Machine Learning in Injection Molding: An Industry 4.0 Method of Quality Prediction

Párizs, Richárd Dominik; Török, Dániel; Ageyeva, Tatyana; Kovács, József

doi:10.3390/s22072704

Cited by 36 publications

(12 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, ML models can also be utilized to predict the quality of polymers, for example, of acrylonitrile butadiene styrene parts produced via injection molding. [ 159 ] The input features used in the models were customized integrals over the cavity pressure curves and the generated parts were classified as “undercompensated,” “acceptable,” or “overcompensated.” Accuracies of about 90% were achieved using decision tree‐based models.…”

Section: Processing and Fabricationmentioning

confidence: 99%

Application of Digital Methods in Polymer Science and Engineering

Schuett,

Endres,

Standau

et al. 2023

Adv Funct Materials

View full text Add to dashboard Cite

The development of new polymer materials is an emerging field for more than 100 years. However, it is currently facing major challenges and the application of digital methods can help to develop new processes, discover new materials and, thus, contribute to the challenges of today and the future. Though, the application of digital methods in the field of polymer science is currently very limited, when compared to other classes of materials such as small molecules or inorganic high‐performance materials. Nevertheless, there are already first, very promising approaches. The current review article focuses on the application of these methods in different aspects of polymer research including polymer design, synthesis, and characterization. Furthermore, the digital discovery of polymer engineering is highlighted in detail showing the broad range of potential applications of these methods in polymer science. Finally, future possibilities and opportunities are derived from the current state‐of‐the‐art and perspectives for a potential evolution of polymer science are provided.

show abstract

Section: Processing and Fabricationmentioning

confidence: 99%

Application of Digital Methods in Polymer Science and Engineering

Schuett,

Endres,

Standau

et al. 2023

Adv Funct Materials

View full text Add to dashboard Cite

show abstract

“…Treebased algorithms, regression-based algorithms, or neural networks are applied to perform classifications or regressions [13] [14]. An interesting approach is described in [15], where the use of sensor data for prediction models is outlined. The authors identified the cavity pressure as essential for quality, which is why sensor data provide important information.…”

Section: Introduction and Motiv A T Ionmentioning

confidence: 99%

Application of Machine Learning Methods for Process Optimization in Electronic Packaging Processes

Niegisch,

Haag,

Braun

et al. 2024

IMAPSource Proceedings

View full text Add to dashboard Cite

Epoxy resins are commonly used as encapsulation materials in electronic packaging processes. Fluctuations in the materials lead to both changes in processability and to varying quality. Ideally, these variations should be identified, and measures taken as quickly as possible to reduce scrap parts and thus costs. A promising optimization approach for encapsulation processes are machine learning models, which have already shown good results in quality predictions, especially for injection molding. Subsequent quality measurements are avoided with good quality prediction models. With this type of models, not only predictions can be made, but also optimal parameter combinations can be found. In this paper, models for predicting quality criteria warpage and residual enthalpy of the epoxy molding compound were set up, trained and validated. Time series of in-situ sensors were used, from which relevant features were extracted and which, together with machine parameters, provide a dataset for prediction. The most promising prediction models are random forest regression and gradient boosting regression. T hey predict warpage with an accuracy of 90% to 91% and the residual enthalpy with an accuracy of 95%. Subsequently, optimization models of the machine parameters were set up. All relevant target variables were considered in a cost function, through the minimization of which an optimal parameter set was found. The gradient boosted tree and Bayesian optimization were determined to be the most promising models, as they lead to the lowest values of the respective cost function.

show abstract

“…Therefore, finding a relation between product quality and process parameters is a non-trivial task that requires a comprehensive approach, which often involves a combination of sensoring and information techniques [ 6 ]. Statistical analysis [ 7 , 8 ] and artificial intelligence (AI) [ 9 ], especially machine learning (ML) [ 10 , 11 , 12 ], are used more and more frequently today for the process optimization and quality control of industrial manufacturing processes. However, these methods are data-driven and often do not consider the physical aspects of injection molding.…”

Section: Introductionmentioning

confidence: 99%

Multiple In-Mold Sensors for Quality and Process Control in Injection Molding

Párizs

Török

Ageyeva

et al. 2023

Sensors

Self Cite

View full text Add to dashboard Cite

The simultaneous improvement of injection molding process efficiency and product quality, as required by Industry 4.0, is a complex, non-trivial task that requires a comprehensive approach, which involves a combination of sensoring and information techniques. In this study, we investigated the suitability of in-mold pressure sensors to control the injection molding process in multi-cavity molds. We have conducted several experiments to show how to optimize the clamping force, switchover, or holding time by measuring only pressure in a multi-cavity mold. The results show that the pressure curves and the pressure integral are suitable for determining optimal clamping force. We also proved that in-channel sensors could be effectively used for a pressure-controlled SWOP. In the volume-controlled method, only the sensors in the cavity were capable of correctly detecting the end of the filling. We proposed a method to optimize the holding phase. In this method, we first determined the integration time of the area under the pressure curve and then performed a model fit using the relationship between the pressure integral and product mass. The saturation curve fitted to the pressure data can easily determine the gate freeze-off time from pressure measurements.

show abstract

Machine Learning in Injection Molding: An Industry 4.0 Method of Quality Prediction

Cited by 36 publications

References 38 publications

Application of Digital Methods in Polymer Science and Engineering

Application of Digital Methods in Polymer Science and Engineering

Application of Machine Learning Methods for Process Optimization in Electronic Packaging Processes

Multiple In-Mold Sensors for Quality and Process Control in Injection Molding

Contact Info

Product

Resources

About