Detecting anomalies in the injection molding process remains a challenging task, demanding significant resources, data, and expertise due to their impact on cost and time reduction. While traditional methods like statistical process control (SPC) using control charts are widely used for detecting irregularities, they can catch predefined patterns such as systematic, upward shift, downward shift, cyclic, and mixture patterns. However, they still have limitations in identifying anomalies beyond theses common patterns. Numerous unnatural patterns may exist in process data, indicating that the process is out of control. In our study, we propose an innovative strategy to enhance anomaly detection by integrating Statistical Process Control (SPC) with a Long Short-Term Memory (LSTM) based Autoencoder (AE). The main objective is to detect variations in the Melt cushion parameter, a crucial aspect of the injection molding process. The LSTM-AE model determines optimal threshold levels based on reconstruction loss rates across all time-series sequences, complementing traditional control charts’ upper and lower limits. With a model achieving a coefficient of determination (R-squared) of 0.993 and a Mean Absolute Error of 0.0146 through training, and incorporating multiple process limits—Upper Limit Control (ULC), Lower Limit Control (LLC), Maximum Threshold, and Minimum Threshold—the enhanced control chart exhibits significant advancements in anomaly detection. In four distinct scenarios, the integrated model demonstrates its capability to detect anomalies in the Melt cushion that exceed predefined limits, as well as anomalies showing ascending and descending trends. This ultimately enhances the robustness and efficiency of anomaly detection in injection molding processes.
13 April 2024 in Progress in Artificial Intelligence
Improving operational processes significantly relies on choosing the right suppliers, which emphasizes the importance of using effective evaluation methods. This study explores supplier evaluation through a detailed framework involving five Multi-Criteria Decision-Making (MCDM) approaches. These methods consider seven main criteria and twenty-one sub-criteria. In this research, the determination of weights for each criterion and sub-criterion is carried out through a multifaceted approach. Subjective methods, such as the Analytic Hierarchy Process (AHP) and Fuzzy AHP, which rely on expert judgments, are employed alongside objective techniques including Entropy, which is based on empirical data. Hybrid methodologies like AHP-Entropy and Fuzzy AHP-Entropy, which combine elements of both subjective and objective methods, are also utilized. The integration of these diverse methodologies ensures a thorough assessment of supplier performance. Furthermore, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is utilized to rank suppliers according to the weighted criteria. In order to validate the reliability and robustness of the employed methodologies, a comprehensive sensitivity analysis consisting of 21 distinct scenarios is conducted. This exhaustive examination allows for a thorough exploration of how variations in different parameters and conditions may impact the outcomes of the supplier evaluation process. The findings indicate that hybrid methods exhibit stability across nineteen scenarios, while objective methods demonstrate stability in three scenarios, and subjective methods in none. This underscores the importance of integrating both subjective and objective methods in the decision-making process.
26 February 2025 in Engineering Applications of Artificial Intelligence
In injection molding manufacturing, selection of the optimal machine from various alternatives is a crucial strategy for enhancing productivity, cost-effectiveness, and maintaining performance standards. This article presents an approach that combines two techniques to make the best choice from three presented options. Firstly, it employs the Analytic Hierarchy Process method to determine the weights of five main criteria and eleven sub-criteria, considering both cost and performance. Secondly, it utilizes the Technique for Order Preference by Similarity to Ideal Solution to rank the three machines by comparing each machine's performance against ideal and anti-ideal solutions to determine their relative suitability. The final model is validated through three distinct scenarios, illustrating how key criteria such as cost breakdown and scrap rate can influence the ultimate selection ranking. Through a presented numerical example, the paper provides decision-makers with a scientifically robust decision support system, aiding in strategic and complex decision-making processes for selecting the most suitable machine.