Process optimisation of transfer moulding for electronic packages using artificial neural networks and multiobjective optimisation techniques

TensorFlow**Other names and brands may be claimed as the property of others. is a popular deep learning framework used by data scientists to solve a wide-range of machine learning and deep learning problems such as image classification and speech recognition. It also operates at a large scale and in heterogeneous environments - it allows users to train neural network models or deploy them for inference using GPUs, CPUs, (Intel <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">®</sup> Xeon <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">®†</sup> Intel, the Intel logo and Xeon ® are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries. © Intel Corporation. CPUs) and deep learning specific custom-designed hardware such as TPUs. Even though TensorFlow supports a variety of optimized backends, realizing the best performance using a backend may require additional efforts. For instance, getting the best performance from a CPU backend requires careful tuning of its threading model. Unfortunately, the best tuning approach used today is manual, tedious, time-consuming, and, more importantly, may not guarantee the best performance. In this paper, we develop an automatic approach, called TensorTuner, to search for optimal parameter settings of TensorFlow's threading model for CPU backends. We evaluate TensorTuner on both Eigen and Intel's MKL CPU backends using a set of neural networks from TensorFlow's benchmarking suite. Our evaluation results demonstrate that the parameter settings found by TensorTuner produce 2% to 123% performance improvement for the Eigen CPU backend and 1.5% to 28% performance improvement for the MKL CPU backend over the performance obtained using their best-known parameter settings. This highlights the fact that the default parameter settings in Eigen CPU backend are not the ideal settings; and even for a carefully hand-tuned MKL backend, the settings are sub-optimal. Our evaluations also revealed that TensorTuner is efficient at finding the optimal settings - it is able to converge to the optimal settings quickly by pruning more than 90% of the parameter search space.

Genetic algorithms (GAs) are general-purpose heuristic search algorithms that mimic the evolutionary process in order to find the most fitting solutions. The algorithms were introduced by Holland in 1975. Since then, they have received growing interest due to their ability to discover good solutions quickly for complex searching and optimization problems. The traditional GAs have been converted to multi-objective GAs to solve multi-objective optimization problems successfully. However, GAs require parameter tunings (such as population size, mutation probabilities, crossover probabilities, selection rates) in order to achieve the desirable solutions. The task of tuning GA parameters has been proven to be far from trivial due to the complex interactions among the parameters. It takes trial and error experiments to obtain the optimal GA parameter settings for an arbitrary real-world problem, Many researchers have been trying to understand the interdependencies of GA parameters in order to determine their optimal settings. The objective of this research is to develop the elitist non-dominated sorting GA (NSGA-II) for multi-objective optimization as a parameter-less multiobjective GA. The research then evaluates and discusses the performance of the parameter-less NSGA-II against the original NSGA-II with optimal parameter settings using the experimental result for a test problem borrowed from the literature.

A Computer-aided Mold Design for Transfer Molding Process in Semiconductor Packaging Industry

In Injection Molding process, energy consumption for the product will be more and optimal setting up of parameters is very important for reducing the energy consumption and also in controlling the quality of the injection molded product. It is much difficult to set optimal process parameter levels which may cause defects such as shrinkage, line defects etc. Determining optimal process parameter setting critically influences the productivity, quality and production cost in Plastic Injection Molding industry. The main purpose of this paper is to minimize the energy consumption and it is done using Minitab. The approach is based on Taguchi's Method and Analysis of Variance (ANOVA) to analyze and optimize the process parameters.

The process of sheet metal forming is characterized by various process parameters. Accurate prediction of springback is essential for the design of tools used in sheet metal forming operations. In this paper, an evolutionary algorithm is presented that is capable of handling single/multiobjective, unconstrained and constrained formulations of optimal process design problems. To illustrate the use of the algorithm, a relatively simple springback minimization problem (hemispherical cup-drawing) is solved in this paper, and complete formulations of the algorithm are provided to deal with the constraints and multiple objectives. The algorithm is capable of generating multiple optimal solutions in a single run. The evolutionary algorithm is combined with the finite element method for springback computation, in order to arrive at the set of optimal process parameters. To reduce the computational time required by the evolutionary algorithm due to actual springback computations via the finite element method, a neural network model is developed and integrated within the evolutionary algorithm as an approximator. The results clearly show the viability of the use of the evolutionary algorithm and the use of approximators to derive optimal process parameters for metal forming operations.

In this paper, a system for LTE Cell Outage Compensation (COC) based on hybrid Genetic Algorithms (GA) and Artificial Neural Networks (ANN) has been proposed. COC aims to minimize the impact of cell outage which leads to decrease in operator revenue and/or the customer satisfaction. The proposed system adopts an optimization module to search for an optimal setting of a set of LTE operational parameters to achieve a targeted set of key performance indicators. The optimization process always leads to good enough solutions, but it also requires a huge number of trials. So, in the proposed system, a huge set of outage scenarios is collected along with their optimal argument settings that are acquired by the optimization module and they are used to train an artificial neural network (ANN) module, which acts as an expert that can optimally act on the different situations in real-time mode. Simulation environment is set to evaluate different LTE measures and Key Performance Indicators (KPIs) on different outage scenarios. Simulation results proved the capability and robustness of the proposed system to minimize the number of users experiencing outage. Simulation results also show that the proposed system achieves optimal parameter settings without violating the overall system performance and with minimal processing time, while introducing significant impact on the performance of LTE.

In this paper, three dimensional finite element analysis has been carried out using FEM-based DEFORMTM 3D software on hot forging of connecting rod. The influence of design parameters and process parameters are investigated for the responses like effective strain rate and forging load during forging operation. In order to optimise both the responses simultaneously, grey relational analysis (GRA) embedded with Taguchi method is employed. Grey relational grade is used as a performance index to determine optimal setting of process parameters for both the responses simultaneously. Analysis of variance (ANOVA) is employed to determine significant parameters. Process parameters such as flash thickness, flash width, and corner radii are found to be the most significant parameters affecting grey relational grade. Optimal process setting leading to a higher effective strain rate with the least forging load has been verified through a confirmation experiment for validation of the results. [Received 15 October 2015; Revised 25 February 2016; Accepted 13 March 2016]

Experimental investigation and optimization of cobalt bonded tungsten carbide composite by hybrid AHP-TOPSIS approach

In this paper, the combined use of decision trees and artificial neural networks is examined in the area of quality improvement of industrial processes. The main goal is to achieve a better understanding of different settings of process parameters and to be able to predict more accurately the effect of different parameters on the final product quality. This paper also presents results from the application of the combined decision tree - neural network method to the transformer manufacturing industry. In the environment considered, quality improvement is achieved by increasing the classification success rate of transformer iron losses. The results from the application of the proposed method on a transformer industry demonstrate the feasibility and practicality of this approach for the quality improvement of industrial processes.

Molded interconnect device (MID) technology is a key enabling technology with growing markets in automotive, communications, consumer electronics through integration with lighting and sensor technologies. The MID technology is yet to be explored for high temperature applications in automotive or consumer lighting. One of the hindering factors for such implementation in the serial production and time to market is the improper electronic and thermal packaging of the light emitting diodes (LEDs) on the MID substrates. This paper addresses the optimization of mold design, surface metallization and soldering process for the effective thermal management of the high-power LED systems. By using a simulation model, the thermal distribution and the resultant decrease in temperatures for varying forward electrical currents in high-power LEDs by design optimization is demonstrated. In addition, the optimization of solder process with respect to solder profile in context of vacuum vapor-phase soldering for void-free solder connections is discussed.

Experiments, analysis and parametric optimization of cylindrical traverse cut grinding of aluminium bronze

In this article, a boron carbide particle (B4C) reinforced AA2014 surface composite was first fabricated by friction-stir processing (FSP) to investigate the impact of the volume percentage of B4C, tool rotational speed and table speed on the tensile strength (TS) and wear rate (WR). The AA2014 composite is one of the important candidates for making defense and aerospace components due to its high strength and minimum weight. Taguchi orthogonal array was employed to design and predict the maximum tensile strength and minimum wear rate. The volume percentage of B4C is the most momentous parameter for both the tensile strength and wear rate. The optimum parameter settings for attaining the maximum tensile strength of 605 MPa and a minimum wear rate of 1.2 mm3/Nm are a B4C volume of 15 %, tool rotational speed of 900 min–1 and table speed of 60 mm/min. The optimum process-parameter settings were used to make a specimen for validating the estimated results. The microstructure and chemical composition of the surface composite of the optimum specimen were illustrated using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS), respectively. The surface profile and microscopic view of the worn-out surface composite were also examined using SEM images.

The thermal reflow process is widely used in microlens array fabrication. However, the resulting arrays are commonly criticized for their low fill factor. In this work, caulking is applied to fill the gaps between adjacent lenses. The experimental results prove that a gapless microlens array with a 100% fill factor could be successfully produced and the caulking time precisely controlled. Furthermore, an artificial neural network and genetic algorithm are used to achieve high quality using the thermal reflow and caulking. The L18 orthogonal array is used as the learning data for the artificial neural network to construct a system model that could predict the results (e.g., S/N, focal length, and roughness) for arbitrary parameter settings. The genetic algorithm is then applied to obtain the optimal parameter settings. The major objectives in using the optimal design are to reduce the variation in the focal length and the surface roughness for a microlens array. This allows improved focus and enhanced illumination brightness. The results show that microlens array quality could be significantly improved in comparison with the original design.

Design optimum frac jobs using virtual intelligence techniques

The additive manufacturing processes are increasing their importance in fabrication of composite like structures. Friction stir processing is classified as a solid state additive manufacturing technique that is used for lamination or powder deposition in a matrix substrate. In the present work an attempt was made to add the silica as secondary phase into the acrylonitrile butadiene styrene (ABS) surface for enhancing the mechanical and surface properties of fabricated composite. The parameters such as amount of silica fraction, pass number, tool rotational speed and linear speed were considered to be involved in the study. The responses surface method (RSM) was used here to assess relationship between aforementioned inputs and outputs vs bending strength and deflection before fracture. Optimal setting of process parameters was determined by maximizing bending strength and deflection. The obtained optimal results showed that maximum strength and deflection can be achieved by selecting of 100% silica volume fraction, 1 pass number, 1400 rpm tool rotation and 80 mm/min travel speed. The optimum results were further verified through confirmatory experiment and the result showed that the values of prediction error for bending strength and deflection are 6.1% and 8.5%, respectively. This proves the validity of the proposed approach for empirical modeling and optimization of friction stir additive manufacturing process.