A Deep Neural Network Model for Learning Runtime Frequency Response Function Using Sensor Measurements

The frequency response function (FRF), defined as the ratio between the Fourier transform of the time-domain output and the Fourier transform of the time-domain input, is a common tool to analyze the relationships between inputs and outputs of a mechanical system. Learning the FRF for mechanical systems can facilitate system identification, condition-based health monitoring, and improve performance metrics, by providing an input-output model that describes the system dynamics. Existing FRF identification assumes there is a one-to-one mapping between each input frequency component and output frequency component. However, during dynamic operations, the FRF can present complex dependencies with frequency cross-correlations due to modulation effects, nonlinearities, and mechanical noise. Furthermore, existing FRFs assume linearity between input-output spectrums with varying mechanical loads, while in practice FRFs can depend on the operating conditions and show high nonlinearities. Outputs of existing neural networks are typically low-dimensional labels rather than real-time high-dimensional measurements. This paper proposes a vector regression method based on deep neural networks for the learning of runtime FRFs from measurement data under different operating conditions. More specifically, a neural network based on an encoder-decoder with a symmetric compression structure is proposed. The deep encoder-decoder network features simultaneous learning of the regression relationship between input and output embeddings, as well as a discriminative model for output spectrum classification under different operating conditions. The learning model is validated using experimental data from a high-pressure hydraulic test rig. The results show that the proposed model can learn the FRF between sensor measurements under different operating conditions with high accuracy and denoising capability. The learned FRF model provides an estimation for sensor measurements when a physical sensor is not feasible and can be used for operating condition recognition.

Collaborative Platform for Remote Manufacturing Systems Using Industrial Internet and Digital Twin in the COVID-19 Era

In the wake of COVID-19, significant influence on the manufacturing industries has been observed in the past year due to the restrictions of in-person communications and interactions. As a consequence, manufacturing efficiency has reduced remarkably all over the world. Despite the great harm to the industrial operations under the pandemic, the opportunities for remote collaborative manufacturing system also arise. Effective and efficient remote manufacturing systems for the real industries have been highly demanded. Through the integration of industrial internet and digital twin systems, the remote manufacturing system can be largely facilitated. This paper proposes a general framework for the remote manufacturing system during the COVID-19 era. The concept of the intelligent collaborative remote manufacturing system is firstly reviewed, as well as discussions of the current pandemic situation and its influence on the industries. The current commercial platforms of the systems are also presented. A case study on the lighthouse factories at the Foxconn Technology Group is finally presented for understanding the implementation of the proposed strategy. The effectiveness of the framework has been validated in the real industrial scenarios, and great economic and operational benefits have been obtained. The proposed framework offers a promising solution for the remote manufacturing system under the current pandemic.

Fabrication of Multilayered Polymer Composite Fibers for Enhanced Functionalities

Carbon nanoparticles-based polymer composites have wide applications across different fields for their unique functional properties, durability, and chemical stability. When combining nanoparticle morphologies with micro- or macro-scale morphologies, the hierarchal structure often would greatly enhance the composites’ functionalities. Here in this work, a thermoplastic polyurethane (TPU) and graphene nanoplatelets (GnPs) based multilayered fiber is fabricated through the combination of dry-jet-wet spinning, based on an in-house designed spinneret which accommodates three layers spinning solution, and hot isostatic pressing (HIP), at 220 °C. The multilayered spinneret enables the spinnability of a high GnPs loaded spinning dope, highly elastic, with great mechanical strength, elongation, and flexibility. The HIP process resulted in superior electrical properties as well as a newly emerged fourth hollow layer. Together, such a scalable fabrication method promotes a piezoresistive sensor that is sensitive to uniaxial strain and radial air pressure. The hollow fiber is characterized based on surface morphologies, layer formation, percolation threshold, piezoresistive gauge factor, mechanical stability and reversibility, and air-pressure sensitivity and reversibility. Such facile fabrication methods and unique structures have combined the mechanically robust outer shell with a highly conductive middle sensing layer for a new sensor with great potentials in wearable, robotics, biomedical, and other areas.

Estimation of Maximum Flow Length for CF-PEEK Overmolded Grid Structures Using the Finite Element Method

Injection overmolding process is a high versatile process that permits, when used in combination with fiber reinforced thermoplastic composites, the obtaining of high mechanical properties structures with complex geometries in short time cycles. The maximum flow length is a parameter that reflects the success of filling in a polymer injection molding process. Geometry of the part, rheological properties of the polymer and process parameters, such as injection pressure and temperature, are involved on the value of this parameter and therefore on the viability of a certain configuration. For injection molding manufacturing, the understanding of the relation between maximum flow length and main geometrical parameters of the molded part is fundamental to approach the product design, which is conditioned severely by processing capabilities. In this work, the maximum flow length is obtained for different geometries of an overmolded rectangular stiffener grid of carbon fiber filled polyether eter ketone (CF-PEEK) using the software Moldflow© Adviser© for calculations. Value of maximum flow length is provided as a function of cross section aspect ratio for gate diameters between 0.8 mm and 1.4 mm and cross section areas from 10 to 50 mm2. An exponential decrement of maximum flow length has been observed with the increment of aspect ratio of the cross section as well as a linear increment with the increment of cross section area. Gate diameter variation is slightly related with maximum flow length for the simulated values. These results provide a support tool for geometry sizing in overmolded rectangular grid parts at preliminary design stages.

Conical Thermoplastic Composite Anisogrid Lattice Structure by Innovative Out-of-Autoclave Molding Process

A new manufacturing process for thermoplastic (TP) composite parts has been used to produce conical anisogrid composite lattice structure (ACLS). An out-of-autoclave (OOA) process has been prototyped by using the compaction exerted by a heat-shrink tube after its exposition to heat in oven. Narrow thermoplastic prepreg tapes have been wounded on a metallic conical patterned mold at room temperature; then, the conical structure has been inserted in the heat-shrink tube and heated. TP unidirectional prepreg tapes have been used with polypropylene matrix and glass fibers. After molding, the TP ACLS has been tested under axial and transverse compression. Conical adapters were used in the transverse loading condition to allow uniform application of the load. Density measurement has been also performed to assess the quality of the OOA process. Results of this study show that TP ACLS with complex shape may be produced with OOA solutions without affecting mechanical performance. In fact, porosity levels of the consolidate ACLS are comparable with the initial prepreg despite of the absence of vacuum during molding. Moreover, high compressive stiffness was measured along both directions without observing damages, buckling or cracks in multiple tests. In the future, this kind of technology could be used for larger ACLSs by substituting the heat-shrink tube with a narrow tape to be wound as well after lamination.

Human Motion Position Prediction for Human-Robot Collaboration in Manufacturing Considering Point Cloud Completion Of Human Joint

The research of human-robot collaboration for intelligent manufacturing is being paid gradually increasing attention due to high flexibility and high manufacturing efficiency. Comparing with the traditional manufacturing with low flexibility, human-robot collaboration in manufacturing system provides more personalized and flexible way to cover the shortages of traditional manufacturing mode. In human-robot collaboration system, human motion position prediction in the collaborative space is an essential prerequisite for ensuring the safety of workers. In this paper, 3D sensor Kinect is utilized to directly obtain human joint information. A partial circle delimitation method is used to solve the offset phenomenon of human joint obtained by Kinect, so as to achieve accurate estimation of human joint points. On this basis, an algorithm combing multilayer perceptron and long short-term memory network is explored to predict human motion position accurately. It not only helps to avoid complex feature extraction due to its end-to-end characteristic, but also provide natural interaction manner between human and robot without wearable devices or tags that may become a burden for the former. After that, the experimental results demonstrate that the proposed method makes predicting results accurate, and provides the reliable basis for human position prediction in the human-robot collaboration. This research could be applied to the human motion position prediction in human-robot collaboration process.

A FPGA-Based Cloud-Edge Collaboration Platform in Cloud Manufacturing

To address the issue that many devices are connected to the cloud during the manufacturing process, which causes severe delays in analyzing massive manufacturing data in the cloud, an FPGA-based architecture of cloud edge collaboration is proposed. In this architecture, manufacturing equipment is connected to the cloud through an FPGA-based embedded edge node. The device data obtained by the edge node is processed by the FPGA module and the embedded system module according to the time-sensitivity. Considering the limited computing power of a single edge node, to realize cloud-edge collaborative computing, a communication-oriented task model and a computing model for edge nodes are designed. The task model learns cloud to edge and edge-edge communication, and the task model realizes the function of migrating computing tasks to other nodes. The edge node system’s design is realized based on the communication-oriented task model and the computing model for edge nodes. The cloud edge collaboration method is researched and explored based on this system. A series of comparative experiments, comparing the time delay of the FPGA module and embedded system module processing the same data, the framework’s usability and data processing ability can be verified.

New Perspectives Regarding the Chip Formation Process of Ti-6Al-4V

The mathematical, physical and morphological characteristics of the chip formation process during cutting of Ti-6Al-4V will be analyzed and presented in this paper. In recent years titanium has received more attention due to their unique material properties, such as light weight by height strength, small deformation at high temperatures, low brittleness at low temperatures, and nearly no oxidation at high temperatures, but with the disadvantage that it is difficult to machine. A lot of investigations have been conducted to solve the complex process of machining. But the real complex phenomena at the cutting edge can’t be explained with the help of simplified models. This paper presents a new mathematical-physical model describing the process mechanics leveraging two kinds of friction to explain the metal behavior to strain and stress with self-hardening or softening effects, and the dynamic chip formation behavior due to strain rate discontinuity. All these influencing parameters have an interdependent relationship; thus, they cannot be analyzed separately. The resultant deformation process leads to a grid deformation pattern in the relevant region of the transversal section of a chip that can be used for comparing the theoretical solution with the experimental result. This deformation pattern is the only characteristic that will not disappear after machining. As long as the theoretical results are found to be in agreement with the experimental data of the produced segmented chip, we can be sure, that the models integrating the friction conditions, strain-stress, and metallurgical conditions are correctly developed. In approaching these problems, it is difficult to choose the relevant machining conditions, because a “quick-stop” test is difficult to produce. The reason might be the existing contact conditions at the tool-chip interface, which has an intensive connection due to the diffusion process. Therefore, two different cutting velocities were chosen with the hope that the diffusion is not too intensive; (one slow velocity with vc = 12.5 m/min and a higher velocity with vc = 100 m/min). In addition, a photomicrograph of a chip was taken for the validation process between theoretical and experimental results. Furthermore, the existing temperatures in the contacting zone as well as in the chip formation area could be developed and are discussed and presented in this paper.

Tool Wear Analysis of MTConnect Production Data

This paper presents the analysis of data collected using the MTConnect protocol from a lathe with a Computer Numerical Control (CNC). The purpose of the analysis is to determine an estimated cutting tool life and generate a model for calculating a real-time proxy of cutting tool wear. Various streams were used like spindle load, NC program blocks, the mode, execution etc. The novelty of this approach is that no information about the machining process, beyond the data provided by the machine, was necessary to determine the tool’s expected life. This method relies on the facts that a) it is generally accepted cutting loads increase with tool wear and b) that many CNC machines rely on a small set of regularly run CNC programs. These facts are leveraged to extract the total load for each run of each program on the machine, creating a dataset which is a good indicator of tool wear and replacement. The presented methodology has four key steps: extracting cycle metadata from the machine execution data; computing the integrated spindle loads for every cycle; normalizing the integrated spindle loads between different programs; extracting tool wear rates and changes from the resulting dataset. It is shown that the method can successfully extract the signature of tool wear under a common set of circumstances which are discussed in detail.

Energy Efficient Multi-Robotic 3D Printing for Large-Scale Construction – Framework, Challenges, and a Systematic Approach

An emerging trend in smart manufacturing of the future is robotic additive manufacturing or 3D printing which introduces numerous advantages towards fast and efficient printing of high-quality customized products. In the case of the construction industry, and specifically in large-scale settings, multi-robotic additive manufacturing (i.e., adopting a team of 3D printer robots) has been found to be a promising solution in order to overcome the existing size limitations. Consequently, several research efforts regarding the development and control of such robotic additive manufacturing solutions have been reported in the literature. However, given the increasing environmental concerns, establishing novel methodologies for energy-efficient processing and planning of these systems towards higher sustainability is necessary. This paper presents a novel framework towards energy-efficient multi-robotic additive manufacturing and describes the overall challenges with respect to the energy efficiency. The energy module of the proposed framework is implemented in a simulation environment. In addition, a systematic approach for energy-aware robot positioning is introduced based on the novel concept of reciprocal energy map. The reciprocal energy map is established based on the original energy map calculated by the energy module and can be used for identifying the low energy zones for positioning and relocation of robots during the printing process.