A Review on Machine Learning Models in Injection Molding Machines

Selvaraj, Senthil Kumaran; Raj, Aditya; Rishikesh Mahadevan, R.; Chadha, Utkarsh; Paramasivam, Velmurugan

doi:https://doi.org/10.1155/2022/1949061

Advances in Materials Science and Engineering

On this page

Abstract Introduction Conclusions Disclosure Conflicts of Interest Authors’ Contributions References Copyright Related Articles

Special Issue

Sustainable Manufacturing, Optimization, and Improvement of Mechanical Properties of High-Strength Steel

View this Special Issue

Review Article | Open Access

Volume 2022 | Article ID 1949061 | https://doi.org/10.1155/2022/1949061

A Review on Machine Learning Models in Injection Molding Machines

Senthil Kumaran Selvaraj,¹Aditya Raj,²R. Rishikesh Mahadevan,¹Utkarsh Chadha,¹and Velmurugan Paramasivam³

Academic Editor: Fuat Kara

Received07 Nov 2021

Accepted10 Dec 2021

Published05 Jan 2022

Abstract

One of the most suitable methods for the mass production of complicated shapes is injection molding due to its superior production rate and quality. The key to producing higher quality products in injection molding is proper injection speed, pressure, and mold design. Conventional methods relying on the operator’s expertise and defect detection techniques are ineffective in reducing defects. Hence, there is a need for more close control over these operating parameters using various machine learning techniques. Neural networks have considerable applications in the injection molding process consisting of optimization, prediction, identification, classification, controlling, modeling, and monitoring, particularly in manufacturing. In recent research, many critical issues in applying machine learning and neural network in injection molding in practical have been addressed. Some problems include data division, collection, and preprocessing steps, such as considering the inputs, networks, and outputs, algorithms used, models utilized for testing and training, and performance criteria set during validation and verification. This review briefly explains working on machine learning and artificial neural network and optimizing injection molding in industries.

1. Introduction

Injection molding is one of the widely used manufacturing processes for manufacturing plastic products [1]. More than 30% of the weight of all plastic products is manufactured using the injection molding process. Besides plastic products, injection molding is mainly used in the automotive, medical, and electronics industries [2]. High production rate and quality are some of the significant advantages of injection molding. The finished products from injection molding require minor finishing operations, which is another advantage [3]. However, the equipment used in injection molding can be quite expensive and sophisticated, so injection molding is not suitable for small-scale production [4–7].

Injection molding is melting plastic consisting of two types of thermoplastic polymers and thermosetting polymers. The machine injects the pressure onto the mold cavity that fills and solidifies the molten plastic to produce an end product. A whole injection molding process consists of three stages that are filling, postfilling, and mold opening. Some of the interesting properties of plastic are as follows: transparency, ability to make complex sizes and shapes quickly, ability to integrate with other materials easily with an excellent thermal and electrical insulator, lightweight, and anticorrosion. Some of the issues regarding the quality are flow marks, volumetric shrinkage, jetting, weld lines, flash, sink marks, dimension shrinkage variation, and warpage. Recently, neural network and machine learning have grown drastically to solve problems that will possibly occur in the injection molding process to improve its characteristics [8–11].

Injection molding consists of four stages: filling stage, holding stage, cooling stage, and demolding stage [12]. The material is fed into the injection molding machine and is then heated. The molten material is forced under pressure into the mold and then cooled and hardened inside the mold. It is then ejected out of the mold. The cycle starts from the filling stage and ends at the demolding stage. As shown in Figure 1, a significant portion of the total time is taken by the cooling and recovery stage. Many industries are working on reducing the cooling time by implementing modifications in mold design [13]. Cooling channels in mold helps to reduce the cooling time and decrease the cycle time. More production cycles can be completed by decreasing cycle time, improving the industry’s productivity [14]. Injection molding is prone to defects, and some of the common defects are flow lines, weld lines, flash, and sink marks. Injection speed, pressure, and mold design are essential in reducing defects [15]. Most of the defects are caused by improper injection pressure and temperature control. So, control over these parameters is crucial in reducing defects [16]. Traditional methods rely on the operator’s expertise and conventional defect detection techniques. These techniques are unreliable and time-consuming, and hence they increase the lead time and reduce the industries’ productivity. So, there is a need for faster monitoring [17]. Implementing machine learning techniques in injection molding can be a better alternative to traditional methods. Although establishing a machine learning network can be expensive, it is profitable in the long run. Machine learning techniques can also be used in optimizing the mold design to reduce the cooling rate and improve the quality of the final product [18–20].

Machine learning is an excellent application in the present era, which provides systems the ability to automatically learn and improve from previous training given to them without being explicitly hard coded or programmed. Machine learning explicitly emphasizes the computer program’s improvement, which runs the data, using it for their learning.

A significant intention of machine learning is to allow the systems or computers to learn correctly without any human intervention. This review intends to accumulate all the ML models implemented in injection molding.

2. Injection Molding

Injection molding is a manufacturing process in which molten material is injected into a mold to produce parts (Figure 2) [22]. Different materials can be molded into desired shapes using injection molding. Some widely used materials are polyvinyl chloride (PVC) (Figure 3), nylon, polyamide, polyester, and acrylic [23]. Except for aluminum and magnesium, almost all metals can be molded using the metal injection molding process. The adherent oxide film on the surface of aluminum and magnesium powder hinders the sintering step in the molding process. Titanium alloys, low alloy steels, cobalt alloys, and stainless steel are widely used in metal injection molding. So, injection molding offers an excellent material choice.

(a)

(b)

One of the significant advantages of injection molding is its higher production rate [24]. The injection process can be automated, and as a result, the production capacity can be tremendously improved. Usually, injection molding is utilized in the production of smaller parts [25]. More extensive parts required more giant molds and more material to fill the mold, impacting its production rate and efficiency. So, the speed of injection molding depends on the size and complexity of the mold. Labor costs are also drastically reduced due to the automation of injection molding. Another advantage of injection molding is that complex parts can also be molded. Thus, injection molding offers design flexibility [26]. Injection molding is capable of producing accurate parts, maintaining good dimensional stability. So, as a result, the scrap resulting from production is meager [27]. Since injection molding is mainly used to produce parts from non-biodegradable materials like plastics and polymers, keeping the scrap rate minimum is essential. The postproduction or finishing work required after production is also low as high dimensional stability is maintained, and surface finish is also good.

2.1. Steps in Injection Molding

The main parts of an injection molding machine are mold, clamping unit, hydraulic system, control system, hopper, cooling channels, injection unit, and drive unit. Injection molding consists of four stages.

From Figure 4, in the first stage, the material is heated until it is melted, and it is then injected into a mold under pressure. Material is fed from the hopper and is then melted [28]. A screw is used to transfer the molten material to the injection site. The melted material takes the shape of the mold, and it is allowed to cool to solidify the material [29]. Different methods are used in the preparation of ceramic powders and metallic powders. Carbonyl synthesis, high-pressure water atomization, and atomization using gas are standard methods for preparing metal powders. The median particle size obtained from these three processes is shown in Table 1.

The purity of gas atomized powder is superior to those prepared using water atomization [33]. Silicon nitride is obtained by nitriding of silicon, and aluminum oxide is obtained from the leaching of bauxite. Different chemical methods are used in the manufacture of ceramic powders. In general, ceramic powders are finer than metal powders. Finer powders offer better strength in comparison to coarse powders. Binders are used to provide easy separation from mold and retain the shape and strength of the part formed [34]. The majority of the defects occur during the molding or injection phase. Injection pressure, mold temperature, and mold complexity are significant factors that affect the quality of the produced parts, so to reduce defects and improve the quality of the produced parts, reasonable control is required over these parameters. Hardened steel, aluminum, and beryllium-copper alloys are used for manufacturing molds. Steel is used for making molds that require a longer lifespan, but it is comparatively expensive [35]. Ejection gap, cooling circuit design, and allowances are considered while designing the mold. A good surface finish is required in the core, cavity, and runner.

Pressure is applied continuously by the screw, which melts the material and injects the molten material into the mold [36]. The screw of the injection molding machine moves forward at a slower rate, and the injection rate is also slower. As the material starts to cool down and solidify, the gate solidifies. After this, the screw of the injection molding machine moves back to its original position [37]. To compensate for shrinkages and ensure the quality of the final product, the holding pressure is maintained for some time after the material in the mold completely solidifies. About 60% of the total cycle time of injection molding is in the cooling phase, as shown in Figure 1 [38]. Cooling channels are provided in the mold design to accelerate the cooling process. Incomplete cooling produces defective parts and is not suitable for the application.

The mold release or demolding is the next and final stage in injection molding. When the mold reaches a particular value, the mold is opened by the clamping unit for extracting the final product. Improper molding methods result in uneven surface stress and product deformation. So, the method of the demolding has to be chosen carefully. Stripper plate demolding and ejector demolding are the most widely used demolding methods [39]. When ejector pins or pressurized air cannot remove a part from the mold core, a stripper plate is used. A stripper plate is utilized to push the part from the injection mold core, and complete contact is made between the outer parts of the part and the stripper plate for the injection process. It is more reliable than ejector pins or pressurized air, and it is usually used for thin-wall injection molding. Incomplete molding can increase cycle time, incur high costs, and waste material. After demolding, one cycle of injection molding is completed. Usually, a cycle time can vary from a couple of seconds to a couple of minutes.

Many injection molding processes are widely used for different sets of applications, as shown in Table 2 [33]. The most commonly used injection molding techniques are thermoplastic injection molding and metal injection molding [40]. Thermoplastic injection molding plays a vital role in manufacturing products from plastic and polymers, whereas metal injection molding is mainly used for automotive and electronic applications [41]. The other commonly used injection molding processes are gas-assisted injection molding, microinjection, and thin-wall injection molding. These molding techniques find their use in medical, electronics, and telecommunication fields. There are various parameters to consider while choosing the desired injection molding process for a particular application [24]. Some factors to consider are production capacity, product size, design considerations, and cost of operation and materials.

Injection molding also has certain disadvantages [42]. Designing and modeling injection molding components take much time, and it is a complex process. One such complicated process is designing the mold. Mold design is critical to the accuracy and finishing of the final product, and hence much time is invested in this process. So, this leads to a temporary increase in lead time and hence affects the industry’s productivity [43]. The equipment and necessary apparatus are pretty expensive compared to other manufacturing processes. So, injection molding is not suitable for small-scale production. Also, there are limitations in the size of the product because it is challenging to manufacture large-size products. High tool cost and design cost are incurred in the manufacture of large-sized products, and hence it is not feasible. So, these are some of the limitations of injection molding.

2.2. Defects in Injection Molding

Injection molding is also prone to defects like any other manufacturing process [44]. The majority of the defects can be mitigated by redesigning the molding process and following design guidelines [45]. Commonly observed defects in injection molding are flow lines, weld lines, flash, and sink marks. The causes and prevention of these defects are listed in Table 3.

From Table 3, it can be seen that most of these defects can be eliminated if we have reasonable control over the operating parameters of the injection molding process. So, to establish a reasonable control on operating parameters, we need to implement specific machine learning techniques into the molding process.

Flow lines are lines or patterns appearing on the final product, and they are usually off-toned in color. Small craters formed in dense areas of the final product are called sink marks [45]. They are mainly caused by the improper gate and runner design and inadequate cooling time. Condition in which thin surface layers form on the surface of the part due to the presence of contaminants is called surface delamination. Weld lines are similar to flow lines, but they appear similar to a plane than a line when compared to flow lines. They are caused by improper bonding and low injection pressure and speed [46]. Non-uniform cooling of mold causes warping, creating internal stresses within the part and resulting in distortions. When mold is subjected to high injection pressure and high screw speed, it can degrade the mold’s material, which can cause discolorations in the final product. This type of defect is called a burn mark [16]. Traditional methods of inspection of the parts include visual inspection and various destructive and non-destructive tests. However, the main problem with these techniques is that they are inefficient and rely on the operator’s experience. It is time-consuming and increases the product’s production cycle, affecting the productivity of the industries [47]. Online monitoring can offer a better alternative than conventional methods. It utilizes machine learning techniques in defect detection, and it is more efficient and improves the industry’s productivity. The initial cost of an online monitoring system can be pretty expensive, but they are profitable in the long run [48].

3. Machine Learning Models

Machine learning algorithms are mainly categorized into supervised and unsupervised.

3.1. Supervised Learning

These algorithms can predict the future or unknown data based on their learning in the previous training. The comparison will be made with the expected output and output that the machine has generated, and then the model can be changed accordingly [49].

3.2. Unsupervised Learning

These algorithms are quite the opposite of the supervised learning algorithms, where they will not be trained to classify, but these will get inference from a function and describe the unlabelled dataset [50].

The following is the list of widely used machine learning algorithms. The following algorithms can be used for most of the data problems.

3.3. Linear Regression

It is used to predict the values of one variable based on the values based on other variables. A variation in guessing is called a dependent variable. A variable that we use to predict other variables’ values is called an independent variable. For example, we use linear regression for understanding that test performance could be predicted based upon the updated timing of how to use tobacco based on smoking time [51].

3.4. Logistic Regression

Logistic regression is a supervised learning set of rules using binary type troubles. Though “regression” contradicts classifying the point of interest here in the phrase, logistic refers to the logistic feature doing the classification assignment on these algorithms. Logistic regression is an easy and productive algorithm, and that is why it is usually utilized as much as binary class obligation. Customer churns, e-mail spamming, and website and advert click prediction are few examples wherein logistic regression offers a practical answer. A base about logistic regression is the logistic function, also known as sigmoid function, which takes into any of the actual value numbers, mapping it to the values between zero and one [52].

3.5. Decision Tree

A decision tree classifier is a nonparametric supervised learning algorithm that can be used for classification and regression. A tree is a piecewise constant approximation. The main element of the decision tree classifier is its ability to separate intricate, complex, and dynamic decision-making processes into an assortment of less complex decisions and, along these lines, provides a solution that is simpler to interpret [53].

3.6. SVM

Support vector machine intakes data points and generates the maximum marginal hyperplane (MMH) that best separates the tags. A line can be an example of MMH in 2 dimensions which can be called as the decision boundary. This decision boundary classifies both sides of lines with different classes [54].

3.7. Naive Bayes

Naive Bayes is a supervised set of rules used for type responsibilities. Therefore, it is known as the Naive Bayes classifier.

Naive Bayes assumes that capability is independent of each difference; there is no correlation among features. However, that is not the case in actual lifestyle. The naive assumption about capabilities is not correlated which is why the set of rules is called naive [55].

3.8. KNN

The K-nearest neighbor (KNN) is a nonparametric classification method, which is easy yet powerful in the vast majority of the cases. The algorithm has a significant dependency on the value of K. In practice, for every unseen instance, its K-nearest neighbors within the training set are first of all recognized. From that point forward, the dominant occurring class on this produced neighborhood set of K elements is recognized as a label set for the unseen instance. Predicting was achieved consistently with most of the people classes. Similarly, KNN regressions take mean values about five close factors.

3.9. K-Means

K-means clustering is a vector quantization method that seeks to partition n observations into k clusters, with each observation belonging to the cluster with the closest mean cluster centres, which serves as the cluster's prototype. As a result, the data space is partitioned into Voronoi cells. Squared Euclidean distances are minimised by K-means clustering within cluster variances [57].

3.10. Random Forest

In this machine learning algorithm, a combination of tree predictors is used to predict the output data. These trees depend on the value of a random vector which is sampled independently and with similar distribution for all the trees in the forest. These are also called decision forests for classification and regression by constructing a multitude of decision trees at training time [58].

A neural network is a subset of machine learning (ML) and is the core of deep learning. Neural networks are also called artificial neural networks or SNN. This neural network is mainly inspired by the human brain and how the neurons send the signals to each other. The artificial neural networks consist of three components: the input layer, the hidden layer, and finally the output layer. The neural networks mainly rely on the training data for learning and improving their accuracy over time [53, 59–62].(1)Single-layer feed-forward network: in the single-layer feed-forward network, there will be two layers: the input layer and the output layer. In the input layer, computations are not done, so they will not be counted. When different weights are applied to the input nodes, the cumulative effect for each node is taken, and the output layer is formed.(2)Multilayer feed-forward network: the multilayer feed-forward network has a hidden layer and does not directly contact external layers. Having more hidden layers makes the neural network computationally stronger.(3)Single node with its feedback: in the single node with its feedback, the output could be directed back like input to the same layers or the preceding layer node resulting in a feedback network. Recurrent networks are the feedback networks with a closed loop.(4)Single-layer recurrent network: recurrent neural network is a subset of an artificial neural network in which the connections between the nodes form a directed graph with sequence. The neural network will exhibit dynamic behavior for a time sequence. The single-layer recurrent network will use the internal state or memory to process the input sequences.(5)Multilayer recurrent network: the processing output element may be directed to the processing element in the same and preceding layers in the recurrent multilayer network. The same task is performed for all sequence elements, and the output depends on previous computations. The main feature of RNN is the hidden state that captures information about the sequence.

4. Discussion: ML Implementations and Outcomes

As technology advances, more advanced injection molding types are invented. Some of the examples include metal injection molding and ceramic injection molding. In metal injection molding, powdered metals are fused with alloys of relatively softer metals injected into the mold. Gas-assisted injection molding and water-assisted injection molding are used when the molding parts are very thick and in cases where making hollow parts is necessary to save both cycle time and maintain high product quality. Even existing injection molding processes are also developed to improve performance. For example, in plastic injection molding, new resin formulations are developed to improve the strength and decrease the weight of the product. Many injection molding tools are also developed, which can improve the quality of the products. These tools use machine learning to improve the precision and accuracy of the final products.

The paper’s main focus was on the algorithm used, AI model taxonomy, depth layer sizes, training time, testing time, dataset, framework, core language, interface, advantages, and disadvantages. Based on the algorithm used, classification was done if it was a supervised or unsupervised algorithm. Depth layer size is found to determine the number of layers and the number of neurons in the hidden layers for the most optimized result. Dataset size is also considered to determine if the amount of input training data was sufficient or not to the model to give the most accurate results and predictions. The framework and software also considered which software gives the best visualization and correct prediction of the results. It was found that the most used software programs were MATLAB and C-MOLD. At last, the advantages and disadvantages of the model proposed in the research works are given to date to help the future models to be analyzed accurately and chosen for the right task based on the requirements met by their advantages and disadvantages.

In machine learning, supervised learning (Figure 5) is when the machine will be trained on input data using labels to be given or predict the output with the required accuracy. The scheme like the supervisor or a teacher first teaches the student how to solve and then makes the student more accurate. After training, the model is then tested on test data, which is a part or subset of the training data in different ratios. Since the machine is already trained, it gives the output more reliability. The method in [63] gave the best output by predicting the perfect time when the machine should be changed to minimize damage cost by getting trained using multiple supervised algorithms, compared with other obtained results from various research works shown in Table 4.

Unsupervised learning (Figure 6) is a section of machine learning used in many real-world problems in which the labels are not mentioned, and the hidden patterns from the dataset should be found out. The model by itself tries to generate a pattern from all the datasets without supervision. Based on the similarities in the dataset, the machine tries to find common elements and group or cluster those data together. This method is similar to how a normal human being learns on his or her own. The method in [72] shows the best results using unsupervised cluster analysis that finds the molds wearing out and predicts in real-time, and it has an excellent early warning system which predicts before they wear out and cause trouble. The comparison of such models with other obtained results from various research works is given in Table 5.

Reinforcement learning (Figure 7) is a segment of machine learning where the machine learns the classification of the dataset by the trial and error method like getting rewards for giving correct output and penalties for wrong output to the machine. Using this concept, the model finds how it should be doing a task to maximize rewards and minimize penalties by starting with random data trials. The paper [77] used reinforcement learning by predicting the operating conditions and molding conditions in a short period and performing molding with less consumption power, compared with other obtained results from various research works shown in Table 6.

The artificial neural network (Figure 8) is a segment of artificial intelligence used to simulate or replicate the working of human brain functionality. The ANN consists of processing units that contain inputs and outputs. Through the input, the ANN learns and then produces the required output. The method in [43] gives the best result using ANN by predicting high accuracy and recommending the injection molding process condition on the real and correct time in fix material condition. Table 7 compares research works done in ANN implementation in injection molding.

Backpropagation algorithms (Figure 9) are rules used to guide artificial neural networks (ANNs). It is the most basic block in the neural network. This algorithm can effectively train the neural network by chain rule technique. The backpropagation performs a back pass after every forward pass, which adjusts the model parameter. Here, the method in [13] using the backpropagation algorithm predicts the faults even with less important variables than desired, causing an ANN system to diagnose faults that can be successfully built, compared with other obtained results from various research works shown in Table 8.

The hybrid algorithms in the artificial neural network use different algorithms like swarm optimization, greedy gradient, and backpropagation and combine them to get an improved hybrid algorithm for prediction. The research works have shown that using the hybrid algorithm for ANN gave high process reliability.

They are cost-effective due to reduced production costs and high-quality parameters. Here, the method in [18] gave the best result by predicting the parameters ideally at the test run without consuming and wasting many resources, compared with other obtained results from various research works shown in Table 9.

A genetic algorithm (Figure 10) is a segment of neural network algorithms which are based on nature-inspired property of natural selection. It is a subset of evolutionary computation. This tedious process follows the Darwinian theory. This algorithm's main advantage is that it is faster and efficient than many traditional methods and has an excellent list of solutions rather than just one solution. In the papers using many parameters and increasing the search space, the genetic algorithms fit the best. The method in [108] gives the best result using genetic algorithms by taking the time needed to determine the starting process parameter for injection molding which could be significantly reduced compared with other obtained results from various research works shown in Table 10.

Recurrent neural networks (Figure 11) are a subset of neural networks mainly used for their sequential data algorithm in the famous Apple Siri and the Google search by voice. It contains an internal memory due to which it can remember its inputs that help significantly solve the sequential data problems. In the research by Nagorny et al. [113], the recurrent network gave the best result by showing the best performance about the LSTM layer architectures, compared with other obtained results from various research works shown in Table 11.

5. Summary and Conclusions

For reducing defects in injection molding, control on operating parameters is required [48]. These operating parameters include injection speed and pressure, mold temperature, and mold design [16]. Traditional defect detection methods are unsuitable for detecting defects and controlling operating parameters. They mostly rely on the operator’s expertise, and they are unreliable at times and are not efficient [47]. We can establish reasonable control over the operating parameters with machine learning techniques and reduce errors. One such application of machine learning techniques is in mold design [14]. The mold design takes a great deal of time, and it is a very complex process. Mold design can affect the quality and strength of the final product. Machine learning can be applied in designing and optimizing mold design. It helps to save a lot of time and preproduction delays [47]. The initial setup cost for implementing machine learning to injection molding can be costly, and it is not suitable in low production applications [2]. However, in high production, applications can be profitable in the long run and improve productivity [118–123]. However, in injection molding, there is a constraint on the size of the product. For manufacturing large products, the equipment required is costly and takes a lot of floor space. So, it is not feasible for the manufacture of large products [20, 124–127].

These are some of the significant conclusions that can be drawn from the recent research works which may benefit researchers:(1)Machine learning algorithms can be used in injection molding processes, starting from the molding process until the final step stages.(2)Multiple regression equations that were obtained through design of experiments can help the requirement of many train datasets for ANN.(3)Machine learning counters the drawback of the simulation by reducing repetitive processes significantly and reducing substantial computational power, amount, and time, which replaces the need for professionals to interpret results that again help in the implementation.(4)The most commonly used simulation programs were MATLAB and C-MOLD.(5)Neural network is a perfect model that can be used in the modeling complex interactions among many inputs and outputs along with high prediction with excellent accuracy.(6)The only drawback is the need for a good amount of training data and computational power, time, and cost.(7)The best selection of the hidden layer architecture is not correctly decided because it fails for other applications.(8)A minimal quantity of datasets has been used by many during training. Large datasets should be used to avoid overfitting.(9)The neural networks can be applied for many applications like control, identification, modeling prediction, optimization, monitoring, and classification.(10)ANN can be used to develop a reverse process model effectively.(11)The best predictor was Gaussian process regression with the highest coefficient of determination of 0.995.(12)Even the multiple linear regression is a simple model structure, and it did not perform worst on many datasets.(13)K-nearest neighbor astonishingly produced results below expectation as compared to rest of the algorithms studied(14)Some other algorithms like neural networks, support vector machine, and decision tree have average predictive quality.

6. Future Directions

The future directions to guide the researchers ahead are as follows:(1)The future work for this review concentrates on applying the different methodologies in machine learning and neural networks to many different types of machine data comprising contrasting values. In the future, more machine learning techniques may be made which are best suited for a particular injection molding machine for corresponding variables that gives the best results after analyzing with different test datasets. The self-made customized learning model might give a better result for a particular injection molding machine and its parameters. In this manner, future work might be continued on this review.(2)In the future, the model’s scalability should be taken into consideration. The study’s main aim was to find the prediction models that can find out good and bad values for injection molding parameters. Many different parameters have been taken into consideration while feeding the models with data, and multiple machine learning models, including supervised and unsupervised learning, have been considered. Their outputs were compared with respect to the best and worst predictions to find the best models that give the most accurate results. While feeding and training the model, the dataset ratio should be balanced between several good and bad parts. The data size should also be considered; it should not cause overfitting or underfitting, which might result in wrong predictions in the future.(3)For the future, the work can concentrate mainly on integrating more and more factors or parameters affecting the performance of injection molding machines. It is strongly believed that the optimization strategy is effective. Many materials might be included in the future, and different molds might also be used for the workpieces to give a great application where automated machines are primarily used in the industries. More stress can be given on the parameter optimization and the tolerance design, which are essential activities related to quality; by considering different molds, the scalability of the model increases.(4)In the future, the work corresponding to this review study can prepare or generate various kinds of learning models and algorithms. It includes the complete implementation of machine learning-based models or artificial intelligence-based models or control systems to manufacture different objects using injection molding. It should also involve the validation part of the injection molding machine with different kinds of data (including the edge cases and exceptions) and experiments that make the model complete for usage in terms of prediction and validation of the injection molding machine output using given process control parameters.

Disclosure

Senthil Kumaran Selvaraj and Aditya Raj are first authors.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

Senthil Kumaran Selvaraj and Aditya Raj contributed equally to this study.

References

R. Raviwongse and V. Allada, “Artificial neural network based model for computation of injection mould complexity,” International Journal of Advanced Manufacturing Technology, vol. 13, no. 8, pp. 577–586, 1997.
View at: Publisher Site | Google Scholar
O. Veligorskyi, R. Chakirov, M. Khomenko, and Y. Vagapov, “Artificial neural network motor control for full-electric injection moulding machine,” in Proceedings of the IEEE International Conference on Industrial Technology (ICIT), pp. 60–65, IEEE, Melbourne, Australia, February 2019.
View at: Publisher Site | Google Scholar
R. Ventura and X. Berjaga, “Comparison of multivariate analysis techniques in plastic injection moulding process,” in Proceedings of the IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA), pp. 1–6, IEEE, Luxembourg, UK, September 2015.
View at: Publisher Site | Google Scholar
A. Sharma, A. Chouhan, L. Pavithran, U. Chadha, and S. K. Selvaraj, “Implementation of LSS framework in automotive component manufacturing: a review, current scenario and future directions,” Materials Today Proceedings, vol. 46, pp. 7815–7824, 2021.
View at: Publisher Site | Google Scholar
K. Virmani, C. Deepak, S. Sharma, and U. Chadha, “Nanomaterials for automotive outer panel components: a review,” European Physical Journal Plus, vol. 136, 2021.
View at: Publisher Site | Google Scholar
A. Raj, K. S. Ram, L. Jose, and A. Karn, “A survey of electromagnetic metal casting computation designs, present approaches, future possibilities, and practical issues,” European Physical Journal Plus, vol. 136, 2021.
View at: Publisher Site | Google Scholar
S. K. Selvaraj, R. Ramesh, M. Tharun et al., “New developments in carbon-based nanomaterials for automotive brake pad applications and future challenges,” Journal of Nanomaterials, vol. 2021, Article ID 6787435, 24 pages, 2021.
View at: Publisher Site | Google Scholar
T. Pati, “IOP conference series,” Materials Science and Engineering, vol. 1206, Article ID 012022, 2021.
View at: Publisher Site | Google Scholar
M. Dharnidharka, U. Chadha, L. M. Dasari, and A. Paliwal, “Optical tomography in additive manufacturing: a review, processes, open problems, and new opportunities,” European Physical Journal Plus, vol. 136, 2021.
View at: Publisher Site | Google Scholar
S. D. MamdiwarR. Akshith, Z. Shakruwala, U. Chadha, K. Srinivasan, and C.-Y. Chang, “Recent advances on IoT-assisted wearable sensor systems for healthcare monitoring,” Biosensors, vol. 11, no. 10, p. 372, 2021.
View at: Publisher Site | Google Scholar
R. Sivasubramani, A. Verma, G. Rithvik, U. Chadha, and S. S. Kumaran, “Influence on nonhomogeneous microstructure formation and its role on tensile and fatigue performance of duplex stainless steel by a solid-state welding process,” Materials Today Proceedings, vol. 46, 2021.
View at: Publisher Site | Google Scholar
C. Yen, J. C. Lin, W. Li, and M. F. Huang, “An abductive neural network approach to the design of runner dimensions for the minimization of warpage in injection mouldings,” Journal of Materials Processing Technology, vol. 174, no. 1-3, pp. 22–28, 2006.
View at: Publisher Site | Google Scholar
N. Lopes and B. Ribeiro, “Part quality prediction in an injection moulding process using neural networks,” Proceedings of WMC, ISM., 1999.
View at: Google Scholar
K. S. Lee and J. C. Lin, “Design of the runner and gating system parameters for a multi-cavity injection mould using FEM and neural network,” International Journal of Advanced Manufacturing Technology, vol. 27, no. 11-12, pp. 1089–1096, 2006.
View at: Publisher Site | Google Scholar
Y. C. Lam, L. Y. Zhai, K. Tai, and S. C. Fok, “An evolutionary approach for cooling system optimization in plastic injection moulding,” International Journal of Production Research, vol. 42, no. 10, pp. 2047–2061, 2004.
View at: Publisher Site | Google Scholar
P. Zhao, J. Zhang, Z. Dong et al., “Intelligent injection molding on sensing, optimization, and control,” Advances in Polymer Technology, vol. 2020, Article ID 7023616, 2020.
View at: Publisher Site | Google Scholar
P. K. Yarlagadda and C. A. T. Khong, “Development of a hybrid neural network system for prediction of process parameters in injection moulding,” Journal of Materials Processing Technology, vol. 118, no. 1-3, pp. 109–115, 2001.
View at: Publisher Site | Google Scholar
T. Ghimire, A. Joshi, S. Sen, C. Kapruan, U. Chadha, and S. K. Selvaraj, “Blockchain in additive manufacturing processes: recent trends & its future possibilities,” Materials Today Proceedings, 2021, in Press.
View at: Publisher Site | Google Scholar
S. K. Selvaraj, K. Srinivasan, U. Chadha et al., “Contemporary progresses in ultrasonic welding of aluminum metal matrix composites,” Frontiers in Materials, vol. 8, p. 126, 2021.
View at: Publisher Site | Google Scholar
C. Ajmi, J. Zapata, S. Elferchichi, A. Zaafouri, and K. Laabidi, “Deep learning technology for weld defects classification based on transfer learning and activation features,” Advances in Materials Science and Engineering, vol. 2020, Article ID 1574350, 16 pages, 2020.
View at: Publisher Site | Google Scholar
P. D. Kale, “IOP conference series,” Materials Science and Engineering, vol. 1017, Article ID 012031, 2021.
View at: Publisher Site | Google Scholar
C. C. Zhou, G. F. Yin, and X. B. Hu, “Multi-objective optimization of material selection for sustainable products: artificial neural networks and genetic algorithm approach,” Materials & Design, vol. 30, no. 4, pp. 1209–1215, 2009.
View at: Publisher Site | Google Scholar
P. K. Yarlagadda, “Development of an integrated neural network system for prediction of process parameters in metal injection moulding,” Journal of Materials Processing Technology, vol. 130, pp. 315–320, 2002.
View at: Publisher Site | Google Scholar
A. E. Smith, “Predicting product quality with backpropagation: a thermoplastic injection moulding case study,” International Journal of Advanced Manufacturing Technology, vol. 8, no. 4, pp. 252–257, 1993.
View at: Publisher Site | Google Scholar
I. Ribeiro, P. Peças, and E. Henriques, “Modelling the energy consumption in the injection moulding process,” International Journal of Sustainable Manufacturing, vol. 3, no. 4, pp. 289–309, 2015.
View at: Publisher Site | Google Scholar
K. C. Ke and M. S. Huang, “Quality classification of injection-molded components by using quality indices, grading, and machine learning,” Polymers, vol. 13, no. 3, p. 353, 2021.
View at: Publisher Site | Google Scholar
W. Michaeli, A. Spennemann, and R. Gärtner, “New plastification concepts for micro injection moulding,” Microsystem Technologies, vol. 8, no. 1, pp. 55–57, 2002.
View at: Publisher Site | Google Scholar
U. Maradia, E. Filisetti, M. Boccadoro, M. Roten, J. M. Dutoit, and S. Hengsberger, “Increasing the injection moulding productivity through EDM surface modulation,” Procedia CIRP, vol. 68, pp. 58–63, 2018.
View at: Publisher Site | Google Scholar
Y. Lockner and C. Hopmann, “Induced network-based transfer learning in injection molding for process modelling and optimization with artificial neural networks,” International Journal of Advanced Manufacturing Technology, vol. 112, no. 11, pp. 3501–3513, 2021.
View at: Publisher Site | Google Scholar
Y. Seki, S. Okamoto, H. Takigawa, and N. Kawai, “Effect of atomization variables on powder characteristics in the high-pressured water atomization process,” Metal Powder Report, vol. 45, no. 1, pp. 38–40, 1990.
View at: Publisher Site | Google Scholar
A. Ünal, “Effect of processing variables on particle size in gas atomization of rapidly solidified aluminium powders,” Materials Science and Technology, vol. 3, no. 12, pp. 1029–1039, 1987.
View at: Publisher Site | Google Scholar
L. Xiong and T. He, “Synthesis and characterization of carbon supported PtW catalysts from carbonyl complexes for oxygen electroreduction,” Electrochemistry Communications, vol. 8, no. 10, pp. 1671–1676, 2006.
View at: Publisher Site | Google Scholar
V. M. Kryachek, “Injection moulding,” Powder Metallurgy and Metal Ceramics, vol. 43, no. 7, pp. 336–348, 2004.
View at: Publisher Site | Google Scholar
J. O. Osarenmwinda and D. D. Olodu, “Optimization of injection moulding process parameters in the moulding of High Density Polyethylene (HDPE),” Journal of Applied Sciences & Environmental Management, vol. 22, no. 2, pp. 203–206, 2018.
View at: Publisher Site | Google Scholar
O. Ogorodnyk, O. V. Lyngstad, M. Larsen, K. Wang, and K. Martinsen, “Application of machine learning methods for prediction of parts quality in thermoplastics injection molding,” in International Workshop of Advanced Manufacturing and Automation, Springer, Berlin, Germany, 2018.
View at: Publisher Site | Google Scholar
S. J. Huang and T. H. Lee, “Application of neural networks in injection moulding process control,” International Journal of Advanced Manufacturing Technology, vol. 21, no. 12, pp. 956–964, 2003.
View at: Publisher Site | Google Scholar
L. V. Dihoru, L. N. Smith, and R. M. German, “Experimental analysis and neural network modelling of the rheological behaviour of powder injection moulding feedstocks formed with bimodal powder mixtures,” Powder Metallurgy, vol. 43, no. 1, pp. 31–36, 2000.
View at: Publisher Site | Google Scholar
D. E. Dimla, M. Camilotto, and F. Miani, “Design and optimisation of conformal cooling channels in injection moulding tools,” Journal of Materials Processing Technology, vol. 164, pp. 1294–1300, 2005.
View at: Publisher Site | Google Scholar
W. Michaeli and R. Gärtner, “New demolding concepts for the injection molding of microstructures,” Journal of Polymer Engineering, vol. 26, no. 2-4, pp. 161–178, 2006.
View at: Publisher Site | Google Scholar
H. L. Shu and H. Shu, “Simulation study of PID neural network temperature control system in plastic injecting-moulding machine,” in Proceedings of the International Conference on Machine Learning and Cybernetics, vol. 1, pp. 492–497, IEEE, Hong Kong, China, August 2007.
View at: Publisher Site | Google Scholar
C. C. Chiu, C. T. Su, G. S. Yang, J. S. Huang, S. C. Chen, and N. T. Cheng, “Selection of optimal parameters in gas‐assisted injection moulding using a neural network model and the Taguchi method,” International Journal of Quality Science, vol. 2, 1997.
View at: Publisher Site | Google Scholar
C. Lee, J. Na, K. Park et al., “Development of artificial neural network system to recommend process conditions of injection molding for various geometries,” Advanced Intelligent Systems, vol. 2, no. 10, Article ID 2000037, 2020.
View at: Publisher Site | Google Scholar
R. Sedighi, M. S. Meiabadi, and M. Sedighi, “Optimisation of gate location based on weld line in plastic injection moulding using computer-aided engineering, artificial neural network, and genetic algorithm,” International Journal of Automotive and Mechanical Engineering, vol. 14, no. 3, 2017.
View at: Publisher Site | Google Scholar
M. Khomenko, O. Veligorskyi, R. Chakirov, and Y. Vagapov, “An ANN-based temperature controller for a plastic injection moulding system,” Electronics, vol. 8, no. 11, 2019.
View at: Publisher Site | Google Scholar
H. Zhou and D. Li, “Modelling and prediction of weld line location and properties based on injection moulding simulation,” International Journal of Materials and Product Technology, vol. 21, no. 6, pp. 526–538, 2004.
View at: Publisher Site | Google Scholar
D. Torkar, S. Novak, and F. Novak, “Apparent viscosity prediction of alumina–paraffin suspensions using artificial neural networks,” Journal of Materials Processing Technology, vol. 203, no. 1-3, pp. 208–215, 2008.
View at: Publisher Site | Google Scholar
T. Wuest, D. Weimer, C. Irgens, and K. D. Thoben, “Machine learning in manufacturing: advantages, challenges, and applications,” Production & Manufacturing Research, vol. 4, no. 1, pp. 23–45, 2016.
View at: Publisher Site | Google Scholar
I. Meekers, P. Refalo, and A. Rochman, “Analysis of process parameters affecting energy consumption in plastic injection moulding,” Procedia CIRP, vol. 69, pp. 342–347, 2018.
View at: Publisher Site | Google Scholar
A. Onan, “A fuzzy-rough nearest neighbor classifier combined with consistency-based subset evaluation and instance selection for automated diagnosis of breast cancer,” Expert Systems with Applications, vol. 42, no. 20, pp. 6844–6852, 2015.
View at: Publisher Site | Google Scholar
A. Onan, “Ensemble of classifiers and term weighting schemes for sentiment analysis in Turkish,” Scientific Research Communications, vol. 1, no. 1, 2021.
View at: Publisher Site | Google Scholar
A. Onan and S. Korukoğlu, “A feature selection model based on genetic rank aggregation for text sentiment classification,” Journal of Information Science, vol. 43, no. 1, pp. 25–38, 2017.
View at: Publisher Site | Google Scholar
A. Onan, “Sentiment analysis on product reviews based on weighted word embeddings and deep neural networks,” Concurrency and Computation: Practice and Experience, 2020.
View at: Publisher Site | Google Scholar
A. Onan and M. A. Toçoğlu, “A term weighted neural language model and stacked bidirectional LSTM based framework for sarcasm identification,” IEEE Access, vol. 9, pp. 7701–7722, 2021.
View at: Publisher Site | Google Scholar
A. Onan, “Sentiment analysis on massive open online course evaluations: a text mining and deep learning approach,” Computer Applications in Engineering Education, vol. 29, no. 3, pp. 572–589, 2021.
View at: Publisher Site | Google Scholar
A. Onan and S. Korukoğlu, “Exploring performance of instance selection methods in text sentiment classification,” in Artificial Intelligence Perspectives in Intelligent Systems, Springer, Berlin, Germany, 2016.
View at: Publisher Site | Google Scholar
M. A. Toçoğlu and A. Onan, “Sentiment analysis on students’ evaluation of higher educational institutions,” in International Conference on Intelligent and Fuzzy Systems, Springer, Berlin, Germany, 2020.
View at: Google Scholar
A. Onan, “Topic-enriched word embeddings for sarcasm identification,” in Computer Science On-Line Conference, Springer, Cham, Switzerland, 2019.
View at: Publisher Site | Google Scholar
A. Onan, S. Korukoğlu, and H. Bulut, “Ensemble of keyword extraction methods and classifiers in text classification,” Expert Systems with Applications, vol. 57, pp. 232–247, 2016.
View at: Publisher Site | Google Scholar
A. Onan, “An ensemble scheme based on language function analysis and feature engineering for text genre classification,” Journal of Information Science, vol. 44, no. 1, pp. 28–47, 2018.
View at: Publisher Site | Google Scholar
A. Onan and M. A. Toçoğlu, “Satire identification in Turkish news articles based on ensemble of classifiers,” Turkish Journal of Electrical Engineering and Computer Sciences, vol. 28, no. 2, pp. 1086–1106, 2020.
View at: Publisher Site | Google Scholar
A. Onan, “Hybrid supervised clustering based ensemble scheme for text classification,” Kybernetes, vol. 46, 2017.
View at: Publisher Site | Google Scholar
A. Onan, S. Korukoğlu, and H. Bulut, “A hybrid ensemble pruning approach based on consensus clustering and multi-objective evolutionary algorithm for sentiment classification,” Information Processing & Management, vol. 53, no. 4, pp. 814–833, 2017.
View at: Publisher Site | Google Scholar
P. Pierleoni, L. Palma, A. Belli, and L. Sabbatini, “Using plastic injection moulding machine process parameters for predictive maintenance purposes,” in Proceedings of the International Conference on Intelligent Engineering and Management (ICIEM), pp. 115–120, IEEE, London, UK, June 2020.
View at: Google Scholar
C. Hopmann, S. Jeschke, T. Meisen et al., “Combined learning processes for injection moulding based on simulation and experimental data,” in Proceedings of the The 33rd International Conference of the Polymer Processing Society, vol. 33, 2019.
View at: Publisher Site | Google Scholar
A. Bettoni, E. Montini, M. Righi et al., “Mutualistic and adaptive human-machine collaboration based on machine learning in an injection moulding manufacturing line,” Procedia CIRP, vol. 93, pp. 395–400, 2020.
View at: Publisher Site | Google Scholar
H. Tercan, A. Guajardo, J. Heinisch, T. Thiele, C. Hopmann, and T. Meisen, “Transfer-learning: bridging the gap between real and simulation data for machine learning in injection molding,” Procedia CIRP, vol. 72, pp. 185–190, 2018.
View at: Publisher Site | Google Scholar
A. Tellaeche and R. Arana, “Machine learning algorithms for quality control in plastic molding industry,” in Proceedings of the IEEE 18th Conference on Emerging Technologies & Factory Automation (ETFA), Cagliari, Italy, September 2013.
View at: Publisher Site | Google Scholar
M. Werner, M. Schneider, N. Greifzu et al., “Development of a digital assistance system for continuous quality assurance in the plastic injection moulding process with a focus on self-learning algorithms,” SPE-Inspiring Plastics Professionals, 2020.
View at: Google Scholar
F. Finkeldey, J. Volke, J. C. Zarges, H. P. Heim, and P. Wiederkehr, “Learning quality characteristics for plastic injection molding processes using a combination of simulated and measured data,” Journal of Manufacturing Processes, vol. 60, pp. 134–143, 2020.
View at: Publisher Site | Google Scholar
M. Charest, R. Finn, and R. Dubay, “Integration of artificial intelligence in an injection molding process for online process parameter adjustment,” in Proceedings of the Annual IEEE International Systems Conference (SysCon), Vancouver, BC, Canada, April 2018.
View at: Publisher Site | Google Scholar
Z. P. Fitrianti, K. Bunte, and M. Biehl, “AI. For predictive maintenance for injection moulding machines,” 2020, Doctoral dissertation.
View at: Google Scholar
F. D. Frumosu, G. Ø. Rønsch, and M. Kulahci, “Mould wear-out prediction in the plastic injection moulding industry: a case study,” International Journal of Computer Integrated Manufacturing, vol. 33, no. 12, pp. 1245–1258, 2020.
View at: Publisher Site | Google Scholar
A. V. Hulagadri, “Quality classification of defective parts from injection moulding,” 2020, https://arxiv.org/abs/2008.02872.
View at: Google Scholar
R. Vrabič, D. Kozjek, and P. Butala, “Knowledge elicitation for fault diagnostics in plastic injection moulding: a case for machine-to-machine communication,” CIRP Annals, vol. 66, no. 1, pp. 433–436, 2017.
View at: Publisher Site | Google Scholar
C. Hopmann, A. Ressmann, M. Reiter, S. Stemmler, and D. Abel, “A self-optimising injection moulding process with model-based control system parameterisation,” International Journal of Computer Integrated Manufacturing, vol. 29, no. 11, pp. 1190–1199, 2016.
View at: Publisher Site | Google Scholar
H. S. Park, D. X. Phuong, and S. Kumar, “AI based injection molding process for consistent product quality,” Procedia Manufacturing, vol. 28, pp. 102–106, 2019.
View at: Publisher Site | Google Scholar
W. Shiraishi and K. Yamamoto, “Injection molding system for the calculation of optimum operating,” US Patent and Trademark Office, Washington, DC, USA, 2019, US Patent No. 10416617.
View at: Google Scholar
T. Sabiston, K. Inal, and P. Lee-Sullivan, “Application of Artificial Neural Networks to predict fibre orientation in long fibre compression moulded composite materials,” Composites Science and Technology, vol. 190, Article ID 108034, 2020.
View at: Publisher Site | Google Scholar
N. Costa and B. Ribeiro, “Artificial neural networks for data modelling of a plastic injection moulding process,” in Proceedings of the ICONIP’99. ANZIIS’99 & ANNES’99 & ACNN’99.6th International Conference on Neural Information Processing. Proceedings (Cat. No.99EX378), Perth, Australia, November 1999.
View at: Publisher Site | Google Scholar
R. J. Bensingh, R. Machavaram, S. R. Boopathy, and C. Jebaraj, “Injection molding process optimization of a Bi-aspheric lens using hybrid artificial neural networks (ANNs) and particle swarm optimization (PSO),” Measurement, vol. 134, 2018.
View at: Publisher Site | Google Scholar
W.-C. Chen, P.-H. Tai, M.-W. Wang, W.-J. Deng, and C.-T. Chen, “A neural network-based approach for dynamic quality prediction in a plastic injection molding process,” Expert Systems with Applications, vol. 35, no. 3, pp. 843–849, 2008.
View at: Publisher Site | Google Scholar
R. Manickam, “Back propagation neural network for prediction of some shell moulding parameters,” Periodica Polytechnica - Mechanical Engineering, vol. 60, no. 4, pp. 203–208, 2016.
View at: Publisher Site | Google Scholar
F. Shi, Z. L. Lou, Y. Q. Zhang, and J. G. Lu, “Optimisation of plastic injection moulding process with soft computing,” International Journal of Advanced Manufacturing Technology, vol. 21, no. 9, pp. 656–661, 2003.
View at: Publisher Site | Google Scholar
P. K. Yarlagadda, “Prediction of processing parameters for injection moulding by using a hybrid neural network,” Proceedings of the Institution of Mechanical Engineers - Part B: Journal of Engineering Manufacture, vol. 215, no. 10, pp. 1465–1470, 2001.
View at: Publisher Site | Google Scholar
N. Costa, A. Cunha, and B. Ribeiro, “Monitoring an industrial plastic injection moulding machine using neural networks,” in Artificial Neural Nets and Genetic Algorithms, Springer, Berlin, Germany, 1999.
View at: Publisher Site | Google Scholar
Ç. Karataş, A. Sözen, E. Arcaklioğlu, and S. Ergüney, “Modelling of yield length in the mould of commercial plastics using artificial neural networks,” Materials & Design, vol. 28, no. 1, pp. 278–286, 2007.
View at: Publisher Site | Google Scholar
S. E. Everett and R. Dubay, “A sub-space artificial neural network for mold cooling in injection molding,” Expert Systems with Applications, vol. 79, pp. 358–371, 2017.
View at: Publisher Site | Google Scholar
B. H. M. Sadeghi, “A BP-neural network predictor model for plastic injection molding process,” Journal of Materials Processing Technology, vol. 103, no. 3, pp. 411–416, 2000.
View at: Publisher Site | Google Scholar
A. B. Johnston, L. P. Maguire, and T. M. McGinnity, “Downstream performance prediction for a manufacturing system using neural networks and six-sigma improvement techniques,” Robotics and Computer-Integrated Manufacturing, vol. 25, no. 3, pp. 513–521, 2009.
View at: Publisher Site | Google Scholar
D. Torkar, S. Novak, and F. Novak, “Apparent viscosity prediction of alumina–paraffin suspensions using artificial neural networks,” Journal of Materials Processing Technology, vol. 203, no. 1-3, pp. 208–215, 2008.
View at: Publisher Site | Google Scholar
Y. F. Zhang, J. Y. H. Fuh, and W. T. Chan, “Feature-based cost estimation for packaging products using neural networks,” Computers in Industry, vol. 32, no. 1, pp. 95–113, 1996.
View at: Publisher Site | Google Scholar
E. B. Garvey and B. Eng, Online Quality Control of Injection Molding Using Neural Networks, 1997.
M. A. Sellés, T. Boronat, S. Ferrándiz, and M. A. Peydró, “Using A neural network as A method to improve the quality in the thermoplastic injection moulding process,” AIP Conference Proceedings, vol. 1181, no. 1, pp. 769–774, 2009.
View at: Publisher Site | Google Scholar
S. J. Liao, W. H. Hsieh, J. T. Wang, and Y. C. Su, “Shrinkage and warpage prediction of injection-molded thin-wall parts using artificial neural networks,” Polymer Engineering & Science, vol. 44, no. 11, pp. 2029–2040, 2004.
View at: Publisher Site | Google Scholar
L. N. Smith, R. M. German, and M. L. Smith, “A neural network approach for solution of the inverse problem for selection of powder metallurgy materials,” Journal of Materials Processing Technology, vol. 120, no. 1-3, pp. 419–425, 2002.
View at: Publisher Site | Google Scholar
M. Abas, K. Naeem, T. Habib et al., “Development of prediction model for conductive pattern lines generated through positive displacement microdispensing system using artificial neural network,” Arabian Journal for Science and Engineering, vol. 46, no. 3, pp. 2429–2442, 2021.
View at: Publisher Site | Google Scholar
S. N. Huang, K. K. Tan, and T. H. Lee, “Neural-network-based predictive learning control of ram velocity in injection molding,” IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 34, no. 3, pp. 363–368, 2004.
View at: Publisher Site | Google Scholar
B. Ribeiro, “Fault detection in a thermoplastic injection molding process using neural networks,” in Proceedings of the IJCNN’99. International Joint Conference on Neural Networks. Proceedings (Cat. No.99CH36339), Washington, DC, USA, July 1999.
View at: Publisher Site | Google Scholar
E. Moritzer, E. Müller, Y. Martin, and R. Kleeschulte, “Artificial neural networks to model formulation-property correlations in the process of inline-compounding on an injection moulding machine,” in Proceedings of the The 30th International Conference of the Polymer Processing Society, 2015.
View at: Publisher Site | Google Scholar
H. Zhou and D. Li, “Modelling and prediction of weld line location and properties based on injection moulding simulation,” International Journal of Materials and Product Technology, vol. 21, no. 6, p. 526, 2004.
View at: Publisher Site | Google Scholar
M. Patel and P. Krishna, “A review on application of artificial neural networks for injection moulding and casting processes,” International Journal of Advances in Engineering Sciences, vol. 3, no. 1, pp. 1–12, 2013.
View at: Google Scholar
I. A. Alvarado, A. J. L. García, and B. M. Rodríguez, “Optimization of injection molding process parameters by a hybrid of artificial neural network and artificial bee colony algorithm,” Revista Facultad de Ingeniería Universidad de Antioquia, vol. 67, pp. 43–51, 2013.
View at: Google Scholar
M. Valtonen, A. Lensu, J. Silén, and P. Koikkalainen, “Analyzing and optimizing the injection molding process with a neural network model,” Laboratory of Plastics and Elastomers Technology, Tampere University of Technology, Tampere, Finland, 2000.
View at: Google Scholar
H. T. Su, W. B. Xie, and H. Zeng, “Multiple response optimisation based on the ANN theory of complex injection moulding process,” International Journal of Computer Applications in Technology, vol. 50, no. 3-4, pp. 186–190, 2014.
View at: Publisher Site | Google Scholar
K. C. Ke and M. S. Huang, “Quality prediction for injection molding by using a multilayer perceptron neural network,” Polymers, vol. 12, no. 8, p. 1812, 2020.
View at: Publisher Site | Google Scholar
M. H. Chiang, Y. P. Yeh, F. L. Yang, and Y. N. Chen, “Integrated control of clamping force and energy-saving in hydraulic injection moulding machines using decoupling fuzzy sliding-mode control,” International Journal of Advanced Manufacturing Technology, vol. 27, no. 1-2, pp. 53–62, 2005.
View at: Publisher Site | Google Scholar
W. He, Y. F. Zhang, K. S. Lee, J. Y. H. Fuh, and A. Y. C. Nee, Journal of Intelligent Manufacturing, vol. 9, no. 1, Springer, Berlin, Germany, 1998.
View at: Publisher Site
S. L. Mok, C. K. Kwong, and W. S. Lau, “A hybrid neural network and genetic algorithm approach to the determination of initial process parameters for injection moulding,” International Journal of Advanced Manufacturing Technology, vol. 18, no. 6, pp. 404–409, 2001.
View at: Publisher Site | Google Scholar
O. A. Abdalla, A. O. Elfaki, and Y. M. AlMurtadha, “Optimizing the multilayer feed-forward artificial neural networks architecture and training parameters using genetic algorithm,” International Journal of Computer Application, vol. 96, no. 10, pp. 42–48, 2014.
View at: Publisher Site | Google Scholar
A. Kamoun, M. Jaziri, and M. Chaabouni, “The use of the simplex method and its derivatives to the online optimization of the parameters of an injection moulding process,” Chemometrics and Intelligent Laboratory Systems, vol. 96, no. 2, pp. 117–122, 2009.
View at: Publisher Site | Google Scholar
O. Ahmed, M. Nordin, S. Sulaiman, and W. Fatimah, “Study of genetic algorithm to fully-automate the design and training of artificial neural network,” IJCSNS, vol. 9, no. 1, pp. 217–226, 2009.
View at: Google Scholar
W.-C. Chen, G.-L. Fu, P.-H. Tai, and W.-J. Deng, “Process parameter optimization for MIMO plastic injection molding via soft computing,” Expert Systems with Applications, vol. 36, no. 2, pp. 1114–1122, 2009.
View at: Publisher Site | Google Scholar
P. Nagorny, M. Pillet, E. Pairel et al., “Quality prediction in injection molding,” in Proceedings of the IEEE International Conference on Computational Intelligence and Virtual Environments for Measurement Systems and Applications (CIVEMSA), pp. 141–146, IEEE, Annecy, France, June 2017.
View at: Publisher Site | Google Scholar
C.-H. Lu, C.-C. Tsai, C.-M. Liu, and Y.-H. Charng, “Predictive control based on recurrent neural network and application to plastic injection molding processes,” in Proceedings of the IECON 2007 - 33rd Annual Conference of the IEEE Industrial Electronics Society, Taipei, Taiwan, November 2007.
View at: Publisher Site | Google Scholar
N. Lopes and B. Ribeiro, “A data pre-processing tool for neural networks (dptnn) use in a moulding injection machine,” in Proceedings of the Second World Manufacturing Congress, Durham, England, 1999.
View at: Google Scholar
N. Costa and B. Ribeiro, “A neural prediction model for monitoring and fault diagnosis of a plastic injection moulding process,” in Proceedings of the European Control Conference (ECC), pp. 2381–2385, IEEE, Karlsruhe, Germany, August 1999.
View at: Publisher Site | Google Scholar
P. G. Manjunath and P. Krishna, “Prediction and optimization of dimensional shrinkage variations in injection molded parts using forward and reverse mapping of artificial neural networks,” in Advanced Materials Research, vol. 463, Trans Tech Publications Ltd, Kapellweg, Switzerland, 2012.
View at: Publisher Site | Google Scholar
P. Ganeshan, S. S. Kumaran, K. Raja, and D. Venkateswarlu, “An investigation of mechanical properties of madar fiber reinforced polyester composites for various fiber length and fiber content,” Materials Research Express, vol. 6, 2019.
View at: Publisher Site | Google Scholar
S. Kannan, S. S. Kumaran, and L. A. Kumaraswamidhas, “Optimization of friction welding by taguchi and ANOVA method on commercial aluminium tube to Al 2025 tube plate with backing block using an external tool,” Journal of Mechanical Science and Technology, vol. 30, pp. 2225–2235, 2016.
View at: Publisher Site | Google Scholar
K. S. Senthil, S. Muthukumaran, and R. C. Chandrasekhar, “Effect of tube preparations on joint strength in friction welding of tube-to-tube plate using an external tool process,” Experimental Techniques, vol. 37, pp. 24–32, 2013.
View at: Publisher Site | Google Scholar
V. V. Kumar and S. S. Kumaran, “Friction material composite: types of brake friction material formulations and effects of various ingredients on brake performance-a review,” Materials Research Express, vol. 6, 2019.
View at: Publisher Site | Google Scholar
K. S. Senthil, S. Muthukumaran, D. Venkateswarlu, G. K. Balaji, and S. Vinodh, “Eco-friendly aspects associated with friction welding of tube-to-tube plate using an external tool process,” International Journal of Sustainable Engineering, vol. 5, pp. 120–127, 2012.
View at: Publisher Site | Google Scholar
K. S. Senthil and D. A. Daniel, “Friction welding joints of SA 213 tube to SA 387 tube plate boiler grade materials by using clearance and interference fit method,” Materials Today Proceedings, vol. 5, pp. 8557–8566, 2018.
View at: Publisher Site | Google Scholar
S. I. Butt, U. Asgher, U. Mushtaq et al., “Intelligent machine vision based modeling and positioning system in sand casting process,” Advances in Materials Science and Engineering, vol. 2017, Article ID 3192672, 11 pages, 2017.
View at: Publisher Site | Google Scholar
A. Eser, E. A. Ayyıldız, M. Ayyıldız, and F. Kara, “Artificial intelligence-based surface roughness estimation modelling for milling of AA6061 alloy,” Advances in Materials Science and Engineering, vol. 2021, Article ID 5576600, 10 pages, 2021.
View at: Publisher Site | Google Scholar
E. A. Ayyıldız, M. Ayyıldız, and F. Kara, “Optimization of surface roughness in drilling medium-density fiberboard with a parallel robot,” Advances in Materials Science and Engineering, vol. 2021, Article ID 6658968, 8 pages, 2021.
View at: Publisher Site | Google Scholar
S. K. Selvaraj, A. Raj, M. Dharnidharka et al., “A cutting-edge survey of tribological behavior evaluation using artificial and computational intelligence models,” Advances in Materials Science and Engineering, vol. 2021, Article ID 9529199, 17 pages, 2021.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Senthil Kumaran Selvaraj et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

6350

Downloads

2559

Citations