Critical Machine Research Articles

Improving Production Performance Through Multi-Plant Cross Learning

Abstract The advancement in Web-/Internet-based technologies and applications in manufacturing sector has increased the tilization of cyber workspace to enable more efficient and effective ways of doing manufacturing from distributed locations. This work introduces a novel continuous improvement framework to enhance the performance of production lines through multi-plant comparison and learning among identical or similar production lines in different locations by leveraging the information stored on factory cloud. In this work, production data from multiple identical production lines are collected and analyzed to learn the “best” feasible action on critical machines, which offers a new way to optimize the management of product lines. Machine learning and system model are used to find the relationships between the performance index and the available data. A real case study based on multiple similar automotive plants is provided to demonstrate the method and the increases of throughput are predicted.

Effect of Preventive Maintenance on Machine Reliability in a Beverage Packaging Plant

This paper investigates the effect of preventive maintenance on the reliabilities of devices in a bottling plant. Six months of real-time maintenance data were analyzed quantitatively. Based on the breakdown events obtained for each machine, mean time between failure (MTBF), mean time to repair (MTTR), and failure rate (λ) values for individual equipment are calculated. The bottle packer, empty bottle inspector (EBI), and palletizer are identified as the plant's critical machines. A breakdown analysis (BDA) is then performed on the bottle packer and from the failure mode of all the reoccurring problems affecting this machine as a result of ineffective PM. An autonomous maintenance (AM) team is set up as part of establishing an effective PM program to improve the reliabilities of the critical machines that were continually falling. A significant reduction in machine breakdowns is observed after two months of rolling out the AM program. As a result, the reliability of bottle packer increased from 55.30% to 70.80%, while EBI and palletizer increased from 89.20% and 87.20% to 92% and 90.50%, respectively.

The Automation of Critical Path Method using Machine Learning: A Conceptual Study

This research aims to shed light on the use of machine learning in improving, developing and automating the critical path method, solving its problems, studying this effect and its dimensions, and discussing that from many aspects. The research is divided into two theoretical and practical parts. The theoretical part is concerned with studying the critical path method and its advantages, problems and challenges, as well as studying machine learning and artificial intelligence and its dimensions, reviewing materials and sources related to this, and then presenting suggestions and future solutions based on this study. As for the practical section, it is a questionnaire that targeted a segment of engineers, in particular, and others who have sufficient experience in both the critical path method and machine learning, and seeking their opinions on both topics. The result of the theoretical research was 14 theories or proposals that were presented based on the foregoing study. As for the practical questionnaire, a sample of 127 was taken. Through statistical analysis, the results were analyzed and discussed separately, and then a conclusion was drawn regarding them.

Slowdown of Water Dynamics from the Top to the Bottom of the GroEL Cavity.

The GroE molecular chaperone system is a critical protein machine that assists the folding of substrate proteins in its cavity. Water in the cavity is suspected to play a role in substrate protein folding, but the mechanism is currently unknown. Herein, we report measurements of water dynamics in the equatorial and apical domains of the GroEL cavity in the apo and football states, using site-specific tryptophanyl mutagenesis as an intrinsic optical probe with femtosecond resolution combined with molecular dynamics simulations. We observed clearly different water dynamics in the two domains with a slowdown of the cavity water from the apical to equatorial region in the football state. The results suggest that the GroEL cavity provides a unique water environment that may facilitate substrate protein folding.

Study of Diebond Machine Platform on Theta Rotation Performance

Theta rotation on die during diebond process is one of the critical machine responses especially for land grid array (LGA) device with tight tolerances requirement. The paper focuses on the die theta rotation tolerance capability with critical design for LGA device evaluated on two different diebond machine platforms. The evaluation was narrowed down into two main diebond machines with the objective of attaining the best performance in terms of die theta rotation tolerance capability. The study used a side-by-side comparison analysis in terms of theta rotation on the two machines and presented the effect of machine selection on the theta rotation response. Theta rotation was monitored and both machines satisfied the specification of 1 degree of maximum rotation, though diebond Machine 1 was able to produce a more stable diebonding with only around less than 0.15 degree of theta rotation variation. For future works, the selected diebond machine could be used for devices with critical requirement.

International Conference on Advanced and Competitive Manufacturing Technologies milling tool wear prediction using unsupervised machine learning

Degraded or defect machine components and consumables negatively impact manufacturing quality and productivity. Diagnosing and predicting the wear or degradation status of critical machine components or parts are therefore of general interest. To tackle this challenge, data-driven approaches based on supervised machine learning principles have demonstrated promising results. However, supervised learning models capable of degradation identification require large quantities of data. In practice, run-to-failure data in large amounts is usually not available and expensive to obtain. To overcome this issue, this study proposes an unsupervised learning approach for degradation prognostics of machine tool components and consumables. It uses time series of multi-sensor signal data, which are transformed into a feature representation. The features consist of various characterizations of the time series, allowing to make different signal measurements comparable, and cluster them according to their feature values. The herewith obtained density-based clustering model is used to diagnose and predict the degradation states of components and parts in unknown conditions. The novelty in the proposed approach lies within the identification of continuous component and part degradation states based on unsupervised learning principles. The proposal is verified and demonstrated on an exemplary data set containing a small sample of run-to-failure multi-sensor signals of milling inserts and their corresponding wear state. By the application of the proposed procedure on the exemplary data set, we demonstrate that an unsupervised clustering approach is capable of separating wear data such that meaningful and accurate estimations of the part condition are possible. The advantages are its ability to cope with scarce data sets, its limited engineering and hyperparameter tuning effort, and its straightforward implementation to a multitude of degradation and wear diagnostics scenarios.

Performance Analysis of Spinning Machines Using Machine Learning

The amount of data captured during the manufacturing process is usually huge. So monitoring and classifying the performance of machines have become strenuous. As the world is technologically advancing, the industries should also grow along with it like automation of decision making process instead of manual work. „Machine learning‟ is one of the advanced technologies to go for it. This project is about developing an application to reduce the manual work on Spinning Machine Performance Analysis usually done by the use of traditional softwares like MS Excel and other Spread sheets. Data visualization has been done that helped us to better understand the influencing features of Machine performance and the performance level of each machine is predicted with the help of supervised and unsupervised machine learning algorithms. The objectives of this project include- reducing the manual work to retrieve machine wise breakage rate for a period, to analyze machine wise OEE % in order to improve the MTTR, MTBF and to identify critical machines for proactive actions, which can be achieved by predicting the performance level of each machine.

SME 4.0: Machine learning framework for real-time machine health monitoring system

Over the past ten years, manufacturers and consumers have become increasingly interested in the applications of smart, sustainable, and autonomous systems in the industry and in everyday life. Due to the recent industrial revolution (Industry 4.0), most of the existing Small and Medium-sized Enterprises (SMEs) also want to adapt their work environment into the smart system by the applications of these technologies such as the Industrial Internet of Things (IIoT), Artificial Intelligence (AI) techniques, Machine Learning Algorithm (MLA), Internet Communication Technology (ICT), and Cyber-Physical System (CPS). Because they are very much interested in maximizing productivity, machine availability, reliability, and customer satisfaction in this competitive industrial world. This research study particularly focuses on the Predictive Maintenance (PdM) activity of critical machines and their components in the SME based on the maintenance history dataset through the application of the supervised machine learning algorithm such as Logistics Regression (LR) and K-Means (K-Nearest Neighbor) approaches. In accordance, the real-time case study is presented in SMEs in the southern region of Tamil Nadu, India with two-phase activities. Initially, the optimal failure rate of the machines is predicted by the utilization of LR trained models. Then trigger the man-machine communication and suitable decision-making process of service and maintenance activity through the application of the K-Mean approaches. The main objective of this research study is to organize the smart PdM activity of the smart factory systems in SMEs with the application of MLA based on the real-time maintenance history dataset.

A hybrid many-objective evolutionary algorithm for flexible job-shop scheduling problem with transportation and setup times

This paper addresses a many-objective flexible job-shop scheduling problem with transportation and setup times (MaOFJSP_T/S) where the objective is to minimize the makespan, total workload, workload of the critical machine, and penalties of earliness/tardiness. We first present a mathematical model as the representation of the problem, and then establish a network graph model to describe the structural characteristics of the problem and develop a new neighborhood structure. The neighborhood structure defines four move types for different objectives. Next, we propose a hybrid many-objective evolutionary algorithm (HMEA), which is designed to better balance exploitation and exploration. In this algorithm, the tabu search with the neighborhood structure is proposed to improve the local search ability. A reference-point based non-dominated sorting selection is presented to guide the algorithm to search towards the Pareto-optimal front and maintain diversity of solutions. Through three sets of experiments based on 28 benchmark instances, the partial and overall effects of this algorithm are evaluated. The experimental results demonstrate the effectiveness of the proposed HMEA in solving the MaOFJSP_T/S.

Optimal Contract for Machine Repair and Maintenance

Maintenance outsourcing is quite common in industries that rely on complex and critical equipment. Instead of investing in the maintenance facilities, firms outsource maintenance activities to specialized companies. However, it may be hard for firms (i.e., principal) to observe whether maintenance companies (i.e., agent) put sufficient resources into providing the best service, which gives rise to agency issues. In a dynamic environment in which an agent is responsible for both maintenance and repair of a critical machine, how the principal uses payments and termination to tackle agency issues is a challenging problem. In “Optimal Contract for Machine Repair and Maintenance,” F. Tian, P. Sun, and I. Duenyas provide theoretical guidance on designing the optimal contract to induce efforts from an agent to efficiently operate a machine. Although they consider the very general contract forms, the optimal contracts demonstrate simple and intuitive structures, making them easy to describe and implement in practice.

A Pareto based discrete Jaya algorithm for multi-objective flexible job shop scheduling problem

Research on the development of Pareto-based multi-objective algorithms to address scheduling problems has attracted a lot of attention in recent years. In this work, a multi-objective discrete Jaya algorithm (MODJA) is proposed to address the flexible job shop scheduling problem (FJSSP) considering the minimization of makespan, total workload of machines, and workload of critical machine as performance measures. A discrete Jaya algorithm is proposed to handle the problem under consideration. A problem specific neighborhood-based local search technique is integrated into the proposed approach to enhance its exploitation capability. Further, a dynamic mutation operator and a modified crowding distance measure are proposed to enhance the diversity in the search process. Extensive computational experiments are carried out considering 203 instances of FJSSP from literature. The Taguchi method of design is employed to identify the best set of key parameters based on three instances. In the experimentation phase, initially, the contribution of the proposed local search technique and crowding distance measure is investigated. Then, a comparison of MODJA with the weighted sum version of the approach and other multi-objective evolutionary algorithms is performed. Computational results demonstrate the effectiveness of the proposed MODJA in obtaining diverse and improved Pareto-optimal solutions.

Development of strategic maintenance prediction model for critical equipment maintenance

Abstract Manufacturing industries deploy a number of equipment which are functionally critical to the daily operations of the system and as such requires effective maintenance, and monitoring for high productivity to be achieved. Though maintenance of facilities has gained impressive recognition in many industries, the implementation of inappropriate maintenance practices has increased level of frustrations due to little result from huge financial commitment to maintenance in many production systems. In this work a programmable method was developed for the determination of appropriate maintenance strategy for critical machine in a manufacturing company. The system tagged C3MPS (critical machine maintenance management program software) was developed on the strength of Parallel Criteria Model (PCM), Equipment Criticality Rating (ECR) and Analytic Hierarchy Process (AHP) as subroutines for the establishment of effective maintenance practice. An application of the resulting software in a plastic production company suggested a predictive maintenance as the most effective compared to the existing strategies in practice by the company. The software developed has capacity to effectively evaluate the criticality of equipment and maintenance practice.

Transient Energy of an Individual Machine PART I: Stability Characterization

The individual-machine-equal-area-criterion method already shows potential in the transient stability assessment of multimachine systems. This two-paper series systematically analyzes the transient characteristics of the system trajectory from a genuine individual-machine transient energy perspective. In the first paper, the stability property that characterizes a critical machine is investigated by defining the residual kinetic energy, which only occurs at the maximum individual-machine potential energy point. We indicate that the transient energy conversion inside a critical machine can delineate the trajectory stability of the machine. In this way, the mapping between individual-machine kinetic energy and individual-machine trajectory can hence be established. Simulation results show that the individual-machine transient energy describes the system trajectory more precisely than the superimposed global transient energy in terms of transient stability analysis, and we also demonstrate that the conjecture in early individual-machine studies might fail in special simulation cases.

A maintenance scheduling optimization model for a multi-component machine in a digitalized manufacturing context

Abstract Manufacturing systems are facing a digital transformation that supports data-driven decision making and the improvement of the efficiency of condition-based or predictive maintenance systems. In this context, international standards and previous works related to the architecture of condition-based maintenance (CBM) systems are available, and the interest in maintenance scheduling optimization, triggered by failure predictions, are increasing. This work is framed within the scope of the PICOePRO project and considers a critical multi-component welding machine. This paper first provides a predictive maintenance architecture, coherent with the OS A-CBM standard and proposed within the project, and then develops one of the modules of this system, in which the main component is a maintenance scheduling optimization model. The defined architecture allows real-time decision-making, and the proposed model minimizes the machine maintenance cost, exploring the potential for cost-saving by grouping maintenance and using machine downtime to perform maintenance activities. The functioning of the module and the potential gains from the adoption of the developed model are illustrated using a numerical example. The architecture and the model provide a viable approach to guide maintenance decision-making in the era of digitalized manufacturing.

Methodology for the assessment of potentials, selection, and design of Predictive Maintenance solutions

PM solutions are able to predict and thus minimize production downtimes. Concerning high implementation costs of PM solutions, it is not always clear, when potentials outweigh additional investment. This paper presents a methodology to systematically assess potentials and benefits of specific PM solutions in production processes. The methodology includes an identification of critical machines and machine components in production processes and the quantification of their downtime costs, which are compared to simulation-, experience- and literature-based target costs enabling an evaluation of failure consequences, thus the component’s criticality, and a goal-oriented selection and designing of PM models.

Nondestructive testing magnetic methods adaptation to the metal products accumulated strain and residual stresses determine

The first kind residual stresses (RS) is one of the key metal characteristics, that significantly affect the critical machine parts durability. To date, most of the RS experimental determination methods is based on layer-by-layer surface layer (SL) metal removal and has a number of disadvantages, the main of which are high labor intensity, the quickly RS determine inability without destroying the workpiece. In this regard, it is important to develop the RS and accumulated strain determining experimental Express on the basis of non-destructive testing (NDT) methods. The paper presents the experimental results and describes the relationship between the samples relative strain and RS values after 3-point bending and uniaxial tension with the parameters measured by magnetic NDT methods.

A hybrid reliability-centered maintenance approach for mining transportation machines: a real case in Esfahan

PurposeThe purpose of this study is to propose a hybrid reliability-centered maintenance (RCM) approach for mining transportation machines of a limestone complex, a real case in Esfahan, Iran.Design/methodology/approachCriteria for selecting critical machines were collected within literature and selected by decision-makers (DCs), and critical machines have been identified using the preference ranking organization method for enrichment of evaluations (PROMETHEE). Also, multi-criteria decision-making (MCDM) methods were used in addition to failure mode, effects and criticality analysis (FMECA) for selecting and prioritizing high-risk failures as well as optimizing the RCM performance. More specifically, the criteria of severity, detectability and frequency of occurrence were selected for risk assessment based on the previous studies, and were weighted using the analytic hierarchy process (AHP) method. Also, the technique for order of preference by similarity to ideal solution (TOPSIS) has been applied to prioritize failures' risk. Finally, the critical failures were inserted in the RCM decision-making worksheet and the required actions were determined for them.FindingsAccording to the obtained values from PROMEHTEE method, the machine with code 739-7 was selected as the first priority and the most critical equipment. Further, based on results of TOPSIS method, the failure mode of “Lubrication hole clogging in crankpin bearing due poor quality oil,” “Deformation of main bearing due to overwork” and “The piston ring hotness due to unusual increase in the temperature of cylinder” have the highest risks among failure modes, respectively.Originality/valueRCM has been deployed in various studies. However, in the current study, a hybrid MCDM-FMECA has been proposed to cope with high-risk failures. Besides, transportation machineries are one of the most critical equipment in the mining industry. Due to noticeable costs of this equipment, effective and continuous usage of this fleet requires the implementation of proper maintenance strategy. To the best of our knowledge, there is no research which has used RCM for transportation systems in the mining sector, and therefore, the innovation of this research is employment of the proposed hybrid approach for transportation machineries in the mining industry.

A Single Machine Equivalent-Based Approach for Online Tracking of Power System Transient Stability

This paper presents a systematic single machine equivalent system based scheme for fast and accurate prediction of impending angular stability using measurements from generator terminals only. In a post-fault scenario, Kalman filter is designed to estimate rotor positions in real-time, based on damper winding observer with measurements gathered from generator terminals. To identify critical machines, short time-ahead prediction of rotor position is made by designing another Kalman filter. A general method is suggested to differentiate between critical and non-critical machines. To further enhance the accuracy of predicted time to instability, an extrapolation algorithm is developed, based on the knowledge of estimated rotor angles and short time-ahead predicted rotor angles. The proposed methods are successfully implemented on WSCC 9-bus, IEEE 39-bus, and NRPG 246-bus test systems.

Bearing defect diagnosis based on semi-supervised kernel Local Fisher Discriminant Analysis using pseudo labels

In bearings defect diagnosis applications, information fusion has been widely used to improve identification accuracy for different types of faults, which may lead to high-dimensionality and information redundancy of the data and thus degenerate the classification performance. Therefore, it is a major challenge for machinery fault diagnosis to extract optimal features from high-dimensional and redundant data for classification. In addition, in order to guarantee the performance of fault diagnosis, conventional supervised methods usually require a large amount of labeled data available for learning. However, it is extremely difficult, costly and time-consuming to collect faulty labeled samples with class information, especially for expensive and critical machines, which often results in only a few labeled data available with a large amount of unlabeled data redundant. In this paper, we propose a novel bearing defect diagnosis model based on semi-supervised kernel local Fisher Discriminant Analysis (SSKLFDA) using pseudo labels, which can effectively extract optimal features for classification and simultaneously utilize unlabeled data for regularizing the supervised dimensionality reduction. The proposed SSKLFDA first adopts Density Peak Clustering technique to generate pseudo cluster labels for the labeled and unlabeled data and then regularizes the between-class scatter and within-class scatter according to two corresponding regularization strategies associated with the generated pseudo cluster labels. This regularization can further improve the discriminant performance of the extracted features and also make it suitable for the cases with the multimodality and noises. In order to accommodate for non-linear feature extraction, the kernel version of the proposed method is also provided with the introduction of kernel trick. The experimental results under different feature dimensions, numbers of labeled data, and subsequent classifiers scenarios demonstrate that the proposed SSKLFDA based bearings fault diagnosis model achieves higher classification performance than other existing dimensionality reduction methods-based models.

Perceptual bias and technical metapictures: critical machine vision as a humanities challenge

In many critical investigations of machine vision, the focus lies almost exclusively on dataset bias and on fixing datasets by introducing more and more diverse sets of images. We propose that machine vision systems are inherently biased not only because they rely on biased datasets but also because their perceptual topology, their specific way of representing the visual world, gives rise to a new class of bias that we call perceptual bias. Concretely, we define perceptual topology as the set of those inductive biases in machine vision systems that determine its capability to represent the visual world. Perceptual bias, then, describes the difference between the assumed “ways of seeing” of a machine vision system, our reasonable expectations regarding its way of representing the visual world, and its actual perceptual topology. We show how perceptual bias affects the interpretability of machine vision systems in particular, by means of a close reading of a visualization technique called “feature visualization”. We conclude that dataset bias and perceptual bias both need to be considered in the critical analysis of machine vision systems and propose to understand critical machine vision as an important transdisciplinary challenge, situated at the interface of computer science and visual studies/Bildwissenschaft.