Sequence-Dependent Robotic Disassembly Line Balancing Problem Considering Disassembly Path

For the realization of environmental protection and resource conservation, remanufacturing is of great significance. Disassembly is a key step in remanufacturing, the disassembly line system is the main scenario for product disassembly, usually consisting of multiple workstations, and has prolific productivity. The application of the robots in the disassembly line will eliminate various problems caused by manual disassembly. Moreover, the disassembly line balancing problem (DLBP) is of great importance for environmental remanufacturing. In the past, disassembly work was usually done manually with high cost and relatively low efficiency. Therefore, more and more researches are focusing on the automatic DLBP due to its high efficiency. This research solves the sequence-dependent robotic disassembly line balancing problem (SDRDLBP) with multiple objectives. It considers the sequence-dependent time increments and requires the generated feasible disassembly sequence to be assigned to ordered disassembly workstations according to the specific robotic workstation assignment method. In robotic DLBP, due to the special characteristics of robotic disassembly, we need to consider the moving time of the robots’ disassembly path during the disassembly process. This is also the first time to consider sequence-dependent time increments while considering the disassembly path of the robots. Then with the help of crossover and mutation operators, multi-objective evolutionary algorithms (MOEAs) are proposed to solve SDRDLBP. Based on the gear pump model, the performance of the used algorithm under different cycle times is analyzed and compared with another two algorithms. The average values of the HV and IGD indicators have been calculated, respectively. The results show the NSGA-II algorithm presents outstanding performance among the three MOEAs, and hence demonstrate the superiority of the NSGA-II algorithm.

A Novel Real-Time Thermal Analysis and Layer Time Control Framework for Large Scale Additive Manufacturing

Production efficiency and product quality need to be addressed simultaneously to ensure the reliability of large scale additive manufacturing. Specifically, print surface temperature plays a critical role in determining the quality characteristics of the product. Moreover, heat transfer via conduction as a result of spatial correlation between locations on the surface of large and complex geometries necessitates the employment of more robust methodologies to extract and monitor the data. In this paper, we propose a framework for real-time data extraction from thermal images as well as a novel method for controlling layer time during the printing process. A FLIR™ thermal camera captures and stores the stream of images from the print surface temperature while the Thermwood Large Scale Additive Manufacturing (LSAM™) machine is printing components. A set of digital image processing tasks were performed to extract the thermal data. Separate regression models based on real-time thermal imaging data are built on each location on the surface to predict the associated temperatures. Subsequently, a control method is proposed to find the best time for printing the next layer given the predictions. Finally, several scenarios based on the cooling dynamics of surface structure were defined and analyzed, and the results were compared to the current fixed layer time policy. It was concluded that the proposed method can significantly increase the efficiency by reducing the overall printing time while preserving the quality.

Localized Acoustic-Event Measurement Probe: Connector Confirmation Utilizing Acoustic Signatures

Electrical connections in consumer products are typically made manually rather than through automated assembly systems due to the high variety of connector types and connector positions, and the soft flexible nature of their structures. Manual connections are prone to failure through missed or improper connections in the assembly process and can lead to unexpected downtime and expensive rework. Past approaches for registering connection success such as vision verification or Augmented Reality have shown limited ability to verify correct connection state. However, the feasibility of an acoustic-based verification system for electrical connector confirmation has not been extensively researched. One of the major problems preventing acoustic based verification in a manufacturing or assembly environment is the typically low signal to noise ratio (SNR) between the sound of an electrical connection and the diverse soundscape of the plant. In this study, a physical means of background noise mitigation and signature amplification are investigated in order to increase the SNR between the electrical connection and the plant soundscape in order to improve detection. The concept is that an increase in the SNR will lead to an improvement in the accuracy and robustness of an acoustic event detection and classification system. Digital filtering has been used in the past to deal with low SNRs, however, it runs the risk of filtering out potential important features for classification. A sensor platform is designed to filter out and reduce background noise from the plant without effecting the raw acoustic signal of the electrical connection, and an automated detection algorithm is presented. The solution is over 75% effective at detecting and classifying connections.

Vision-Based Precision Deburring and Edge Breaking Process Quality Control

This paper presents a computer vision system for evaluating the quality of deburring and edge breaking on aluminum and steel blocks. This technique produces both quantitative (size) and qualitative (quality) measures of chamfering operation from images taken with an off-the-shelf camera. We demonstrate that the proposed computer vision system can detect edge chamfering geometry within a 1–2mm range. The proposed technique does not require precise calibration of the camera to the part nor specialized hardware beyond a macro lens. Off-the-shelf components and a CAD model of the original part geometry are used for calibration. We also demonstrate the effectiveness of the proposed technique on edge breaking quality control.

Modeling of Dry-EDM Plasma Discharge

There is a growing interest in developing the dry EDM process as a sustainable alternative to the conventional liquid dielectric-based EDM process. It is shown that the dry EDM process possesses advantages over the conventional process in terms of thermal damage, recast layer, and tool wear. However, there is a need to increase the productivity of the dry EDM process for its successful adaptation in the industry. This paper presents a model of dry EDM plasma discharge with air as the dielectric medium. The model uses global modeling (‘0D’) approach in which equations of mass balance, energy balance, and plasma expansion are solved simultaneously to obtain a time-dependent description of the plasma in terms of its composition, temperature, diameter, and heat flux to electrodes. The model includes reaction kinetics involving 622 reactions and 55 species to determine the air plasma composition. A single discharge dry EDM operation is successfully simulated using the model, and the effect of discharge current on the plasma is studied. An increase in the discharge current increases the electron density, temperature, and diameter of the plasma linearly, while heat flux to the workpiece increases exponentially. Overall, the model provides an essential tool to study the dry EDM process mechanisms at a fundamental level and devise methods for process improvements.

Laser Sharpening of Carbon Fiber Microelectrode Arrays for Brain Recording

Microwire microelectrode arrays (MEAs) are implanted in the brain for recording neuron activities to study the brain functioning mechanism. Among various microwire materials that had been applied, carbon fiber is outstanding due to its small footprint (6–7 μm), relatively high Young’s modulus, and low electrical resistance. Tips of microwire in MEAs are often sharpened to reduce insertion force. Currently, carbon fiber MEAs are sharpened with either torch burning, which can only give a uniform length of wires in an array, or electrical discharge machining (EDM), which requires circuit connection with each single carbon fiber. The sharp tip results from intense burning induced by a flame or spark, leading to poor repeatability and controllability of the sharp tip geometry. In this paper, a laser-based, non-contact carbon fiber sharpening method is proposed, which enables controllable and repeatable production of carbon fiber MEAs of custom electrode lengths, insulation stripping lengths, and sharpened tips. Path of laser movement is designed according to desired array pattern. Variation in tip geometry can be accomplished by changing laser output power and moving speed. Test with different laser parameters (output power and moving speed) were conducted. Tip sharpening results were evaluated and analyzed in terms of tip geometry and insulation stripping length. Results showed that to achieve the desired MEA with sharper tip and shorter insulation stripping length, a higher laser power with faster moving speed is preferred.

Data-Enabled Modeling and Performance Evaluation of Multistage Serial Manufacturing Systems With Quality Rework

In a multistage serial production line, products with defect can be repaired or reworked to ensure high product quality. This paper studies a multistage serial manufacturing system with quality rework loops. Rework is the activity to repair or repeat the work on the defect parts during manufacturing processes, and it adds to cost and cycle time. This paper introduces an event-based data-enabled mathematical model for a stochastic production line with quality rework loops. The system performance properties are analyzed and permanent production loss due to quality rework loops is identified. The mathematical model and system performance identification methodology are studied analytically through numerical case studies.

Research on Intelligent Optimization Algorithm for Multi-Objective Disassembly Line Balancing Problem

For resource reutilization and environmental protection, remanufacturing gets more and more attention in many countries. Disassembly is a critical part of traditional manufacturing industry, but the traditional disassembly operation is mainly done by workers, which is low-efficiency. Now the use of robots can improve production efficiency a lot, which involves the problem of disassembly line balancing. Due to the constraints such as product complexity and precedence relationship between tasks, when the number of tasks increases, the combination scheme between tasks increases geometrically, and conventional algorithms are difficult to solve the problems, the Disassembly Line Balancing Problem (DLBP) is generally necessary to optimize multiple objectives. In this research, the author selects a variety of intelligent optimization algorithms to resolve the complex disassembly line balancing problem in different dimensional objective space. Four representative algorithms are selected from three angles to be compared through three performance indicators. It is concluded that these algorithms have different search capabilities for different specifications and objective space. Researchers should carefully select the algorithm according to the specific disassembly problem. The appropriate algorithm should be selected according to the scale of the disassembly line problem and the number of optimization objectives in actual production practice.

Sketch-Based Tensor Decomposition for Non-Parametric Monitoring of Electrospinning Processes

Electrospinning is a promising process to fabricate functional parts from macrofibers and nanofibers of bio-compatible materials including collagen, polylactide (PLA), and polyacrylonitrile (PAN). However, the functionality of the produced parts highly rely on quality, repeatability, and uniformity of the electrospun fibers. Due to the variations in material composition, process settings, and ambient conditions, the process suffers from large variations. In particular, the fiber formation in the stable regime (i.e., Taylor cone and jet) and its propagation to the substrate plays the most significant role in the process stability. This work aims to designing a fast process monitoring tool from scratch for monitoring the dynamic electrospinning process based on the Taylor cone and jet videos. Nevertheless, this is challenging since the videos are of high frequency and high dimension, and the monitoring statistics may not have a parametric distribution. To achieve this goal, a framework integrating image analysis, sketch-based tensor decomposition, and non-parametric monitoring, is proposed. In particular, we use Tucker tensor-sketch (Tucker-TS) based tensor decomposition to extract the sparse structure representations of the videos. Additionally, the extracted monitoring variables are non-normally distributed, hence non-parametric bootstrap Hotelling T2 control chart is deployed to handle this issue during the monitoring. The framework is demonstrated by electrospinning a PAN-based polymeric solution. Finally, it is demonstrated that the proposed framework, which uses Tucker-TS, largely outperformed the computational speed of the alternating least squares (ALS) approach for the Tucker tensor decomposition, i.e., Tucker-ALS, in various anomaly detection tasks while keeping the comparable anomaly detection accuracy.

Verification of a Novel Position Verification Sensor to Identify and Isolate Robot Workcell Health Degradation

Manufacturing processes have become increasingly sophisticated leading to greater usage of robotics. Sustaining successful manufacturing robotic operations requires a strategic maintenance program. Maintenance can be very costly, especially when some manufacturers unnecessarily spend resources (i.e., time, money) to maintain their equipment. To reduce maintenance costs, manufacturers are exploring how they can assess the health of their robot workcell operations to enhance their maintenance strategies. Effective health assessment relies upon capturing appropriate data and generating intelligence from the workcell. Multiple data streams relevant to a robot workcell may be available including robot controller data, a supervisory programmable logic controller data, maintenance logs, process/part quality data, and equipment/process fault and/or failure data. This data can be extremely informative, yet the extreme volume and complexity of this data can be both overwhelming, confusing, and paralyzing. Researchers at the National Institute of Standards and Technology have developed a test method and companion sensor to assess the health of robot workcells, which will yield an additional and unique data stream. The intent is that this data stream can either serve as a surrogate for larger data volumes to reduce the data collection and analysis burden on the manufacturer or add more intelligence to assessing robot workcell health. This article will present the immediate effort focused on verifying the companion sensor. Results of the verification test process are discussed along with preliminary results of the sensor’s performance during verification testing.