Professor, Department of Industrial Engineering and Director TEES Institute for Manufacturing Systems, Texas A&M University
Towards Smart and Sustainable Manufacturing: Pushing AI beyond Machine Learning
Distinguished Engineer & Chief Technologist: Supply Chain Optimization, Sustainability and Protection, IBM
IBM and Environmental Sustainability: Making it Part of Our DNA
Professor, Department of Manufacturing and Industrial Engineering, University of Peradeniya, Sri Lanka
Decision Support Systems for Eco Design
Abstract: It has become a common practice to use Lifecycle Assessment (LCA)-based eco design tools to identify and address environmental hotspots. However, a majority of these efforts depend on simplified LCA or LCAs conducted based on global average lifecycle inventories (LCI) from available databases. This leads to inaccurate identification and quantification of the environmental hotspots across the life cycle. A framework is proposed to overcome this gap and conduct comprehensive LCAs based on process data collected manually and identifying life cycle impacts by collected localized or semi localized data through Internet of Things (IoT) capabilities. This presentation will introduce a decision support tool, validated with industry applications, that integrates comprehensive life cycle information different sources to identify the appropriate interventions in the way of eco design options, or eco innovations.
Chief Technology Officer, Clean Energy Smart Manufacturing Innovation Institute
Democratizing Smart Manufacturing
Leader, Life Cycle Engineering Group, National Institute of Standards and Technology
Directions for Sustainable Manufacturing Standards: Towards the Circular Economy
The global movement towards a circular economy holds many hopes for the future by addressing an unsustainable demand for global resources. To make this move manufacturing practices will be impacted from product design, production, and resurrection of materials. Together with ASTM International NIST is developing a roadmap of the standards that will help to facilitate changes in manufacturing practices leading to more resource and energy efficiency and better management materials, extending their use into multiple product lifecycles. This talk will highlight early standards from ASTM for improving the environmental efficiency of manufacturing processes and discuss the future roadmap for standards to support materials management in the circular economy.
David C. Rummler
Managing Director, CleanTech Energy
Making a Business Case for Sustainable Manufacturing
Assistant Professor, University of Idaho
Sustainable Aquaculture Engineering: Novel Remediation of Water Contamination from Fish Farms
The study aims to develop novel technologies to remediate aquaculture-generated water contamination. The idea is to simultaneously increase food production capacity and improve water quality, enabling future expansion of aquaculture. The primary objectives are (1) identification and testing of novel biomaterials engineered to remove water contaminants, (2) evaluation of the impact of these biomaterials on downstream water systems, and (3) optimization of methods to utilize these materials in commercial applications. The motivation is clear: 90% of seafood consumed in the United States (US) is imported from countries whose regulatory frameworks do not meet US standards, and the US has an annual seafood trade deficit of nearly $17 billion. Aquaculture is the fastest-growing food producing sector. Idaho is in the top ten states for aquaculture and value, producing over 70% of all rainbow trout that are raised in the US; however, their ability to continue this level of production could be in jeopardy by exceeding State and Federal regulatory limits for pollution. The results of laboratory research indicate that our engineered biomaterials from different feedstocks are able to remove phosphorus from the polluted waters, especially lodge pine, with a removal rate of 131% in a couple of minutes. The results of our field experiments show that the small particle size biomaterial from lodge pine has the highest total phosphorus removal rate.
Assistant Professor, University of Michigan
Creating rapid, transparent, and updateable environmental impact models of manufacturing systems
For many manufacturing systems, such as emerging process technologies or distributed supply chains, the data with which to form environmental impacts models are often sparse and even contradictory. Despite this, the environmental impact modeling of industrial systems often lacks a rigorous uncertainty quantification (UQ)—a lack of UQ limits insight into the impacts, risks and unintended consequences of system interventions. Making environmentally motivated decisions and policies without uncertainty and confidence measures may lead to either reduced or even negative environmental benefits. On the other hand, inclusion of UQ would provide a mathematically principled procedure to incorporate sparse, noisy, and incomplete data into the impact models—a data-informed model learning approach. This model-and-data relationship can then be leveraged to create intelligent data acquisition strategies for seeking out the most informative data that can tell us what the model structure should look like, and what its parameters values are. These important unmet challenges must be addressed systematically and rigorously before the usefulness and credibility of environmental impact modeling can be completely realized, and for it to be fully adopted into practice. In this talk, I will present some of my group’s early work on developing and applying Bayesian inference techniques to rapidly generate and select candidate model structures and transparently communicate and update the parametric uncertainties as newly collected data are included. These techniques are demonstrated for the case of an emerging unit manufacturing process (laser powder bed fusion) and material supply chain (annual flow of steel in the United States). These methods show great promise in generating transparent environmental impact model uncertainties, providing a more robust basis for environmentally motivated decision-making.
Assistant Professor, University of South Florida
Smart Manufacturing to Enable Sustainability in the Post-COVID-19 Era
Abstract: The unique and unprecedented challenges of the COVID-19 pandemic have resulted in significant disruptions to manufacturing systems and diverse supply chains around the world. High variability in the demand for products and shifts in the manufacturing capabilities needed within already complex production systems have imposed many challenges to maintaining sustainability within the sector. Nonetheless, smart manufacturing technologies have the potential to overcome several challenges as they can offer a multitude of benefits such as the improved visibility and performance of production operations. In this presentation, a multi-criteria decision-making analysis for smart and sustainable machining will be presented. Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) is used to compare alternative manufacturing scenarios with varied technological capabilities. Tool Condition Monitoring (TCM) for CNC machining and the modeling of ambient conditions in the manufacturing environment via Computational Fluid Dynamics (CFD) are included in the MCDA. The case study shows a preference for the application of smart manufacturing capabilities relative to conventional machining when accounting for sustainability and performance indicators. Ultimately, a systematic integration of sustainable manufacturing principles and metrics into the business practices of manufacturing enterprises is envisioned. Realizing the vision for sustainable and resilient manufacturing will inevitably require greater availability and transparency of key production data, to create a future in which our response to emergencies can be realized swiftly and efficiently.
Assistant Professor, University of Kentucky
Recent Progress Towards Digitally-enabled and Physics-informed Modeling and Optimization of Finish Machining Processes
Abstract: In light of recent trends towards near-net-shape manufacturing via precision casting, extrusion and additive manufacturing, high performance finishing is becoming increasingly important. Moreover, functional performance of many engineered components such as turbine blades and biomedical implants is significantly impacted by the surface and sub-surface characteristics induced by finishing processes. However, currently available modeling approaches of surface integrity either lack predictive power or require tremendous computational time. As a result, the manufacturing industry has largely adopted an empirical approach, and designers do not leverage finishing processes for ‘pro-active surface engineering’. Moreover, there are currently no provisions to integrate critical surface integrity parameters, such as sub-surface residual stresses and micro-hardness, in the emerging Digital Thread. This talk will present recent developments on a novel integrated approach of high-speed in-situ characterization and semi-analytical modeling that is envisioned to finally enable real-time modeling of process-induced surface integrity evolution. Using the in-situ characterization testbed developed at the University of Kentucky, previously unavailable workpiece material and process-specific properties can be efficiently captured. This novel characterization approach is enabling the use of computationally-efficient, physics-based analytical models, which are capable of real-time predictive modeling of process-induced surface integrity. Preliminary experimental results, as well as relevant implications for product and process design will be discussed. Finally, ongoing and future work towards implementing real-time modeling and control of process-induced surface integrity within the Digital Thread for closed-loop adaptive machining strategies will be presented.
Assistant Professor, University of Kentucky
Hybrid Machine Learning for Adaptive Robotic Welding Control
Abstract: The first commercial use of robotics in manufacturing was for welding operations. Since then, industrial robots have become more diverse and common for automating manufacturing processes. Even though robotic design and control have rapidly progressed in the past half-century, many welding operations are still done manually. The skills needed for complex welding scenarios possessed by trained professionals have yet to be mimicked by current industrial robots. Existing robotic control is incapable of adaptively adjusting its robotic operation in response to a dynamic welding environment, whereas a skilled human welder can. To enable sophisticated and adaptive robotic control toward improved manufacturing and production efficiency, three elements are needed: perception, prediction, and reaction. Accurate prediction and real-time reaction rely on the effective and efficient processing of perception data and characterization of this highly dynamic system. Emerging machine learning and deep learning techniques have the potential to realize adaptive robotic control mirroring human capabilities. This presentation presents a preliminary study on developing a hybrid Machine Learning (ML) framework for real-time welding quality prediction and adaptive welding speed adjustment for GTAW welding at a constant current. The hybrid ML framework includes three elements: Convolutional Neural Network (CNN)-based welding quality prediction, Multiple Layer Perceptron (MLP)-based process modeling, and Gradient Descent (GD)-based controller. With the CNN, in-situ imaging of top-side weld pool can be analyzed to predict the back-side bead width during active welding control. With the MLP, the effect of welding speed on bead width can be quantitively modeled. Through the trained MLP, a computationally efficient GD algorithm has been developed to adjust the travel speed accordingly to achieve an optimal bead width with full material penetration. Because of the nature of gradient descent, the robot would change faster when the quality is further away and then fine-tune the speed when it was close to the goal. Experimental studies have shown promising results on real-time bead width prediction and adaptive speed adjustment to realize ideal bead width.
Assistant Professor, University of Texas at Arlington
Enhancing the Sustainability of Additive Manufacturing in the Context of Industry 4.0
Additive manufacturing plays a critical role in future manufacturing because of its unique characteristics such as higher material usage efficiency, lower production cost and time, enhanced manufacturing complexity and capability, and simplified supply chain structure. In the context of Industry 4.0, the integration of intelligent production processes/systems with advanced data analytics technologies has offered new opportunities to assess and advance the sustainability of additive manufacturing, attributed to the high degree of automation and in-process sensing. These new technologies will bridge the environment between cyber and physical worlds and form a pathway to the next generation of additive manufacturing: sustainable and smart additive manufacturing. This presentation will discuss some exciting opportunities for enhancing the sustainability of additive manufacturing in the context of Industry 4.0 from different perspectives such as monitoring and analyzing sustainability-related measures, joint fabrication quality and sustainability assessment, and recycling/reusing additive manufacturing waste.