Article

Optimization of injection molding process parameters using combination of artificial neural network and genetic algorithm method

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Injection molding is the most widely used process in manufacturing plastic products. Since the quality of injection molded plastic parts are mostly influenced by process conditions, how to determine the optimum process conditions becomes the key to improving the part quality. In this paper, a combining artificial neural network and genetic algorithm (ANN/GA) method is proposed to optimize the injection molding process. In this method, a BP neural network model is developed to map the complex non-linear relationship between process conditions and quality indexes of the injection molded parts, and a GA is used in the process conditions optimization with the fitness function based on an ANN model. The combining ANN/GA method is used in the process optimization for an industrial part in order to improve the quality index of the volumetric shrinkage variation in the part. The results show that the combining ANN/GA method is an effective tool for the process optimization of injection molding.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... Artificial neural network (ANN) [12,13], genetic algorithm (GA) [14,15], regression methods [16,17], Taguchi experimental design method [18], and fuzzy [19] are the most preferred predictive and optimization methods found in the literature. ANN is an advantageous method for predicting linear and non-linear systems that has been widely used for modeling and prediction purposes in many fields [17]. ...
... H + a15 .H 2 + a i16 .A.B + a i7 .A.C ...
... Polymers 2023,15, 3915 ...
Article
Full-text available
The use of recycled polypropylene in industry to reduce environmental impact is increasing. Design for manufacturing and process simulation is a key stage in the development of plastic parts. Traditionally, a trial-and-error methodology is followed to eliminate uncertainties regarding geometry and process. A new proposal is presented, combining simulation with the design of experiments and creating prediction models for seven different process and part quality output features. These models are used to optimize the design without developing additional time-consuming simulations. The study aims to compare the precision and correlation of these models. The methods used are linear regression and artificial neural network (ANN) fitting. A wide range of eight injection parameters and geometry variations are used as inputs. The predictability of nonlinear behavior and compensatory effects due to the complex relationships between this wide set of parameter combinations is analyzed further in the state of the art. Results show that only Back Propagation Neural Networks (BPNN) are suitable for correlating all quality features in a single formula. The use of prediction models accelerates the optimization of part design, applying multiple criteria to support decision-making. The methodology is applied to the design of a plastic support for induction hobs. Furthermore, this methodology has demonstrated that a weight reduction of 27% is feasible. However, it is necessary to combine process parameters that differ from the standard ones with a non-uniform thickness distribution so that the remaining injection parameters, material properties, and dimensions fall within tolerances.
... The network designer initially determined the weights assigned to these connections among the three layers, but they are subsequently modified for every "epoch" that the network goes through. Shen et al. [22] optimized the injection molding parameter using ANN and genetic algorithm methods. The results reveal that the ANN methodology could effectively model the complex interaction between process conditions and quality index for injection molded parts. ...
... that the network goes through. Shen et al. [22] optimized the injection molding parameter using ANN and genetic algorithm methods. The results reveal that the ANN methodology could effectively model the complex interaction between process conditions and quality index for injection molded parts. ...
Article
Full-text available
Weld line defects, commonly occurring during the plastic product manufacturing process, are caused by the merging of two opposing streams of molten plastic. The presence of weld lines harms the product’s aesthetic appeal and durability. This study uses artificial neural networks to forecast the ultimate tensile strength of a PA6 composite incorporating 30% glass fibers (GFs). Data were collected from tensile strength tests and the technical parameters of injection molding. The packing pressure factor is the one that significantly affects the tensile strength value. The melt temperature has a significant impact on the product’s strength as well. In contrast, the filling time factor has less impact than other factors. According to the scanning electron microscope result, the smooth fracture surface indicates the weld line area’s high brittleness. Fiber bridging across the weld line area is evident in numerous fractured GF pieces on the fracture surface, which enhances this area. Tensile strength values vary based on the injection parameters, from 65.51 MPa to 73.19 MPa. In addition, the experimental data comprise the outcomes of the artificial neural networks (ANNs), with the maximum relative variation being only 4.63%. The results could improve the PA6 reinforced with 30% GF injection molding procedure with weld lines. In further research, mold temperature improvement should be considered an exemplary method for enhancing the weld line strength.
... Further, ML techniques can predict the quality of injection-molded products by analyzing the data generated throughout the production process. Previous research in injection molding has often involved developing single ML models for quality prediction or process optimization [12][13][14][15][16][17][18][19][20]. However, the current trend is toward utilizing ensemble model and, combining multiple models to enhance overall prediction accuracy [21][22][23][24]. ...
... The study using the single model related to injection molding is thus summarized here. Changyu et al. [12] focused on modeling the complex relationship between process conditions and quality indices of injection-molded parts using the ANN method. Specifically, the researchers aimed to improve the quality index associated with volume shrinkage variation and demonstrate the effectiveness of their approach for optimizing the injection molding process. ...
Article
Full-text available
The growing move toward smart factories can leverage industrial big data to enhance productivity. In particular, research is being conducted on injection molding and utilizing machine learning techniques to analyze molding process data, discover optimal molding conditions, and predict and improve product quality. This study aims to identify the key factors influencing the weight defects of injection-molded products and demonstrate the potential use of the double ensemble technique for better prediction accuracy of weight defects. We obtain the key factors influencing weight defects prediction, barrel H2 temp real, metering time, and fill time using gain ratio analysis. Subsequently, we develop single models using machine learning algorithms, including decision tree, random forest, logistic regression, the Bayesian network, and the artificial neural network. Ensemble models, including bagging and boosting and double ensemble models are developed to compare their performance with that of single models. The findings indicate that ensemble models outperform the prediction accuracy of the single models. The double ensemble technique demonstrates the greatest improvements in prediction accuracy over the single models. These results showcase the potential of applying the double ensemble technique to other injection molding areas and suggest that adopting this technique will contribute to establishing other smart factories that will enhance both productivity and cost competitiveness.
... Approaches such as response surface methodology [6,7], design of experiments [8], Taguchi method [9][10][11][12][13][14], surrogate model [15], genetic algorithms [6,7], and artificial neural network model [16][17][18][19][20][21][22][23][24] have been combined with numerical simulation and/or experimental analysis to reduce the time for optimizing the injection molding process. ...
... The artificial neural network (ANN) has been extensively used to predict the relationship between process parameters and the quality of molded parts through shrinkage [17], warpage [16,[23][24][25], mechanical properties [19], part thickness [21], and weight [22]. ...
Article
Full-text available
In this study, computer-aided engineering (CAE) simulation software and the design of experiments (DOE) method were used to simulate the injection molding process in terms of the melt flow length, using a spiral part. Process parameters such as melt temperature, mold temperature, injection pressure and mold cavity thickness were considered as injection molding variables. A predictive model for the flow length was created using a three-layer artificial neural network (ANN). The ANN model was trained with both simulation and experimental data, and the predictive performances were compared in terms of correlation coefficient, root mean square error and mean relative error. The cavity thickness and melt temperature were found to be the most significant factors for both the simulation and the experiment, while the injection pressure and the mold temperature had little effect on the flow length. The ANN model trained with Moldex3D data shows a significantly higher prediction capacity than the ANN model trained with experimental data. However, the melt flow lengths predicted by the ANN model for both Moldex3D and Moldflow simulation data are statistically significant, indicating that the proposed prediction methodology, which combines the ANN model, DOE method and the CAE simulation technology, can effectively predict the flow length of injection molded parts, with a small number of data.
... The algorithm effectively controls and manages the solution set to find the optimal strategy through a process reminiscent of "survival of the fittest" [60,15]. In general, the processing of the genetic algorithm requires the determination of six basic issues: 1. Representative chromosomes; 2. Function selection; 3. Genetic operators; 4. Initial population generation; 5. Terminal conditions; 6. Evaluation function [93,80]. ...
... On this basis, methods combining approximation techniques such as radial basis function (RBF) networks [28], NN models [29,30], and Kriging models [31,32] with optimization techniques, for example, GA and sequential quadratic programming (SQP), have been widely employed in the literature for molding process parameters optimization. These innovative approaches strengthened multiple computational methods to increase the efficiency and precision of the optimization process. ...
Article
Full-text available
An intelligent optimization technique has been presented to enhance the multiple structural performance of PA6-20CF carbon fiber-reinforced polymer (CFRP) plastic injection molding (PIM) products. This approach integrates a deep neural network (DNN), Non-dominated Sorting Genetic Algorithm II (NSGA-II), and Monte Carlo simulation (MCS), collectively referred to as the DNN-GA-MCS strategy. The main objective is to ascertain complex process parameters while elucidating the intrinsic relationships between processing methods and material properties. To realize this, a numerical study on the PIM structural performance of an automotive front engine hood panel was conducted, considering fiber orientation tensor (FOT), warpage, and equivalent plastic strain (PEEQ). The mold temperature, melt temperature, packing pressure, packing time, injection time, cooling temperature, and cooling time were employed as design variables. Subsequently, multiple objective optimizations of the molding process parameters were employed by GA. The utilization of Z-score normalization metrics provided a robust framework for evaluating the comprehensive objective function. The numerical target response in PIM is extremely intricate, but the stability offered by the DNN-GA-MCS strategy ensures precision for accurate results. The enhancement effect of global and local multi-objectives on the molded polymer–metal hybrid (PMH) front hood panel was verified, and the numerical results showed that this strategy can quickly and accurately select the optimal process parameter settings. Compared with the training set mean value, the objectives were increased by 8.63%, 6.61%, and 9.75%, respectively. Compared to the full AA 5083 hood panel scenario, our design reduces weight by 16.67%, and achievements of 92.54%, 93.75%, and 106.85% were obtained in lateral, longitudinal, and torsional strain energy, respectively. In summary, our proposed methodology demonstrates considerable potential in improving the, highlighting its significant impact on the optimization of structural performance.
... In recent years, a surge in studies has focused on control strategies for the IMP, with contributions emerging from both industrial and academic spheres [6][7][8][9][10][11][12][13][14][15]. For example, Tan et al. [16] proposed a learning-enhanced proportionalintegral (PI) control method for periodic control of RAM speed in injection molding machines. ...
Article
Full-text available
Injection molding is a pivotal industrial process renowned for its high production speed, efficiency, and automation. Controlling the motion speed of injection molding machines is a crucial factor that influences production processes, directly affecting product quality and efficiency. This paper aims to tackle the challenge of achieving optimal tracking control of injection speed in a standard class of injection molding machines (IMMs) characterized by nonlinear dynamics. To achieve this goal, we propose a learning-based model predictive control (LMPC) scheme that incorporates Gaussian process regression (GPR) to predict and model uncertainty in the injection molding process (IMP). Specifically, the scheme formulates a nonlinear tracking control problem for injection speed, utilizing a GPR-based learning residual model to capture uncertainty and provide accurate predictions. It learns the dynamics model and historical data of the IMM, automatically adjusting the injection speed according to target requirements for optimal production control. Additionally, the optimization problem is efficiently solved using a control-constrained differential dynamic programming approach. Finally, we conduct comprehensive numerical experiments to demonstrate the effectiveness and efficiency of the proposed LMPC scheme for controlling injection speed in IMP.
... The optimization method for reaching the desired mold temperature, uniform temperature distribution, and limiting cooling time becomes more successful by using this approximate mathematical connection. An artificial neural network (ANN/GA) is combined with a genetic algorithm (ANN/GA) in the publication [21]. A new approach for optimizing the injection molding process is offered. ...
Article
Full-text available
Fabricating conformal cooling channels (CCCs) is now easier and more economical because of recent advances in additive manufacturing. CCCs offer better cooling performance throughout the injection molding process than typical (straight drilled) channels. The main reason for this is that CCCs can follow the courses of molded objects, whereas regular channels cannot. CCCs can be used to speed up cycle times, decrease thermal strains and warpage, and produce a more uniform temperature distribution. Using computer-aided engineering (CAE) simulations, designs that are both effective and economical can be made. The goal of this study is to optimize the design of an injection mold to speed up ejection and improve temperature uniformity. This work developed an optimization approach that improved the position of cooling channels during the mold design stage, enabling the construction of geometrically pre-optimized molds. It is safe to assume that the developed technique is efficient and suitable for the task’s intended objectives.
... Uniformity in cooling was then optimized through GA for varying channel diameter, pitch among others for SCC, which were thought analogous to CCC. Changyu et al. [20] optimized and reduced the extents of volumetric shrinkage of an industrial part using ANN coupled with a GA. ...
Preprint
Full-text available
There is a popular use and demand of plastic products worldwide. Therefore, their quality and production rate ought to be high. Injection molding has gained populace in the manufacture of plastics because it is economical, repeatable and can manufacture complex shaped products. However, straight cooling channels characterize most injection mold tools which leads to slow and non-uniform cooling consequently resulting in low production rate and defective products. This study deals with optimization of circular cross-section conformal cooling channel design. Design of experiments is based on Taguchi method with the variable design factors being diameter, depth and pitch. Solidworks is used for 3D design, numerical simulation is conducted using Solidworks Plastics and grey relational analysis is used for multi-response optimization. Grey relational grade results showed that the optimal design was characterized by a minimum diameter (D), depth and pitch of 8 mm, 1.5 D and 2 D respectively. Analysis of variance (ANOVA) findings depicted the diameter as the only and most significant factor that contributed to all the responses concurrently. The performance of all CCC designs were superior to that of SCC. Overall, Grey relational analysis sufficed multi-response optimization in injection molding process through CCC design. The methodology adopted can be used to optimize the designs of injection mold tool components for efficiency in the production process.
... The hyperbolic tangent sigmoid transfer function was used to connect the input layer with the hidden layer(s), whereas the linear transfer function was utilised to connect the hidden layer(s) with the output layer. The effectiveness of this combination of transfer functions has been empirically shown to provide satisfactory outcomes in most training instances [488]. ...
Thesis
Full-text available
Digital Twins, Artificial Intelligence, Computer Aided Engineering
... Depending on the use, hot work steel dies can be exposed to a wide range of temperatures. On the lower end is plastic injection molding, where the die temperature rarely exceeds 100 °C [3][4][5]. On the high end is hot forging, where during the dwell time, when the tool is in contact with the heated material, the tool can reach a peak temperature of around 700 °C. It is then followed, by rapid cooling, due to the application of lubricant, cooling it down to about 200 -350 °C [6][7][8]. ...
... [1][2][3][4][5] Essentially, IM is a complex phase transition process of polymer melt under constantly changing speeds, pressures, and temperatures, which inevitably leads to various defects in optics, materials, and geometry of molded parts. [6][7][8][9] For molded parts with high optical performance, the iridescent pattern seriously affects the appearance and visibility of molded parts and is an IM defect that must be faced. For example, the iridescent patterns of the car dashboard manufactured by IM disturb the driver's view, which may lead to traffic accidents; The iridescent patterns on the molded lens also affect the imaging quality of optical instruments. ...
Article
Full-text available
Iridescent patterns are insufferable molding defects during mass manufacturing of polymer optic devices by using injection molding (IM) technique and seriously affect the appearance and imaging performance of parts. However, at present, the research on the factors affecting iridescent pattern defects is lacking, resulting in the difficulty of obtaining effective methods to weaken iridescent pattern defects. This article aims to clarify the correlation mechanism between injection process parameters and iridescent patterns and provide a theoretical basis for reducing iridescent pattern defects. Based on the process characteristics of IM and the optical mechanism of iridescent patterns, the correlation between iridescent patterns and influencing factors such as optical characteristics and the material structure is established. The influence mechanism of process parameters on iridescent patterns is further clarified by experiments. The results show the iridescent pattern color is related to the retardation, while the iridescent pattern brightness is affected by the retardation and transmittance. Process parameters change the retardation and transmittance by varying factors such as the thickness, birefringence, and crystallinity. Higher mold temperature, reasonable melt temperature, injection speed, and holding pressure are conducive to reducing the colorful fringe area and iridescent pattern brightness, thereby weakening the iridescent patterns on the part.
... In addition to the influences described, several disturbance variables make the control of production difficult [15,16]. To ensure the production of good parts during the continuous operation of the machine, a variety of supervised machine learning methods can be used to model part characteristics as a function of machine setting parameters. ...
Article
Full-text available
The performance of an injection molding machine (IMM) influences the process and the quality of the parts manufactured. Despite increasing data collection capabilities, their machine-specific behavior has not been extensively studied. To close corresponding research gaps, the machine-specific behavior of two hydraulic IMMs of different sizes and one electric IMM were compared with each other as part of the investigations. Both the start-up behavior from the cold state and the behavior of the machine at different operating points were considered. To complement this, the influence of various material properties on the machine-specific behavior was investigated by processing an unreinforced and glass-fiber-reinforced polyamide. The results obtained provide crucial insights into machine-specific behavior, which may, for instance, account for disparities between computer fluid dynamic (CFD) simulations and experimental results. Furthermore, it is expected that the description of the machine-specific behavior can contribute to transfer knowledge when applying transfer learning algorithms. Looking ahead to future research, it is advised to create what is referred to as a “machine fingerprint”, and this proposal is accompanied by some preliminary recommendations for its development.
... They demonstrated the feasibility of this method through actual experiments. Shen et al. [7] mentioned that injection molding process parameters have a significant impact on product quality. They proposed a model architecture that combines artificial neural networks with genetic algorithms. ...
Article
Full-text available
For a long time, the traditional injection molding industry has faced challenges in improving production efficiency and product quality. With advancements in Computer-Aided Engineering (CAE) technology, many factors that could lead to product defects have been eliminated, reducing the costs associated with trial runs during the manufacturing process. However, despite the progress made in CAE simulation results, there still exists a slight deviation from actual conditions. Therefore, relying solely on CAE simulations cannot entirely prevent product defects, and businesses still need to implement real-time quality checks during the production process. In this study, we developed a Back Propagation Neural Network (BPNN) model to predict the occurrence of short-shots defects in the injection molding process using various process states as inputs. We developed a Back Propagation Neural Network (BPNN) model that takes injection molding process states as input to predict the occurrence of short-shot defects during the injection molding process. Additionally, we investigated the effectiveness of two different transfer learning methods. The first method involved training the neural network model using CAE simulation data for products with length–thickness ratios (LT) of 60 and then applying transfer learning with real process data. The second method trained the neural network model using real process data for products with LT60 and then applied transfer learning with real process data from products with LT100. From the results, we have inferred that transfer learning, as compared to conventional neural network training methods, can prevent overfitting with the same amount of training data. The short-shot prediction models trained using transfer learning achieved accuracies of 90.2% and 94.4% on the validation datasets of products with LT60 and LT100, respectively. Through integration with the injection molding machine, this enables production personnel to determine whether a product will experience a short-shot before the mold is opened, thereby increasing troubleshooting time.
... This integration has been made possible via the development of intelligent information technology and data computing, ushering in a new era of AI-driven solutions. Among them, there is a growing industrial demand for artificial neural networks (ANNs), which have shown strong performance in unraveling complex nonlinear relationships, making them one of the most promising languages in the field of artificial intelligence [3,4]. ...
Article
Full-text available
In this study, a multi-input, multi-output-based artificial neural network (ANN) was constructed by classifying output parameters into different groups, considering the physical meanings and characteristics of product quality factors in the injection molding process. Injection molding experiments were conducted for bowl products, and a dataset was established. Based on this dataset, an ANN model was developed to predict the quality of molded products. The input parameters included melt temperature, mold temperature, packing pressure, packing time, and cooling time. The output parameters included mass, diameter, and height of the molded product. The output parameters were divided into two cases. In one case, diameter, and height, representing length, were grouped together, while mass was organized into a separate group. In the other case, mass, diameter and height were separated individually and applied to the ANN. A multi-task learning method was used to group the output parameters. The performance of the two constructed multi-task learning-based ANNs was compared with that of the conventional ANN where the output parameters were not separated and applied to a single layer. The comparative results showed that the multi-task learning architecture, which grouped the output parameters considering the physical meaning and characteristics of the quality of molded products, exhibited an improved prediction performance of about 32.8% based on the RMSE values.
... The results showed that GA gave better results than GBA in terms of cooling time, temperature uniformity, and warpage. Volumetric shrinkage was also optimized for an industrially produced part using ANN together with GA [21]. ...
Preprint
Full-text available
The widespread use and demand for plastic products worldwide have caused manufacturers to covet high productivity and product quality. Most plastic products are produced using the injection molding technique. This technique is characterized by long cooling times, which affect the production cycle and product quality. According to the literature, cooling during the injection molding process can be significantly affected by the design of the cooling channels. This study is, therefore focused on multi-factor design optimization of circular cross-section conformal cooling channels for multiple responses. The Taguchi design-of-experiments approach was adopted in this study. The key variables of conformal cooling channels that were studied involved diameters, depths, and pitches. Solidworks® was used for 3D design and numerical simulation to determine the cooling time, volumetric shrinkage, warpage, and sink marks. Multi-response optimization was then conducted using the Taguchi-Grey Relational Analysis technique. Results show that the optimal cooling channel design has a minimum; diameter of 8 mm, a depth of 12 mm, and a pitch of 16 mm. Additionally, Analysis of Variance (ANOVA) revealed that the diameter is the significant cooling channel design parameter contributing to all the responses concurrently, with the most significant percentage of 80.26%. Comparing the conformal and straight cooling channel designs, superior performance was noted for the former against the latter, with the optimal design recording an improvement of 29.35%, 5.99%, 19.77%, and 38.85% in the cooling time, volumetric shrinkage, warpage, and depth of sink marks, respectively.
... Over the years, there has been a substantial efort toward developing structural health monitoring (SHM) algorithms for mechanical structures. Several techniques have been applied for damage detection, classifcation, and parameter estimation, such as support vector machines [31][32][33][34], neural networks [35][36][37][38][39][40][41][42], and hierarchical neural networks [43][44][45]. Data compression strategies along with hierarchical networks have also been reported [46][47][48]. ...
Article
Full-text available
A vibration data-driven structural defect identification and classification technique is developed using frequency response under random excitation and a hierarchical neural network. A system of artificial neural networks (ANNs) is trained using finite element simulation-based synthetic data to reduce the need for many sensor measurements required otherwise. Principal component analysis (PCA) is employed to compress the high dimensionality of the vibration response data and eliminate the noise effect in the training and testing. Frequency responses data dimension for the structure with defects such a crack from stress concentration, rivet hole expansion, and attached foreign object mass such as ice accumulation in aircraft wing or fuselage are reduced using PCA and fed to a classifier network. The probabilistic decision output from the classifier network and the compressed data are then fed to the next levels of estimator networks, where each network is dedicated to the individual type of defect for the estimation of the defect parameters corresponding to that class of defect. The methodology is applied to a stiffened panel structure. The cracks and rivet hole expansions are introduced in the rivet line of the stiffener, and the foreign object mass is attached to the panel surface. The results show that it is possible to classify the defects and further estimate the defect parameters with good accuracy and reliability. It was observed that the damage classification network had an accuracy of roughly 95%. The damage localization network for crack as well as rivet expansion had average absolute error of around 2. The damage severity network was also able to perform well with a mean absolute error of about 0.34 for crack length detection and 0.22 for expanded rivet damage. However, the damage localization and severity prediction networks were quite challenging to train in the presence of multiple damages and need further development in the network architecture.
... However, most process parameters can be changed with ease. Process parameters have also been optimized by many researchers (Changyu, Lixia and Qian, 2007;Kitayama et al., 2017Kitayama et al., , 2018 making cooling channel design a desirable study area. ...
Preprint
Full-text available
The widespread use and demand for plastic products worldwide have caused manufacturers to covet high productivity and product quality. Most plastic products are produced using the injection molding technique. This technique is characterized by long cooling times, which affect the production cycle and product quality. Literature reveals that cooling in injection molding can be significantly affected by the design of the cooling channels. This study is, therefore focused on design optimization of circular cross-section conformal cooling channels. The Taguchi design of experiments approach was adopted in this study. The key variables of conformal cooling channels that were studied involved diameters, depths, and pitches. Solidworks® was used for 3D design, and for numerical simulation to determine the cooling time, volumetric shrinkage, warpage, and sink marks. Multi-response optimization was then conducted using the Taguchi-Grey Relational Analysis technique. Results show that the optimal cooling channel design has a minimum; diameter of 8 mm, depth of 12 mm and pitch of 16 mm. Additionally, Analysis of Variance (ANOVA) revealed that the diameter is the significant cooling channel design parameter that contributes to all the responses concurrently with the largest percentage of 80.26%. Comparing the conformal with straight cooling channel designs, superior performance was noted for the former against the later with optimal design recording an improvement of 29.35%, 5.99%, 19.77%, and 38.85% in the cooling time, volumetric shrinkage, warpage, and depth of sink marks respectively.
... Additionally, due to the machine learning (ML) algorithms can construct analytic mappings from input features to output responses (Chen et al. 2008;Wang et al. 2022), some studies applied ML as a surrogate method to rapidly figure out the optimal parameters, such as artificial neural networks (ANN), support vector regression (SVR), and K means. Shen et al. (2007) proposed an ANN combined with an intelligent heuristic algorithm to optimize the process parameters of an injection molding machine. In this method, the CAE simulation data was used as the dataset to train the model between injection process parameters and volume shrinkage of parts. ...
Article
Full-text available
The high energy intensity and rigorous quality demand of injection molding have received significant interest under the background of the soaring production of global plastic industry. As multiple parts can be produced in a multi-cavity mold during one operation cycle, the weight differences of these parts have been demonstrated to reflect their quality performance. In this regard, this study incorporated this fact and developed a generative machine learning-based multi-objective optimization model. Such model can predict the qualification of parts produced under different processing variables and further optimize processing variables of injection molding for minimal energy consumption and weight difference amongst parts in one cycle. Statistical assessment via F1-score and R² was performed to evaluate the performance of the algorithm. In addition, to validate the effectiveness of our model, we conducted physical experiments to measure the energy profile and weight difference under varying parameter settings. Permutation-based mean square error reduction was adopted to specify the importance of parameters affecting energy consumption and quality of injection molded parts. Optimization results indicated that the processing parameters optimization could reduce ~ 8% energy consumption and ~ 2% weight difference compared with the average operation practices. Maximum speed and first-stage speed were identified as the dominating factors affecting quality performance and energy consumption, respectively. This study could contribute to the quality assurance of injection molded parts and facilitate energy efficient and sustainable plastic manufacturing.
... ANNs and GA were used to study shrinkage. Researchers found a relationship between design variables and the target parameter, which showed a great ability of ANN to predict the results [16]. In two separate studies, Kitayama et al. arranged two multi-objective optimizations that both used the RBF method to improve the weld lines. ...
Article
Full-text available
Injection molding is one of the most important processes for the mass production of plastic parts. In recent years, many researchers have focused on predicting the occurrence and intensity of defects in injected molded parts, as well as the optimization of process parameters to avoid such defects. One of the most frequent defects of manufactured parts is blush, which usually occurs around the gate location. In this study, to identify the effective parameters on blush formation, eight design parameters with effect probability on the influence of this defect have been investigated. Using a combination of design of experiments (DOE), finite element analysis (FEA), and ANOVA, the most significant parameters have been identified (runner diameter, holding pressure, flow rate, and melt temperature). Furthermore, to provide an efficient predictive model, machine learning methods such as basic artificial neural networks, their combination with genetic algorithms, and particle swarm optimization have been applied and their performance analyzed. It was found that the basic artificial neural network (ANN), with an average accuracy error of 1.3%, provides the closest predictions to the FEA results. Additionally, the process parameters were optimized using ANOVA and a genetic algorithm, which resulted in a significant reduction in the blush defect area.
... Fernandes et al. (2018) extensively studied the optimization of an injection molding process. Shen et al. (2007) and Yin et al. (2011) used a back-propagation ANN and a genetic algorithm to optimize the process parameters of a plastic injection molding process. In this approach, ANN is utilized as a surrogate model of the objective function in the optimization. ...
Article
Full-text available
This paper proposes a supervised learning with a class-balancing loss function (SL-CBL) approach for fault detection and feature-similarity-based recipe optimization (FSRO) for a plastic injection molding process. SL-CBL is a novel method that can accurately classify an input sample as a normal or fault condition, even when the training data is severely class-imbalanced. The proposed class-balancing loss function consists of the weighted focal loss and the loss of the F1-score; together, these are used to correctly classify even a small number of faulty samples. SL-CBL is investigated with four classifiers of different structures; the classifiers consist of several fully connected and batch normalization layers. FSRO is an optimization scheme that finds the optimal recipe whose feature is similar to the features of normal samples. The optimal solution is obtained by minimizing the Euclidean distance to the centroid of the normal features. In this research, the proposed SL-CBL and FSRO methods are validated by applying them to an industrial plastic injection molding dataset. The validation results show that the proposed SL-CBL approach achieves the highest F1-score with the lowest misclassification rate, as compared to the alternative methods. When visualizing the feature space, the optimal recipe found by the FSRO scheme was found to be close to the centroid of the normal features, even if the initial recipe is classified as a fault. Furthermore, each variable of the optimized recipe lies within the confidence interval of 3σ{\rm{\sigma }} for the normal condition. This indicates that the optimal recipe is statistically similar to the normal samples.
... One of the objectives of the Taguchi method is to reduce the number of tests required, thus improving test efficiency. It was applied to optimize the shrinkage and warpage reduction of plastic processes and then has been compared with the application of finite elements and Moldflow ® [32][33][34][35][36]. Gray Relation Analysis (GRA) constitutes a tool with an approach to solving multi-objective optimization problems. ...
Article
Full-text available
The consumer market has changed drastically in recent times. Consumers are becoming more demanding, and many companies are competing to be market leaders. Therefore, companies must reduce rejects and minimize their operating costs. One problem that arises in producing plastic parts is controlling deformation, mainly in the form of shrinkage due to the material and warpage associated with the geometry of the parts. This work presents a novel extended adaptive weighted sum method (EAAWSM: Extended Adaptive Weighted Summation Method) integrated into a Pareto front model. The performance of this model is evaluated against three other conventional optimization methods—Taguchi–Gray (TG), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), and Model Optimization by Genetic Algorithm (MOGA)—and compared with EAAWSM. Two response variables and three input factors are considered to be analyzed: material melting temperature, mold temperature, and filling time. Subsequently, the performance is compared and its behavior observed using Moldflow® simulation. The results show that with the EAAWSM method, the shrinkage is 15.75% and the warpage is 3.847 mm, regarding the manufacturing process parameters of a plastic part. This proposed deterministic model is easy to use to optimize two or more output variables, and its results are straightforward and reliable.
... The most widely used learning algorithm is the gradient descent algorithm incorporated with back propagation (BP). To minimize the error between the estimated output and observed value, the BP algorithm tries to find the optimized weights between the layers of the network [105,106]. After finding the weights with the least error, the network is called a trained network and will be evaluated with a new dataset to calculate its generalization capability [103]. ...
Article
Full-text available
Gully erosion susceptibility mapping (GESM) through predicting the spatial distribution of areas prone to gully erosion is required to plan gully erosion control strategies relevant to soil conservation. Recently, machine learning (ML) models have received increasing attention for GESM due to their vast capabilities. In this context, this paper sought to review the modeling procedure of GESM using ML models, including the required datasets and model development and validation. The results showed that elevation, slope, plan curvature, rainfall and land use/cover were the most important factors for GESM. It is also concluded that although ML models predict the locations of zones prone to gullying reasonably well, performance ranking of such methods is difficult because they yield different results based on the quality of the training dataset, the structure of the models, and the performance indicators. Among the ML techniques, random forest (RF) and support vector machine (SVM) are the most widely used models for GESM, which show promising results. Overall, to improve the prediction performance of ML models, the use of data-mining techniques to improve the quality of the dataset and of an ensemble estimation approach is recommended. Furthermore, evaluation of ML models for the prediction of other types of gully erosion, such as rill–interill and ephemeral gully should be the subject of more studies in the future. The employment of a combination of topographic indices and ML models is recommended for the accurate extraction of gully trajectories that are the main input of some process-based models.
... GAs have been extensively applied in job scheduling [19,27,[30][31][32]. However, most GA applications have not been well explained [33]. ...
Preprint
Full-text available
Many evolutionary artificial intelligence (AI) technologies have been applied to assist job scheduling in manufacturing. Among them, genetic algorithms (GAs) are one of mainstream methods. However, GA applications in this field may not be easy to understand or communicate, especially to factory workers without relevant background knowledge, preventing widespread acceptance of such applications. To address this problem, the concept of explanatory AI (XAI) has been proposed. This study first reviews existing XAI techniques for explaining GA applications in job scheduling. Based on the review results, the problems faced by existing XAI techniques are summarized. To solve these problems, this study proposes several novel XAI techniques, including decision tree-based interpretation, dynamic transformation and contribution diagrams, and improved bar charts. To illustrate the effectiveness of the proposed methodology, it has been applied to a case in the literature. According to the experimental results, the proposed methodology can make up for the deficiencies of existing XAI methods in processing high-dimensional data and visualizing the contribution of feasible solutions, thereby satisfying all the requirements for an effective XAI technique for explaining GA applications in job scheduling. Furthermore, the proposed methodology can be easily extended to explain other evolutionary AI applications such as ant colony optimization (ACO), particle swarm optimization (PSO), artificial bee colony (ABC) in job scheduling.
... By learning from substantial amounts of trained samples, ANN can model the complicated non-linear system without requiring the use of specific analytic formulations and physics-based derivations. The utility of ANNs for successfully solving complex problems, such as optimization [25], fault diagnose [26], health structural monitoring [27], has been demonstrated in many studies over the years. Here, we adopted this technique to predict the relation of residual strength of CFRP composites after high-temperature exposure and the corresponding FTIR spectra. ...
Article
Full-text available
In this paper, a non-destructive inspection (NDI) method of carbon fiber reinforced polymer (CFRP) composites is proposed for flexural residual strength based on Fourier transform infrared (FTIR) test, using an artificial neural network to predict the flexural residual strength of CFRP composites after high-temperature exposure. The thermogravimetric analysis (TGA) and microstructure results showed that the CFRP composites started to decompose at 220 °C. Correspondingly, after high-temperature exposure of 220 °C, as the exposure temperature increased, the residual strength of the composites decreased. The FTIR spectra of the CFRP composites were tested with a handheld infrared instrument. It was found that when the temperature was higher than 220 °C, the characteristic peak intensity of the CFRP composites decreased significantly with the increase of the exposure temperature. Then, an artificial neural network was applied to establish the relationship between the FTIR intensity and the residual strength. The maximum error of the presented ANN model was less than 5%. The residual strength of CFRP composites after high-temperature exposure could be accurately predicted by the artificial neural network based on FTIR testing results.
Article
Full-text available
Injection molding is one of the dominant methods for mass‐producing short fiber reinforced plastics renowned for their exceptional specific properties. In the utilization of such composite components, optimization of process parameters significantly influences material characteristics and part performance. However, in industrial practice, this process often relies on intuition and iterative experimentation. Prior studies have introduced data‐efficient optimization methods but faced limitations in adopting minor variations in the product development cycle. This study introduces a multi‐fidelity optimization framework aimed at efficiently addressing new problems by leveraging previously acquired knowledge from analogous domains, particularly accommodating alterations in material scenarios. Two data‐driven frameworks are explored: 1) Gaussian process‐based and 2) neural network‐based, each employing distinct information‐transferring techniques, hierarchical Kriging and transfer learning, respectively. Bayesian optimization of process parameters under limited data budget, which is typical in realistic industrial settings, is performed. The results highlight the efficiency of the proposed framework, demonstrating superior performance not only in data‐driven modeling but also in optimization efficiency compared to conventional single‐fidelity approaches. The Pearson correlation coefficient is utilized to assess the applicability of the multi‐fidelity framework in handling the inherent ambiguity of the similarity of problem scenarios. The proposed method is believed to be adaptable and versatile, offering potential application across various challenges in process optimization.
Article
Full-text available
Injection molding is a prevalent method for producing plastic components, yet determining the ideal process parameters has predominantly relied on heuristic approaches. In this research, a data‐driven injection molding process optimization framework is developed to simultaneously minimize warpage, cycle time, and clamping force. Employing multi‐objective Bayesian optimization (MBO), the framework is applied to a fan blade model for verification. Incorporating those objectives enables the selection of an injection molding machine with the proper maximum clamping force within acceptable tolerance and production time. Furthermore, non‐moldable regimes are considered in design space, which allows less prior knowledge for setting design space. The optimized results are compared with those obtained using NSGA‐III, a well‐established genetic algorithm‐based optimization technique, as a benchmark. The optimization frameworks produce a Pareto front for the three‐dimensional outputs, revealing distinct trade‐off relationships between cycle time and warpage, as well as clamping force and warpage. The MBO framework demonstrates a superior Pareto front compared to NSGA‐III when utilizing a limited data set, underscoring its benefits in scenarios where costly simulations or experiments are necessary. These findings are anticipated to contribute to the optimization of manufacturing processes, ultimately enhancing productivity in real‐world industries.
Article
Full-text available
This study presents a novel hybrid optimization approach for intelligent manufacturing in plastic injection molding (PIM). It focuses on globally optimizing process parameters to ensure high-quality products while reducing cycle time, material waste, and energy consumption. The method combines a backpropagation neural network (BPNN) with a genetic algorithm (GA) and employs a multi-objective optimization model based on design of experiments (DoE). A BP artificial neural network captures the relationship between optimization goals and process parameters. Leveraging the genetic algorithm, it effectively optimizes process parameters for achieving global optimization goals. The case study involves a polypropylene product, considering dimensional deviation, weight, cycle time, and energy consumption during the PIM cycle. Design variables include melt temperature, injection velocity, injection pressure, commutation position, holding pressure, holding time, and cooling time. The results demonstrate that this approach efficiently adjusts process parameters to meet quality standards, significantly reducing raw material consumption (2%), cycle time (12%), and energy consumption (16%). This offers substantial benefits for companies in highly competitive markets demanding swift adoption of smart production methods.
Article
In this work, the heating and cooling systems of a rapid thermal response mold (RTRM) are optimized applying the response surface methodology (RSM) and an optimization technique. An experiment design with six factors and three levels was carried out, and variables such as the distance between the heating rods, and the distance between the cooling channel and the heating plate, among others, were analyzed. First, thermal, and thermalstructural resistance analyses were performed using the finite element method (FEM), and response surface mathematical models were developed using the mixed regression and response surface models. Then, the analysis of variance (ANOVA) method was used to verify the precision of these models. With the models obtained, the position of the heating rods and the dimensions of the cooling channels were optimized, minimizing at the same time the heating and cooling time and maintaining a reasonable temperature distribution and structural resistance, using the Particle Swarm Optimization (PSO). The results show that it is possible to reduce the heating time by 43%, which is significantly. The cooling time in the molding cycle by 29%, thus improving the effectiveness of the heating and cooling system with respect to the cooling plate, the Von Mises stress reported an improvement in its structural value of 67% with respect to the initial value.
Article
Full-text available
It is desirable in injection moulding that every yielded product is of high quality and precision. Several studies are conducted on modelling the relationship between the process parameters e.g., holding pressure, mould temperature, etc. that are either time series or non-time series, and the output product quality characteristics e.g., weight, dimensions, and warpage. To build a more comprehensive/generalized quality prediction model with online learning, this study for the first time takes time series process parameters from 2 pairs of in-mould pressure and temperature sensors simultaneously with 18 non-time series parameters and introduces a single deep learning architecture to perform multitask learning i.e., to leverage information from the related sources to simultaneously predict multiple product quality characteristics on a real injection moulding machine. This along with the discussed modern containerization practices develops a fully automated software part of the quality control process. The non-time series process parameters are learned by fully connected neural network layers of the model while the time series process parameters are related through an attention-based encoder–decoder architecture. This model is ablated as well as compared with a benchmark model to validate the performance advantage in terms of the reduced error asymptote. The proposed model offers an order of magnitude lower average mean squared error than the benchmark model, and 1.2 to 3 times smaller mean squared errors than its ablated versions.
Article
Full-text available
The manufactured products that rely on molten Acrylonitrile Butadiene Styrene (ABS) via Tederic machine (450[Formula: see text]t) suffer from some types of defects such as black dots, shrinkage, and bubbles, which weaken the competitiveness. Therefore, this work proposes the Jidoka recruit’s network system (JRNS) to set optimal operating parameters to reduce the processes’ output variations and cut down on total process wastage autonomously. The suggested JRNS consists of two sequential stages. The first stage in front propagation relies on response surface methodology (RSM) in classifying the significant operating parameters from 15 candidates experimentally (DOE) to identify the most crucial one. The second stage predicts the processes’ deviation by integrating two meta-heuristic methods called harmony search (HS) and weighted superposition attraction (WSPA) in the backpropagation to automatically reset the operating parameters to keep the product within the standard specification to avoid 11 defect chances. JRNS is innovative and has been used to upgrade autonomous control for enhancing the Tederic machine (450[Formula: see text]t) process to reduce 70.98% of molding defects. The JRNS (RSM, [Formula: see text]) increases the operation efficiency from 94% to 100% and reduces the defects per million opportunity to the six sigma scale.
Article
Full-text available
With the global quest for improved sustainability, partially realized through the electrification of the transport and energy sectors, battery cell production has gained ever‐increasing attention. An in‐depth understanding of battery production processes and their interdependence is crucial for accelerating the commercialization of material developments, for example, at the volume predicted to underpin future electric vehicle production. Over the last five years, machine learning approaches have shown significant promise in understanding and optimizing the battery production processes. Based on a systematic mapping study, this comprehensive review details the state‐of‐the‐art applications of machine learning within the domain of lithium‐ion battery cell production and highlights the fundamental aspects, such as product and process parameters and adopted algorithms. The compiled findings derived from multi‐perspective comparisons demonstrate the current capabilities and reveal future research opportunities in this field to further accelerate sustainable battery production.
Article
This research aims to enhance the design of an injection mold in order to minimize ejection time and increase temperature uniformity by repositioning cooling channels. For 3D transient thermal simulations, the ANSYS Finite Element Method (FEM) program was employed, while the MATLAB program was used for optimization. It was discovered that the method could be used successfully to optimize the design of CCCs in injection molds. The objective function value of the optimized model represents a near 40 percent improvement in relation to the initial model. The combined parameterization, simulation, and optimization method has proven beneficial for enhancing the arrangement of channels in an injection machine mold. The methodology is generalizable. As such, it can be employed in any industrial applications involving injection molding machines, for the improvement of the quality of the manufactured part.
Article
Full-text available
Sustainable energy strategies, particularly alternatives to fossil fuels, e.g., solar-to-hydrogen production, are highly desired due to the energy crisis. Thereby, materials leading to hydrogen production by utilizing water and sunlight...
Article
Improvement is required in any industry to increase the productivity by reducing the defects rate and remove the overall waste produced during manufacturing process. In this research paper, the working parameters are optimised that are responsible for process variation. A plastics manufacturing company, the major problems were flow marks and air bubbles appeared on the surface of mould part a car side mirror plastic cover. Because of these problems, the overall production of company was decreasing that resulted in customer dissatisfaction and company could not achieve its daily production target. The responsible working parameters for process variation are found to be injection pressure, melting temperature along with viscosity and flow rate of molten material during injection moulding process. The injection speed and screw speed are also considered. The analysis of variance (ANOVA) test was conducted to find most critical working parameters. It is identified that response surface methodology (RSM) can be used to optimise the working parameters to improve the quality of moulded part that will increase process efficiency by 50%. The data was collected from four machines to find which machine is causing more problems that is found to be Tederic 450 tone machine. This machine has more problems as compared to other machines. Acrylonitrile Butadiene Styrene (ABS) material is used for the production of car side mirror plastic cover. The Rejection of this moulded part was counted 35% of total rejection. After optimization, the rejection is reduced to 16% is recorded which is significant improvement.
Chapter
This chapter discusses a new application field of XAI in manufacturing—job sequencing and scheduling. It first breaks down job sequencing and scheduling into several steps and then mentions AI technologies applicable to some of these steps. It is worth noting that many AI applications focus on the preparation of inputs required for scheduling tasks, rather than the process of scheduling tasks, which is a distinctive feature of the field. Nonetheless, many AI techniques have already been explained in other fields or domains. These explanations can provide a reference for explaining the application of AI in job sequencing and scheduling. Therefore, some general XAI techniques and tools for job sequencing and scheduling are reviewed, including: referring to the classification of job scheduling problems; customizing scheduling rules; text description, pseudocode; decision trees, flowcharts. Furthermore, job sequencing and scheduling problems are often formulated as mathematical programming (optimization) models to be optimized. AI technologies can be applied to find the best solution for the model. Applications of genetic algorithm (GA) are of particular interest because such applications are most common in job scheduling. Furthermore, XAI techniques and tools for explaining GA can be easily extended to other evolutionary AI applications, such as artificial bee colony (ABC), ant colony optimization (ACO), and particle swarm optimization (PSO) in job scheduling. Applicable XAI techniques and tools include flowcharts, text description, chromosome maps, dynamic line charts, and bar charts with baseline. Some novel XAI techniques and tools for interpreting GAs are also introduced: decision tree-based interpretation and dynamic transition and contribution diagrams.KeywordsXAIJob sequencing and schedulingGADecision tree-based interpretationDynamic transition and contribution diagrams
Chapter
Full-text available
The thermoplastic injection process is an industrial technique that allows getting a high precision plastics part with high production rate. This process is considered one of the most complexes in the plastic industry due to its complexity and variability. The main problems in this technique can occur during two phases: first, during the initial setting when we try to identify the initial parameters for a new plastic part; and second, during mass production when there is a deviation in the production process. The purpose of this article is divided on three parts: first, is to make a basic review and to present overview of the main issues faced in this process, second part, is to present the contribution of the artificial intelligence methods to resolve this issues and finally to present a general guidelines for future researchers to resolve or reduce the process issues.
Article
Multi-stage stamping process is the promising technology to fabricate the metallic bipolar plate (BPP) for proton exchange membrane (PEM) fuel cell. In the present study, a novel die design in the pre-forming stage is proposed and its effect on the formability of ultra-thin metallic BPP is verified by the finite element (FE) simulations of micro- and macro-scale BPP channels. It reveals that the multi-stage forming with the proposed die approach significantly improve the formability of ultra-thin BPP. As a result, the more uniform thickness distribution and considerable reduction of springback are beneficial to the fabrication of high quality metallic BPP. Furthermore, the relatively high reaction efficiency (∼79.4%) of fuel cell stacks can be predicted, indicating the high fuel consumption. These findings demonstrate the feasibility and efficiency of the proposed die design in the fabrication of ultra-thin metallic BPP based on the perspectives of both the formability and energy efficiency.
Article
Full-text available
INTRODUCTION Algorithms for function optimization are generally limited to convex regular functions. However, many functions are multi-modal, discontinuous, and nondifferenName: Christopher R. Houck Address: North Carolina State University, Box 7906, Raleigh, NC, 27695-7906,USA,(919) 5155188, (919) 515-1543,chouck@eos.ncsu.edu Affiliation: North Carolina State University Name: Jeffery A. Joines Address: North Carolina State University, Box 7906, Raleigh, NC, 27695-7906,USA,(919) 5155188, (919) 515-1543,jjoine@eos.ncsu.edu Affiliation: North Carolina State University Name: Michael G. Kay Address: North Carolina State University, Box 7906, Raleigh, NC, 27695-7906,USA,(919) 5152008, (919) 515-1543,kay@eos.ncsu.edu Affiliation: North Carolina State University Sponsor: This research was funded in part by the National Science Foundation under grant number DMI-9322834. 2 Delta C. Houck et al. tiable. Stochastic sampling methods have b
Article
Decreasing warpage is a very significant topic to improve the quality of injection-molded parts. Dimensional stability is an important factor for the minimum warpage of thin shell plastic part. In this study, efficient minimization of warpage on thin shell plastic parts by integrating finite element analysis, statistical design of experiment method, response surface methodology and genetic algorithm is investigated. A thin shell plastic part model is considered as an example. To achieve the minimum warpage, optimum process condition dimensional parameters are determined. X dimension, Y dimension, and Z dimension are considered as model variables. Another parameters of effecting minimum warpage are taken into consideration as constant, such as mold temperature, melt temperature, injection time, injection pressure, etc. Finite element analyses are conducted for combination of process parameters organized using statistical three-level full factorial experimental design. A predictive model for warpage is created using response surface methodology exploiting finite element analysis results. Response surface model is interfaced with an effective genetic algorithm to find the optimum process parameter values.
Article
The basis for a novel pattern-based closed-loop control strategy for the injection molding process is presented. The strategy uses artificial neural networks (ANNs) embedded within a cascade design to analyze sensor patterns, identify process character and control part quality. The platform for this work, the injection molding process, is an industrially significant, cyclic manufacturing operation. Final part quality of this process is a nonlinear function of many machine and polymer variables. Part quality control of this process is currently attained via single input–single output machine controls supervised by human operators. Presented here is a method that employs ANN technology to improve upon this approach and provide the basis for closed-loop part quality control. In the cascade design, machine controller set-points of an inner loop are updated based on ANN analysis of mold cavity pressure patterns. The controller action maintains the desired pressure pattern set-point of the outer loop associated with desired part quality. Control strategy details are provided along with set-point tracking demonstrations that support feasibility of this pattern-based approach.
Article
In this study, optimum values of process parameters in injection molding of a bus ceiling lamp base to achieve minimum warpage are determined. Mold temperature, melt temperature, packing pressure, packing pressure time and cooling time are considered as process parameters. In finding optimum values, advantages of finite element software MoldFlow, statistical design of experiments, artificial neural network and genetic algorithm are exploited. Finite element analyses are conducted for combination of process parameters designed using statistical three-level full factorial experimental design. A predictive model for warpage is created using feed forward artificial neural network exploiting finite element analysis results. Neural network model is validated for predictive capability and then interfaced with an effective genetic algorithm to find the optimum process parameter values. Upon optimization, it is seen that genetic algorithm reduces the warpage of the initial model of the bus ceiling lamp base by 46.5%.
Article
Plastic injection processes comprise plastication, injection, packing, cooling, ejection and process/part quality control applications. These steps are followed for the parts, which are designed to be produced by plastic injection method. Having initial knowledge about the process that will be undertaken is necessary because of present-day competitive conditions that forces us to produce faster and cheaper with a higher quality, such as minimum warpage, sink marks, etc. Computer-aided analysis and engineering softwares are used in order to meet this necessity. For plastic injection process, one of the commercial computer-aided engineering softwares is the MoldFlow Plastic Insight. In this study, best gate location, filling and flow, warpage applications have done for minimum warpage of plastic part with this tool. Process parameters such as mold temperature, melt temperature, packing pressure, packing time, cooling time, runner type and gate location are considered as model variables. The effects of process parameters for thin shell plastic part were exploited using both design of experiment (DOE), Taguchi orthogonal array and finite element software MoldFlow (FE). The most important process parameters influencing warpage are determined using finite element analysis results based on analysis of variance (ANOVA) method. Artificial neural network (ANN) is interfaced with an effective GA to find the minimum warpage value.
Article
Variable retort temperature (VRT) thermal processing can be an effective technique used for improving the quality of canned foods and reduce the process time. In this study, artificial intelligent techniques were applied to develop prediction models and search for optimal variable retort temperature processing conditions for conduction heated foods. A computer simulation program was used to gather data needed for training and testing of artificial neural network (ANN) models. Two VRT functions, sine and exponential, were used for modeling and optimization. ANN models were implemented to develop prediction models for VRT output variables: average quality retention (Qv), process time (Pt) and surface cook value (Fs). Genetic algorithms (GA) were coupled with trained neural network models to meet different optimization objectives: minimum Pt and Fs, under given constraints. The searching range of each independent variable was based on a sensitivity analysis of effects of function parameters on response variables. The best results for Qv, Pt and Fs under constant retort temperature (CRT) processing conditions were used as constraints. Test results indicated that coupled ANN–GA models could be effectively used for describing the relationships between the operating variables and VRT function parameters, and for identifying optimal processing conditions. VRT processes reduced the process time by more than 20% and surface cook value by about 7–10% as compared to the best CRT process.
Article
A neural-network model has been developed to predict the value of a critical strength parameter (internal bond) in a particleboard manufacturing process, based on process operating parameters and conditions. A genetic algorithm was then applied to the trained neural network model to determine the process parameter values that would result in desired levels of the strength parameter for given operating conditions. The integrated NN–GA system was successful in determining the process parameter values needed under different conditions, and at various stages in the process, to provide the desired level of internal bond. The NN–GA tool allows a manufacturer to quickly determine the values of critical process parameters needed to achieve acceptable levels of board strength, based on current operating conditions and the stage of manufacturing.
Article
A neural network model for predicting the quality or soundness of the injected plastic parts based on key process variables and material grade variations has been developed. The approach uses a backpropagation 4-2-3 net (BPN) trained based on inputs/outputs data which were taken from simulation works carried out through a CAE software. The system reduces the time required for planning and optimizing of process conditions or operating parameters.
Article
The optimal use of fuel and electricity in a direct-fired absorption chiller system is important in achieving economical operation. Previous work on the control schemes mainly focused on the component local feedback control. A system-based control approach, which allows an overall consideration of the interactive nature of the plant, the building and their associated variables is seen to be the right direction. This paper introduces a new concept of integrating neural network (NN) and genetic algorithm (GA) in the optimal control of absorption chiller system. Based on a commercial absorption unit, neural network was used to model the system characteristics and genetic algorithm as a global optimization tool. The results appear promising.
Article
The use of Recurrent Neural Networks is not as extensive as Feedforward Neural Networks. Training algorithms for Recurrent Neural Networks, based on the error gradient, are very unstable in their search for a minimum and require much computational time when the number of neurons is high. The problems surrounding the application of these methods have driven us to develop new training tools. In this paper, we present a Real-Coded Genetic Algorithm that uses the appropriate operators for this encoding type to train Recurrent Neural Networks. We describe the algorithm and we also experimentally compare our Genetic Algorithm with the Real-Time Recurrent Learning algorithm to perform the fuzzy grammatical inference.