 3D metal solenoid and magnet sliding channel. The 150-turn solenoid is fabricated by rapidly casting molten ZnA1 alloy (at its low melting point of 380°C) into pre-micromachined silicon mold, where all the coils (wire cross sectional area= $200\mu \mathrm{m}\times 40\mu \mathrm{m}$ ) in the wafer are formed within 7 minutes. A sliding magnet responding to environmental vibration is put into the channel inside the solenoid to generate electricity. With the wafer-level energy harvester fabrication technique, the size of the harvester chip is as tiny as $10.5\times 2.5\times 1.7\text{mm}$ . Test shows that the harvester collects environmental energy high-efficiently within frequency range of 17-47Hz and generates peak power of $16.1\mu\mathrm{W}$ at 37Hz under 6.5g vibration. Compared to either non-MEMS wire-wound electromagnetic harvesters or normal on-chip piezoelectric/electrostatic ones, present on-chip EM harvester shows great improvement in volume power density.