An efficient strategy for robotic surface finishing of curved surfaces that includes real-time identification of the surface profile, control, and implementation is presented in this paper. Real-time identification of the surface profile in the robot base frame is accomplished by employing a proximity laser sensor mounted on the robot end-effector. This surface profile description allows us to generate trajectories for both motion and force control as it provides the surface normal at each point of the surface. Using the surface profile, a trajectory is generated that would orient the surface finishing tool to the local normal of the surface. An algorithm for simultaneous position and force control is developed for surface finishing of curved surfaces. The integrated robotic surface finishing system consists of a UR5 robot and a custom end-effector that includes a force/torque sensor, an electromechanical sander, and the proximity laser sensor. Robot Operating System (ROS) is utilized for real-time implementation, which would enable easy migration of the developed tools if other industrial robots are used. The effectiveness of the strategy is evaluated by conducting a number of experiments for flat and curved surfaces. A representative sample of results on force regulation and surface finishing are presented and discussed.

Surface finishing is an important manufacturing process. Many parts with complex geometries require finishing of internal regions before they can be used. In small and medium volume productions most of the finishing tasks are non-repetitive in nature, and have to be performed manually. These finishing operations for parts with complex geometries can be quite labor intensive, and may pose risk to humans. We have developed a collaborative finishing system where human operators work on high level decision making, and the robot assistants carry out the labor intensive low level finishing tasks. The human operator guides the robotic system by transferring operator knowledge through a user interface. Our system generates instructions for the robots based on the user inputs and task requirements. We have also developed a planning algorithm that automatically computes the paths for the robots by using the CAD model of the part. This significantly reduces the robot programming time and improves the efficiency of the finishing system. If needed, the system seeks help from the human operator by generating notifications.

Honeycomb sandwich panels are widely used for high performance parts subject to bending loads, but their manufacturing costs remain high. In particular, for parts with non-flat, non-uniform geometry, honeycombs must be machined or thermoformed with great care and expense. The ability to produce shaped honeycombs would allow sandwich panels to replace monolithic parts in a number of high performance, space-constrained applications, while also providing new areas of research for structural optimization, distributed sensing and actuation, and on-site production of infrastructure. Previous work has shown methods of directly producing shaped honeycombs by cutting and folding flat sheets of material. This research extends these methods by demonstrating work towards a continuous process for the cutting and folding steps of this process. An algorithm for producing a manufacturable cut-and-fold pattern from a three-dimensional volume is designed, and a machine for automatically performing the required cutting and parallel folding is proposed and prototyped. The accuracy of the creases placed by this machine is characterized and the impact of creasing order is demonstrated. Finally, a prototype part is produced and future work is sketched towards full process automation.

In this paper, we present an approach to determine probing points for the CMM (coordinate measurement machine) measurement. A surface uncertainty based approach is developed to maximize the amount of information acquired by the touch probing of the surface deviations caused by machining errors. The surface uncertainty is modeled by the Gaussian process model and the probing points are selected to minimize the maximum surface uncertainty (surface variance) conditioned on the touch probings at the selected points. The algorithm has been tested with various numerical examples and has been applied in real machining scenarios. The surface reconstruction error based on the developed algorithm is 50% smaller than uniform sampling. The experiments of on machine probing has validated that the selected points can adequately capture the machining errors.

This paper addresses a bi-objective distribution permutation flow shop scheduling problem (FSP) with setup times aiming to minimize the makespan and the total tardiness. It is very difficult to obtain an optimal solution by using traditional approaches in reasonable computational time. This paper presents an appropriate non-dominated sorting Genetic Algorithm III based on the reference point. The NEH strategy is applied into the generation of the initial solution set. To validate the performance of the NEH strategy improved NSGA III (NNSGA III) on solution quality and diversity level, various test problems are carried out. Three algorithms, including NSGA II, NEH strategy improved NSGA II(NNSGA II) and NNSGA III are utilized to solve this FSP. Experimental results suggest that the proposed NNSGA III outperforms the other algorithms on the Inverse Generation Distance metric, and the distribution of Pareto solutions are improved excellently.

In this paper, a multi-objective hybrid flowshop rescheduling problem (HFRP) is addressed in a dynamic shop environment where two types of real-time events, namely machine breakdown and job cancellation, simultaneously happen. For the addressed problem, two objectives are considered. One objective concerning the production efficiency is minimizing the maximum completion time or makespan, while regarding with the instability, the total number of the jobs assigned to different machines between the revised and the origin schedule is considered. A multi-objective evolutionary algorithm based on decomposition (MOEA/D) is applied to solve this problem. In the algorithm, the weighted sum approach is used as the decomposition strategy. The algorithm is, then, rigorously compared with three state-of-the-art evolutionary multi-objective optimizers, and the computational results demonstrate the effectiveness and efficiency of the algorithm.

Most researches on process planning optimized machining process routings and cutting parameters independently and ignored their comprehensive effects on carbon reduction. In order to further reduce carbon emissions in manufacturing processes, an optimization model of cutting parameters and machining process routings is proposed to minimize total carbon emissions and total processing time of all processes. Carbon emissions include those caused by energy consumptions of machines in cutting state, material consumption of cutting tools and cutting fluid in all processes. As the optimization of cutting parameters is a continuous optimization problem, but the optimization of machining process routings including machining methods, process sequences, machine allocating and cutter selecting are discrete optimization problems, the whole optimization of process planning is divided into two parts. One is continuous optimization of cutting parameters. Another is discrete optimization of machining process routings. A hybrid optimization strategy of bird swarm algorithm (BSA) and NSGA-II algorithm is proposed to optimize the proposed model. Cutting parameters are optimized using BSA aiming at minimizing carbon emissions and machining time of each process. Machining process routings are optimized using NSGA-II under each optimized group of cutting parameters from the Pareto set. Four kinds of mutation operators in NSGA-II are designed for the discrete optimization of machining process routings. A workpiece with six machining features to be machined in a workshop with two CNC lathes, two CNC milling machines and two drilling machines is taken as a case study. The validity of the proposed model and hybrid strategy is verified by computational and analytical results. Several conclusions are yielded.

To estimate the heat flowing into the workpiece in machining processes, an inverse algorithm based on the Conjugate Gradient Method (CGM) is proposed to estimate the unknown boundary heat flux. Outgoing from infrared temperature measurements the heat flowing into the work-piece for an orthogonal cut can be estimated. To increase convergence of the estimated solution, a sensitivity analysis of the direct problem is performed to determine the identifiability of the boundary heat flux on the measurement site. The proposed Fixed Identifiability Conjugate Gradient Method (FIX-CGM) computes a step size function considering the identifiability of the unknown boundary condition to minimize the objective function. In contrast, the CGM computes a scalar step size by integrating the difference between measured and calculated temperature over time. Results show that applying the FIX-CGM for a benchmark case with a step heat flux faster convergence, better accuracy and less sensitivity to noise are achieved.

In this paper, a manufacturing work cell with a gantry that is in charge of moving materials/parts between machines and buffers is considered. With the effect of the gantry movement, the system performance becomes quite different from traditional serial production lines. In this paper, reinforcement learning is used to develop a gantry scheduling policy in order to improve system production. The gantry learns to take proper actions under different situations to reduce system production loss by using Q-Learning algorithm and finds the optimal moving policy. A two-machine one-buffer work cell with a gantry is used for case study, by which reinforcement learning is applied. Compare with the FCFS policy, the fidelity and effectiveness of the reinforcement learning method are also demonstrated.

In order to help manufacturing companies quantify and reduce product carbon footprints in a mixed model manufacturing system, a product carbon footprint oriented multi-objective flexible job-shop scheduling optimization model is proposed. The production portion of the product carbon footprint, based on the mapping relations between products and the carbon emissions within the manufacturing system, is proposed to calculate the product carbon footprint in the mixed model manufacturing system. Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) is adopted to solve the proposed model. In order to help decision makers to choose the most suitable solution from the Pareto set as its execution solution, a method based on grades of product carbon footprints is proposed. Finally, the efficacy of the proposed model and algorithm are examined via a case study.

This work reports the development of robust and efficient algorithms for optimum process planning of Additive Manufacturing (AM) processes needing support structures during fabrication. In particular, it addresses issues like part hollowing, support structure generation and optimum part orientation. Input to the system is a CAD model in STL format which is voxelized and hollowed using the 2D Hollowing strategy. A novel approach to design external as well as internal support structures for the hollowed model is developed considering the wall thickness and material properties. Optimum orientation of the hollowed part model is computed using Genetic Algorithm (GA). The Fitness Function for optimization is the weighted average of process performance parameters like build time, part quality and material utilization. A new performance measure has been proposed to choose the weightages for performance parameters to obtain overall optimum performance. The paper presents, in detail, the design and development of algorithms with results for typical case studies. The proposed methodology will significantly contribute to improving part quality, productivity and material utilization for AM processes.

The goal of this work is online quality monitoring of flexible electronic devices made using Aerosol jet printing (AJP) additive manufacturing (AM) process. In pursuit of this goal, the objective is to recover and quantify the 3D topology of AJP-printed electronic traces (lines) through in situ images. The intent is to use the estimated 3D topology for online prediction of the device electrical performance characteristics. To realize this objective different shape-from-shading (SfS) techniques are tested to recover the 3D topology of lines from high resolution in situ 2D images. These images are obtained from a CCD camera installed on our experimental Optomec AJ300 AJP setup. White light interferometry is used offline to verify the online experimental trends. Three types of SfS algorithms are tested, namely, minimization method, Pentland’s method, and Shah’s method. Tests with synthetic images and experimental data indicate that Shah’s method is more suitable. The correlation between online and offline estimates of line thickness was ∼80%.

In a fully automated manufacturing system, tool condition monitoring system is essential to detect the failure in advance and minimize the manufacturing loses with the increase in productivity. To look for a reliable, simple and cheap solution, this paper proposes a new tool wear monitoring model to detect the tool wear progression and early detection of tool failure in end milling using audible sound signals. In this study, cutting tools are classified into six classes based on different flank wear ranges. A series of end milling experiments are operated with a broad range of cutting conditions for each class to collect sound signals. A machine learning algorithm that incorporates support vector machine (SVM) approach coupled with the application of time and frequency domain analysis is developed to correlate observed sound signals’ signatures to tool wear conditions. The performance evaluation results of the proposed algorithm have shown accurate predictions in detecting tool wear conditions from the sound signals. In addition, the proposed machine learning approach has shown the fastest response rate, which provides the good solution for on-line cutting tool monitoring.

This paper provides a case study of diagnosing helicopter swashplate ball bearing faults using vibration signals. We develop and apply feature extraction and selection techniques in the time, frequency, and joint time-frequency domains to differentiate six types of swashplate bearing conditions: low-time, to-be-overhauled, corroded, cage-popping, spalled, and case-overlapping. With proper selection of the features, it is shown that even the simple k -nearest neighbor ( k -NN) algorithm is able to correctly identify these six types of conditions on the tested data. The developed method is useful for helicopter swashplate condition monitoring and maintenance scheduling. It is also helpful for testing the manufactured swashplate ball bearings for quality control purposes.

Personal thermal comfort is a crucial yet often over-simplified factor in building climate control. Traditional comfort models lack the adaptability to fit individuals’ demand. Recent advances of machine learning and ubiquitous sensor networks enable the data-driven approach of thermal comfort. In this paper, we built a platform that can simulate occupants with different thermal sensations and used it to examine the performance of support vector machine (SVM) and compared with several other popular machine learning algorithms on thermal comfort prediction. We also proposed a hybrid SVM-LDA thermal comfort classifier that can improve the efficiency of model training.

Disruptions happen in the actual manufacturing system under environment of the Internet of Things and they make the system difficult to manage. However, the convenient access to information of orders, equipment and participants make disruption recovery easier. In this paper we build a disruption recovery scheduling integer programming model considering the objective of minimizing total weighted completion-time (as the original objective) and the objectives of maximizing total consumer satisfaction degree & minimizing total deviation degree (as the revising objective). A PVPS (PSO & VNS Parallel Search) algorithm is proposed. The experiments results prove all above are effective.

Fine-scale characterization and monitoring of spatiotemporal processes are crucial for high-performance quality control of manufacturing processes, such as ultrasonic metal welding and high-precision machining. However, it is generally expensive to acquire high-resolution spatiotemporal data in manufacturing due to the high cost of the 3D measurement system or the time-consuming measurement process. In this paper, we develop a novel dynamic sampling design algorithm to cost-effectively characterize spatiotemporal processes in manufacturing. A spatiotemporal state-space model and Kalman filter are used to predictively determine the measurement locations using a criterion considering both the prediction performance and the measurement cost. The determination of measurement locations is formulated as a binary integer programming problem, and genetic algorithm is applied for searching the optimal design. In addition, a new test statistic is proposed to monitor and update the surface progression rate. Both simulated and real-world spatiotemporal data are used to demonstrate the effectiveness of the proposed method.

Given the chord length of equal angle isn’t equal on elliptical section of piston skirt with middle-bulged varying ellipse (PSMVE), a larger theoretical processing error will inevitably be introduced with the interpolation algorithm of equal angle for PSMVE. To improve the manufacturing precision of PSMVE, a novel interpolation algorithm of equal-length-chord and spiral-line (IAES) and a method of tool radius compensation are presented based on symbolic computation. Firstly, a three-dimensional model of PSMVE is generated through the symbolic computation method, meanwhile, the coordinate values of arbitrary cutter-contact point can be expressed accurately. Secondly, the turning spiral trajectory is generated via updated cutter-contact point which can be searched from the obtained cutter-contact point with equal length chord. Besides, this paper proposes a method of tool radius compensation and obtains the cutter location points through appropriate transformation of coordinates. Last, some simulation, which mainly includes the establishment of 3D model, the generation of spiral trajectory with equal-length-chord, the transformation between cutter-contact point and cutter-location points, is carried out. In addition, this paper takes CNC turning center (System: SINUMERIK 802C) as an example to complete the processing of PSMVE. Experiment results verify that the machining method is appropriate for PSMVE.

Artificial intelligence techniques can play a significant role in solving problems encountered in the domain of Total Productive Maintenance (TPM). This paper considers a new reinforcement learning algorithm called iSMART, which can solve semi-Markov decision processes underlying control problems related to TPM. The algorithm uses a constant exploration rate, unlike its precursor R-SMART, which required exploration decay. Numerical experiments conducted here show encouraging behavior with the new algorithm.

Manufacturers have faced an increasing need for the development of predictive models that help predict mechanical failures and remaining useful life of a manufacturing system or its system components. Model-based or physics-based prognostics develops mathematical models based on physical laws or probability distributions, while an in-depth physical understanding of system behaviors is required. In practice, however, some of the distributional assumptions do not hold true. To overcome the limitations of model-based prognostics, data-driven methods have been increasingly applied to machinery prognostics and maintenance management, transforming legacy manufacturing systems into smart manufacturing systems with artificial intelligence. While earlier work demonstrated the effectiveness of data-driven approaches, most of these methods applied to prognostics and health management (PHM) in manufacturing are based on artificial neural networks (ANNs) and support vector regression (SVR). With the rapid advancement in artificial intelligence, various machine learning algorithms have been developed and widely applied in many engineering fields. The objective of this research is to explore the ability of random forests (RFs) to predict tool wear in milling operations. The performance of ANNs, SVR, and RFs are compared using an experimental dataset. The experimental results have shown that RFs can generate more accurate predictions than ANNs and SVR in this experiment.