Search for: All records

Abstract Predictive maintenance in truck fleet management is essential to reduce downtime and maintenance costs, yet traditional approaches often rely on static, rule-based schedules that fail to capture real-time operational variability. In this paper, we propose a robust digital twin (DT) framework for predictive maintenance specifically designed for tire predictive maintenance that integrates real-time tire health data, dynamic decision-making, and adaptive model updates to optimize tire resource allocation and enhance system health. Our framework is unique in its ability to incorporate uncertainty-aware dynamic programming, drift detection, and adaptive surrogate model updates within the digital twin. Specifically, we develop an uncertainty-aware dynamic linear programming (U-DLP) approach to optimize tire placement and servicing schedules based on continuously updated tire health data through surrogate model. To ensure DT reliability, we employ the maximum concept discrepancy (MCD) method to detect drift by identifying discrepancies between predicted and actual tire performance, thereby flagging data for necessary tire health model updates. Subsequently, we introduce an uncertainty-aware low-rank adaptation (U-LORA) method to efficiently update the tire health model by dynamically refining the surrogate model based on measured uncertainty. Simulation results indicate that our framework extends tire lifespan by nearly 50% compared to conventional methods, requiring fewer tires to achieve the same operational mileage, while also reducing tire waste and maintenance costs. This integrated digital twin framework offers a reliable and efficient solution for tire predictive maintenance, enhancing fleet sustainability and operational efficiency. 

Abstract This study models the temperature evolution during additive friction stir deposition (AFSD) using machine learning. AFSD is a solid-state additive manufacturing technology that deposits metal using plastic flow without melting. However, the ability to predict its performance using the underlying physics is in the early stage. A physics-informed machine learning approach, AFSD-Nets, is presented here to predict temperature profiles based on the combined effects of heat generation and heat transfer. The proposed AFSD-Nets includes a set of customized neural network approximators, which are used to model the coupled temperature evolution for the tool and build during multi-layer material deposition. Experiments are designed and performed using 7075 aluminum feedstock deposited on a substrate of the same material for 30 layers. A comparison of predictions and measurements shows that the proposed AFSD-Nets approach can accurately describe and predict the temperature evolution during the AFSD process. 

Abstract Skeletal fixation plates are essential components in craniomaxillofacial (CMF) reconstructive surgery to connect skeletal disunions. To ensure that these plates achieve geometric conformity to the CMF skeleton of individual patients, a pre-operative procedure involving manual plate bending is traditionally required. However, manual adjustment of the fixation plate can be time-consuming and is prone to geometric error due to the springback effect and human inspection limitations. This work represents a first step towards autonomous incremental plate bending for CMF reconstructive surgery through machine learning-enabled springback prediction and feedback bending control. Specifically, a Gaussian process is first investigated to complement the physics-based Gardiner equation to improve the accuracy of springback effect estimation, which is then incorporated into nonlinear model predictive controller to determine the optimal sequence of bending inputs to achieve geometric conformity. Evaluation using a simulated environment for bending confirms the effectiveness of the developed approach. 

Abstract We introduce a novel digital twin framework for predictive maintenance of physical systems with long term operations. Using monitoring tire health as an application, we show how the digital twin framework is used to enhance automotive safety and efficiency, while overcoming technical challenges using a three-step approach. Firstly, for managing the data complexity over a long operation span, we employ data reduction techniques to concisely represent physical tires using historical performance and usage data. Relying on this data, for fast real-time prediction, we train a transformer-based model offline on our concise dataset to predict future tire health over time, represented as Remaining Casing Potential (RCP). Based on our architecture, our model quantifies both epistemic and aleatoric uncertainty, providing reliable confidence intervals around predicted RCP. Secondly, we incorporate real-time data by updating the predictive model in the digital twin framework, ensuring its accuracy throughout its life span with the aid of hybrid modeling and the use of a discrepancy function. Thirdly, to assist decision making in predictive maintenance, we implement a Tire State Decision Algorithm, which strategically determines the optimal timing for tire replacement based on RCP forecasted by our transformer model. This three-step approach ensures that our digital twin not only accurately predicts the health of a system, but also continually refines its digital representation and makes predictive maintenance decisions for removal from service. Our proposed digital twin framework embodies a physical system accurately and leverages big data and machine learning for predictive maintenance, model update and decision making. 

Abstract Environment-induced cracking (EIC) research spanning the last 80 years for ferrous and non-ferrous metals in aqueous environments at ambient and elevated temperatures has concentrated on crack propagation. Studies clearly reveal EIC involves two differentiable processes, one controlling initiation and the other propagation. Utilization of advanced high-resolution electron microscopy over the last 20 years has enabled more focused studies of crack initiation for stainless steel and nickel-based alloys at elevated temperatures exposed to environments associated with the nuclear industry. More recently, when coupled with advancedin-situexperimental techniques such as time-lapse X-ray computed 3D-tomography, progress has also been made for aluminum alloys suffering EIC at ambient temperatures. Conventional wisdom states that chemical processes are typically rate-controlling during EIC initiation. Additionally, experimental evidence based on primary creep exhaustion ahead of the introduction of an aggressive environment indicates that time-dependent mechanically-driven local microstructural strain accommodation processes (resembling creep-like behavior) often play an important role for many metals, even for temperatures as low as 40 % of their melting points (0.4 T_m). EIC studies reveal initial surface conditions and their associated immediate sub-surface alloy microstructures generated during creation (i.e. disturbed layers) can dictate whether or not EIC initiation occurs under mechanical loading conditions otherwise sufficient to enable initiation and growth. The plethora of quantitative experimental techniques now available to researchers should enable significant advances towards understanding EIC initiation. 

Abstract Simultaneous Localization and Mapping (SLAM) is an autonomous localization technique used for mobile robots without GPS. Since autonomous localization relies on pre-existing maps, to use SLAM with the Robotic Operating System (ROS), a map of the surroundings must first be created, and a controller can then use the initial map. The first mapping procedure is mostly carried out manually, with human intervention. When operating manually, the person operating the robot is responsible for avoiding obstacles and moving the robot to different sections of the space to create a full map of the entire environment. The mapping process, if done manually, is time demanding, and often not feasible. To solve this constraint, which is to construct a map of the environment autonomously without human involvement while avoiding obstacles, the Vector Field Histogram (VFH) technique is implemented in this study by integrating it with SLAM. VFH is a real-time motion planning approach in robotics that uses a statistical representation of the robot’s surroundings known as the histogram grid, to place a strong emphasis on handling modeling errors and sensor uncertainty. Furthermore, using range sensor values, the VFH algorithm determines a robot’s obstacle-free driving directions. Aside from its real-time obstacle avoidance function, the VFH method is enhanced in this study to collaborate with SLAM to create maps and reduce localization complexity. While generating maps, the VFH approach uses a two-step data-reduction procedure to calculate the appropriate vehicle control directives. The robot’s temporary location is used to generate a one-dimensional polar histogram, which is the first stage of the histogram grid reduction process. The polar obstacle density in a given direction is represented by a value in each sector of the polar histogram. In the second stage, the robot’s steering is oriented in the direction of the most appropriate sector, which the algorithm determines from all the polar histogram sectors with a low polar obstacle density. Following that, further algorithms, such as Rapidly Exploring Random Tree (RRT) and A*, can be used to plan autonomous pathways using the map provided by VFH. In order to put the concept into practice, MATLAB and ROS are used together in collaboration to autonomously and simultaneously map the environment and localize the robot. The combination of MATLAB and ROS provides many advantages because of their extensive feature set and ability to integrate with each other. Finally, a simulation and a real-time robot are utilized to analyze and validate the study’s findings. 

Abstract Over the last few decades, globalization has weakened the US manufacturing sector. The COVID-19 pandemic revealed import dependencies and supply chain shocks that have raised public and private awareness of the need to rebuild domestic production. A range of new technologies, collectively called Industry 4.0, create opportunities to revolutionize domestic and local manufacturing. Success depends on further refinement of those technologies, broad implementation throughout private companies, and concerted efforts to rebuild the industrial commons, the national ecosystem of producers, suppliers, service providers, educators, and workforce necessary to regain a competitive, innovative manufacturing sector. A recent workshop sponsored by the Engineering Research Visioning Alliance (ERVA) identified a range of challenges and opportunities to build a resilient, flexible, scalable, and high-quality manufacturing sector. This paper provides a strategic roadmap for regaining US manufacturing leadership by briefly summarizing discussions at the ERVA-sponsored workshop held in 2023 and providing additional analysis of key technical and economic issues that must be addressed to achieve dynamic, high-value manufacturing in the USA. The focus of this presentation is on discrete manufacturing of production of structural components, a large subset of total manufacturing that produces high-value inputs and finished products for domestic consumption and export. 

Abstract This paper details the design and operation of a testbed to evaluate the concept of autonomous manufacturing to achieve a desired manufactured part performance specification. This testbed, the autonomous manufacturing system for phononic crystals (AMSPnC), is composed of additive manufacturing, material transport, ultrasonic testing, and cognition subsystems. Critically, the AMSPnC exhibits common manufacturing deficiencies such as process operating window limits, process uncertainty, and probabilistic failure. A case study illustrates the AMSPnC function using a standard supervised learning model trained by printing and testing an array of 48 unique designs that span the allowable design space. Using this model, three separate performance specifications are defined and an optimization algorithm is applied to autonomously select three corresponding design sets to achieve the specified performance. Validation manufacturing and testing confirms that two of the three optimal designs, as defined by an objective function, achieve the desired performance, with the third being outside the design window in which a distinct bandpass is achieved in phononic crystals (PnCs). Furthermore, across all samples, there is a marked difference between the observed bandpass characteristics and predictions from finite elements method computation, highlighting the importance of autonomous manufacturing for complex manufacturing objectives. 

Abstract In contemporary manufacturing processes, reliable but efficient pick-and-place robots are frequently used. The automation and optimization of the pick and place procedures utilizing various path-planning approaches thereby support the expansion of application areas. Yet, the design of a controller faces significant difficulties due to the nonlinearities inherent in robotic manipulators and the unpredictable nature of the ambient factors. In place of the classic model predictive control (MPC), this paper presents the application of the Nonlinear Model Predictive Controller (NLMPC) as an acceptable control mechanism for real-time optimization and robust stability of the KINOVA Gen3 robotic arm. The developed NLMPC-based method ensures that the robotic arm does not run into obstacles in the workplace or with itself while reaching, gripping, selecting, and placing the necessary items. To acquire the control input trajectory, the optimization in NLMPC is solved repeatedly. When input constraints are available, the modeled system tracks reference trajectories to achieve the aim of recognizing and organizing distinct objects. After the NLMPC is successfully developed, a simulation environment is built and finally brought to life by combining all the processes into one using a MATLAB Stateflow chart. 

Abstract Highlights from a member survey and panel discussion outlining the challenges and benefits of advanced manufacturing for the materials community. The article covers research being conducted in academia and industry to develop and integrate advanced manufacturing techniques; methods for in-process monitoring and machine learning; and challenges in adopting new digital manufacturing technologies.