Prediction of Hotspots in Injection Moulding - Using Simulation, In-mould Sensors, and Machine Learning

Aug 4, 2026

Upload by Asher

Injection moulding relies on tightly controlled parameters, yet non-optimised process variables and cooling limitations create hotspot regions that can cause warpage or shrinkage. This study presents a machine-learning framework that predicts the maximum hotspot temperature in an injection-moulded component using process simulation and in-mould sensor data. Hotspot locations and temperatures are verified via in-mould thermocouples, while ANN and SVR models are trained using DOE-extracted data. Gaussian SVR outperforms linear SVR, and ANN shows slightly better overall prediction.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages6

Reading Time15 min

AccessFree to download

Introduction

Injection moulding process and hotspot causes

Related work on ML in injection moulding

Gap: temperature-profile and hotspot prediction

Machine-learning-based prediction approach

Simulation-based hotspot identification

Sensor-based experimental confirmation

ANN and SVR model development and kernel comparison

Prediction of Hotspots in Injection moulding by Using Simulation, In-mould Sensors, and Machine Learning* Mandana Kariminejad 1 , David Tormey2 , Christopher O’Hara2 and Marion McAfee 1 Abstract—Injection moulding is an industrial process for the mass production of plastic components, with many parameters affecting the quality of this process. Hotspot regions in the component occur due to non-optimised process variables or limitations in the cooling system and can lead to warpage or shrinkage. Hotspots should be minimised to avoid part defects and achieve the required dimensional tolerances for precision components. This work outlines a machine-learning-based approach for predicting the maximum hotspot temperature inan injection moulded component using process simulation and in-mould sensor data. The hotspots were identified through software simulation, and then their locations and temperatures were confirmed through an actual experiment using in-mould thermocouples. Two different machine learning approaches, artificial neural network (ANN) and support vector regression (SVR), were developed using the extracted data from the sensors and a design of experiment (DOE) method. The performance of linear and Gaussian kernels was compared for the SVR method. The Gaussian SVR resulted in superior performance compared to the linear kernel. The Gaussian SVR was then compared to the ANN prediction method, where ANN showed a slightly better prediction performance. This study has two primary outcomes. First, we show the simulation results can be used to identify critical areas of the part for real-time monitoring. Secondly, embedding sensors in these locations and applying a machine learning approach to the data, provides a good indication of potential quality issues such as warpage and shrinkage post-production. The use of ANN indicates an accurate prediction performance, facilitating rapid optimisation of the process for the minimisation of hotspots. I. INTRODUCTION Injection moulding (IM) is a widely used process for the rapid manufacturing of plastic components in high volumes. The process contains three main stages: filling, packing, and cooling. Uneven and non-uniform temperature distribution on the cooling of the part can lead to residual stresses and, thereby, part defects such as warpage and shrinkage. Identifying and eliminating the locally heated regions or hotspots are crucial to having a uniform temperature profile and high part quality. Hence, predicting these hotspots and *This research is supported by an ATU Sligo Bursary and also by a research grant from Science Foundation Ireland (SFI) under Grant Number 16/RC/3872 and is co-funded under the European Regional Development Fund and by I-Form industry partners. 1Mandana Kariminejad & Marion McAfee are with Centre for Precision Engineering, Materials and Manufacturing (PEM Centre) & Centre for Mathematical Modelling and Intelligent Systems for Health and Environment (MISHE), Atlantic Technological University, Ash Lane, F91 YW50 Sligo, [Ireland.](Ireland. mandana.kariminejad@research.atu.ie)[ mandana.kariminejad@research.atu.ie](Ireland. mandana.kariminejad@research.atu.ie) , [marion.mcafee@atu.ie](marion.mcafee@atu.ie) 2David Tormey & Christopher O’Hara are with Centre for Precision Engineering, Materials and Manufacturing (PEM Centre), Atlantic Technological University, Ash Lane, F91 YW50 Sligo, Ireland. [david.tormey@atu.ie](david.tormey@atu.ie) , [christopher.ohara@atu.ie](christopher.ohara@atu.ie) mapping their relationship with the key input parameters is a proposed novel approach to improve the part quality. Different machine learning (ML) algorithms have been applied in the injection moulding (IM) process to predict the part quality factors and optimal process variables. Artificial Neural Network (ANN) is one of these well-developed methods that has been shown to be effective for modelling the IM process in a number of recent works. Bensingh et al.

FAQs

Why are hotspots important in injection moulding?

Hotspots form due to non-optimised process variables or cooling limitations and can lead to part defects such as warpage and shrinkage. Minimising them helps meet required dimensional tolerances.

How does the study predict maximum hotspot temperature?

It combines process simulation with in-mould sensor data to identify hotspot locations and then predicts the maximum hotspot temperature using machine learning models trained on DOE-extracted features.

Which machine learning approach performed best in the comparison?

For SVR, the Gaussian kernel delivered superior performance compared with the linear kernel. The ANN method also achieved slightly better prediction performance than Gaussian SVR in the reported comparison.

Aug 4, 2026

Upload by Asher