The objective of this collaborative research project is to significantly improve observability in polymer processing and characterize its contribution to higher product quality and manufacturing productivity. Multivariate, wireless sensors will be designed and structurally integrated into an injection mold to monitor the dynamic variations in four critical polymer states: pressure, temperature, velocity, and viscosity. An embedded ASIC chip (Application-Specific Integrated Circuit) will be incorporated within the sensor package to coordinate the multi-modal sensing actions, enable adaptive sampling rates to capture the process dynamics in an energy-efficient manner, and control the acoustic-based wireless digital transmission of the four process parameters.
If successful, the research will advance knowledge and understanding of energy harvesting means for remote sensing and the optimal interface of sensors with embedded microelectronics, in a harsh manufacturing environment. The design of the sensors represents the first system integration of piezoelectric and infrared transduction principles with mechanistic analysis, and has the potential to improve molding productivity, wherein each percentage point improvement corresponds to cost savings of over $40 million per year. Besides injection molding, the multivariate sensing platform and new analysis capability resulting from this research can be adapted across a wide array of manufacturing scenarios to improve process control and enhance quality assurance. As a result, various manufacturing operations can be better optimized and automated. The research will also contribute to multidisciplinary education and training of students at the PIs' institutions in digital sensing, multiphysics modeling, and nonlinear optimization methods.
The research investigated the feasibility, design, and utility of an advanced sensor for monitoring the process of plastic injection molding for better product quality and higher productivity. The motivation is to assist manufacturers to better compete by making higher quality products at lower costs. For this purpose, this research enabled a sensor with onboard data-processing and wireless data transmission electronics that can be structurally integrated into an injection mold to measure different parameters during the molding process: melt temperature, melt pressure, melt velocity, and melt viscosity. Through coded acoustic waves, these parameters are wirelessly transmitted through the mold steel and processed on-line. This new multi-modal sensor enabled, for the first time, the simultaneous observation of the effect of different sets of process states and product quality, such as dimensions and strength, which have not been directly measurable within the process of injection molding. The intellectual contributions of this research are the analytical and experimental work on both the sensing principles and data analysis. The research has advanced knowledge and understanding of energy harvesting means for remote sensors and their optimization for use with integrated circuits. The sensor design is creative and demonstrates the first integration of multi-physical effects, such as piezoelectric and infrared, and mechanistic analysis within a single sensor package to provide vital information about multiple states during polymer processing. The explicit modeling of the injection molding process is transformative with respect to providing an instrumentation methodology for use on an application-specific basis in polymer processing. Systematic experiments performed on a production-grade injection molding machine has shown that, using the new sensor and advanced analytics, the average accuracy for predicting the part thickness is better than 93.9%, whereas for part thickness, the prediction accuracy is better than 91.9%. Moreover, the new sensor has consistently outperformed multiple commercial sensors in terms of accuracy and robustness. Such results demonstrate that new sensing method directly contributes to advancing the state-of-the-knowledge about polymer processing. Research outcome in analytical and numerical modeling has led to the publication of over fifteen papers in high impact technical journals and international conferences. One the papers has received the Best Paper Award from the ASME International Symposium on Flexible Automation in 2012. Furthermore, an US patent has been filed, based on the multivariate sensing method that is directly originated from this project. The research has also identified limitations of the new sensor, opening up new possibilities for future research. One of the limitations was the acquisition of process data at very fast melt velocities, e.g., 400 mm/s, when the sensor could not respond fast enough to accurately capture the polymer melt velocity and viscosity. We found this was due to the relative slow response of the thermopile that was used as a temperature sensing element. Since alternative temperature sensors do not provide high fidelity signals at low temperatures, a challenging problem arises that motivates our future research on temperature sensing mechanisms, e.g., through imaging. Considering that the plastics industry is the third largest manufacturing industry in the United States with nearly $374 billion in annual shipments, this research has significant broader impact by directly addressing the bottleneck of in-process sensors and the lack of consensus as to their use for improving polymer processing control. The project has demonstrated both an innovative sensing platform and new analysis capability for improved closed-loop process control and enhanced quality assurance within harsh environments. As a result, challenging manufacturing applications can be better optimized and automated.
