Machine Failure Research Articles

Role of Artificial Intelligence in IoT

Artificial Intelligence (AI) and the Internet of Things (IoT) are two revolutionary technologies shaping modern industries. AI enables machines to learn from data and make autonomous decisions, while IoT connects devices, allowing them to communicate and exchange information. The convergence of these technologies, known as AIoT (Artificial Intelligence of Things), has immense potential to transform daily life and business operations. AI enhances IoT by adding intelligence to device networks, improving efficiency and automating processes. For example, AI-powered IoT systems can perform predictive maintenance in industrial settings, preventing costly machine failures. In smart buildings, AI can optimize energy usage, reducing both costs and environmental impact. However, integrating AI with IoT poses significant challenges. One major issue is the vast amount of data generated by IoT devices, which AI requires to function effectively. Managing this data efficiently is critical. Additionally, IoT devices are vulnerable to cyber-attacks, making security a key concern. Protecting these systems from potential threats is essential to safely realize the benefits of AIoT. Despite these challenges, the combination of AI and IoT promises to drive future advancements across industries, enabling smarter, more connected systems that optimize decision-making, reduce costs, and improve overall efficiency.

Integration of lean manufacturing system with novel intuitive fuzzy syncretic lean frame work to improve the overall equipment effectiveness

Lean manufacturing embraces the principle of doing more with less by removing non-value-added activities from manufacturing processes to preserve effectiveness, flexibility, and profitability. Predictive maintenance in production stages are more critical since it involves machines which are prone to failure and a post failure maintenance stagnates the production which creates bottlenecks due to machine failures. Hence in this work a novel intuitive fuzzy syncretic lean frame work has been implemented to incorporate the need of intelligent systems. The framework is divided into phases which employs Fuzzy logic with embedded smart sensors for effective utilization of man machine and materials in order to improve the effectiveness of a production floor. A novel time based forecasting is done in the design phase which implements the lean tool Takt time. The manufacturing phase uses the sensors to determine the predictive maintenance of the machines thereby implementing continuous flow as the lean tool and the inspection phase uses smart sensor system for real time continuous monitoring of the machining process thereby incorporating Poka-yoke as a lean tool. Thus by implementing the framework the overall equipment effectiveness is achieved which helps in achieving continuous flow of products and defect free products in any production firm.

Operational Performance Evaluation Model for Food Processing Machinery Industry Chain

This study aims to create a performance evaluation model for the food processing machinery industry. The goal is to help food processing plants improve both process quality and competitiveness. Additionally, component failures may disrupt the continuous operation of the food processing machine, potentially resulting in insufficient production and delays in delivery, which in turn leads to cost losses. For the sold food processing machinery, decreases in the average number of failures within a unit of time, the average repair response time when a failure occurs, and the average repair duration are three crucial factors in minimizing the total expected loss due to machine failures. Based on these three important factors, this study established the following evaluation indices: (1) the processing performance index, (2) the repair reporting performance index, and (3) the maintenance performance index. These indices serve as tools for assessing the performance of the three key operational aspects. This study employed a radar chart to construct the evaluation model, which can directly compare the critical values with the point estimates of three indices. Consequently, this approach can judge whether the operational performance has achieved the required level. This can maintain the simplicity and usability of point estimates while reducing the risk of misjudgment due to sampling errors.

A sustainable fuzzy economic production quantity model (SFEPQM) with sporadic machine failures, inspection and prepayments under carbon tax and cap policy

Goals for sustainability have caused an upcoming change in the emphasis of Economic Production Quantity models, which was previously limited mainly to financial considerations to include aspects of the environment as well. The primary tools used in many nations to reduce emissions are carbon cap and price legislation. However, companies in the production industry may have difficulties with their planning and oversight procedures as they frequently have to cope with faulty products and machinery failures. Previous research has not addressed the comprehensiveness of green investment in machine breakdowns. Considering this problem, we have developed a sustainable economic production quantity model with sporadic malfunctions of machinery and prepayments under carbon tax and cap regulations. Our aim is to minimise the overall cost of faulty items with the investment in green technologies. An approach that combines with the outstanding effectiveness of green strategies to minimise emissions will culminate in a greener production that generates more profit than the traditional system. This study considers two cases i.e. with and without green technology investment (GTI). It also has many predictions, including the execution of inspections afterwards as well as during manufacturing, faulty goods are scrapped while others are sold at discounted price.

Intraoperative Anesthesia-Related Critical Events in Low-Resource Hospitals During Short-Term Surgical Missions in Tanzania and Democratic Republic of the Congo: An Observational Study.

Critical incidents and mortality related to anesthesia are more frequently observed in low- and middle-income countries in comparison to high-income countries. The difficulties linked to anesthesia in rural areas of the Democratic Republic of the Congo (DRC) and Tanzania have limited documentation. The aim of this study was to comprehensively document anesthesia-related critical events that occurred during surgical missions organized by the nongovernmental organization 2nd Chance in hospitals in DRC and Tanzania. Data were collected during 6 surgical missions in 3 hospitals in the DRC and 1 in Tanzania. All scheduled surgery patients were included. Anesthesia was administered by a local Non-Physician Anesthesia Provider (NPAP), using local resources, under the supervision of an anesthesiologist from the association. The anesthesiologist reported critical events and collected data. Local teams managed critical events initially, with intervention by the anesthesiologist from 2nd Chance on the local team's request, according to preestablished protocol or if the situation was considered dangerous. Critical incidents associated with anesthesia, including bradycardia, hypoxemia, airway management failure, and equipment problems, were documented from induction of anesthesia until discharge from the recovery room. We recruited 201 patients, of whom 192 were evaluated, with 9 patients dropping out due to protocol noncompliance. All patients were American Society of Anesthesiologists (ASA) I (62%; n = 120) or ASA II (38%; n = 72). Among them, 104 individuals (54%) experienced at least 1 critical event, totaling 202 critical events. Hypoxemia emerged as the most common event, affecting 29% of the patients (n = 55) with at least 1 episode. Equipment problems (oxygen supply and/or anesthesia machine failure) occurred in 24% of cases (n = 46), airway management issues in 23% (n = 44), and bradycardia in 6% (n = 12). Hypotension and hypertension were not documented due to the lack of monitoring. The majority of these events (over 60%) required intervention by the anesthesiologist. The occurrence of critical events related to anesthesia appears to be high in this study. Due to numerous limitations, these results cannot be generalized to all hospitals in Tanzania and the DRC. However, this study underscores the challenges faced by anesthesia teams, encompassing inadequate resources, equipment deficiencies, and varying levels of expertise among anesthesia personnel. The research further stresses the significance of addressing these challenges to enhance patient safety.

A New Method for Determining the Safety Distance Between Irregular Karst Cave and Circumferential Shield Tunnel

ABSTRACTWhen a shield tunnel passes through a karst area, the water‐filled cave can easily make the surrounding rock metamorphic, resulting in water inrush, ground collapse, and shield machine failure and other engineering hazards. Natural cavities have a significant degree of geometric irregularity due to groundwater alteration and soluble rock erosion. Considering the difficulties in describing the shape of a natural irregular cavity, circular, rectangular, and elliptical geometries have been simplified in most related studies. Based on the upper bound theorem of limit analysis, we established a three‐dimensional failure model including the karst caves located directly above and below the circumferential side of the tunnel. Then, we deduced the corresponding analytical solution of the critical safety distance (CSD). Furthermore, the effects of rock mass parameters, cave parameters, and geometric parameters on the CSD were analyzed. Then we designed the numerical simulation considering the irregular geometry shape at the circumferential side of tunnel using the Fourier descriptors. In addition, we estimated the CSDs for two failure models using the revised dichotomy and failure criterion. The findings demonstrated a quantifiable association between CSD and Fourier descriptors of irregular cave shape, resulting in the development of a CSD prediction model. These test results can provide a theoretical foundation and direction for predicting water inrush caused by the constrained irregular cave.

Deep learning-based anomaly detection using one-dimensional convolutional neural networks (1D CNN) in machine centers (MCT) and computer numerical control (CNC) machines

Computer numerical control (CNC) and machine center (MCT) machines are mechanical devices that manipulate different tools using computer programming as inputs. Predicting failures in CNC and MCT machines before their actual failure time is crucial to reduce maintenance costs and increase productivity. This study is centered around a novel deep learning-based model using a 1D convolutional neural network (CNN) for early fault detection in MCT machines. We collected sensor-based data from CNC/MCT machines and applied various preprocessing techniques to prepare the dataset. Our experimental results demonstrate that the 1D-CNN model achieves a higher accuracy of 91.57% compared to traditional machine learning classifiers and other deep learning models, including Random Forest (RF) at 89.71%, multi-layer perceptron (MLP) at 87.45%, XGBoost at 89.67%, logistic regression (LR) at 75.93%, support vector machine (SVM) at 75.96%, K-nearest neighbors (KNN) at 82.93%, decision tree at 88.36%, naïve Bayes at 68.31%, long short-term memory (LSTM) at 90.80%, and a hybrid 1D CNN + LSTM model at 88.51%. Moreover, our proposed 1D CNN model outperformed all other mentioned models in precision, recall, and F-1 scores, with 91.87%, 91.57%, and 91.63%, respectively. These findings highlight the efficacy of the 1D CNN model in providing optimal performance with an MCT machine’s dataset, making it particularly suitable for small manufacturing companies seeking to automate early fault detection and classification in CNC and MCT machines. This approach enhances productivity and aids in proactive maintenance and safety measures, demonstrating its potential to revolutionize the manufacturing industry.

A hybrid LSTM random forest model with grey wolf optimization for enhanced detection of multiple bearing faults

Bearing degradation is the primary cause of electrical machine failures, making reliable condition monitoring essential to prevent breakdowns. This paper presents a novel hybrid model for the detection of multiple faults in bearings, combining Long Short-Term Memory (LSTM) networks with random forest (RF) classifiers, further enhanced by the Grey Wolf Optimization (GWO) algorithm. The proposed approach is structured in three stages: first, time and frequency domain features are manually extracted from vibration signals; second, these features are processed by a dual-layer LSTM network, which is specifically designed to capture complex temporal relationships within the data; finally, the GWO algorithm is employed to optimize feature selection from the LSTM outputs, feeding the most relevant features into the RF classifier for fault classification. The model was rigorously evaluated using a dataset comprising six distinct bearing health conditions: healthy, outer race fault, ball fault, inner race fault, compounded fault, and generalized degradation. The hybrid LSTM-RF-GWO model achieved a remarkable classification accuracy of 98.97%, significantly outperforming standalone models such as LSTM (93.56%) and RF (98.44%). Furthermore, the inclusion of GWO led to an additional accuracy improvement of 0.39% compared to the hybrid LSTM-RF model without optimization. Other performance metrics, including precision, kappa coefficient, false negative rate (FNR), and false positive rate (FPR), were also improved, with precision reaching 99.28% and the kappa coefficient achieving 99.13%. The FNR and FPR were reduced to 0.0071 and 0.0015, respectively, underscoring the model’s effectiveness in minimizing misclassifications. The experimental results demonstrate that the proposed hybrid LSTM-RF-GWO framework not only enhances fault detection accuracy but also provides a robust solution for distinguishing between closely related fault conditions, making it a valuable tool for predictive maintenance in industrial applications.

An integrated method of extended STPA and BN for safety assessment of man-machine phased-mission system

Man-Machine Phased-Mission System (MMPMS) usually demands the cooperation of operators with different responsibilities and machines to accomplish multi-phase missions. Its machine configuration and human organization structure may change across phases, and phase dependencies of machine failures and human errors may exist. In current studies, the safety of man-machine system is usually analyzed qualitatively by System Theoretic Process Analysis (STPA) and assessed quantitatively by the integration of STPA with Bayesian Networks (BN). These studies only focus on single-phase systems and conduct single-phase BN while cannot address the features of MMPMS. In this paper, a qualitative analysis and quantitative assessment method for phase dependencies is proposed and integrated into the method that combines STPA and BN. Firstly, four types of phase dependencies in MMPMS are identified. Secondly, new mapping rules for phase dependencies are proposed to integrate single-phase BN into a multi-phase BN. Thirdly, the quantitative assessment method for phase dependencies considering the effects of human organization structure changes are proposed to quantify the parameters of multi-phase BN. Fourthly, the safety of MMPMS can be assessed through multi-phase BN. Finally, an Unmanned Aerial Vehicle system with three-phase missions is presented as a case study to demonstrate the effectiveness of the proposed method.

Web-Based Maintenance Prediction of Machine Conditions and Failure Modes Using Machine Learning

Machine failure may result in unplanned downtime, production losses, safety risks, and increased costs. Hence, predicting machine failures is a significant challenge faced by many industries. Predictive Maintenance (PdM) techniques can help mitigate these risks by predicting machine failures before they occur. This research, therefore, develops two supervised Machine Learning (ML) algorithms to predict failure conditions and the associated failure modes for a milling machine. Six machine features were investigated. The ML models were then developed using a four-stage process including data pre-processing, training, evaluation, and deployment. Several ML algorithms were applied then the results were evaluated using accuracy, recall, precision, and [Formula: see text]1-score. The results showed that the Random Forest algorithm was the most effective in predicting machine conditions and failure modes of the milling machine. A web application for PdM was finally developed and tested for the PdM of the milling machine. In conclusion, the developed web application including ML algorithms can support the effective PdM for the machine, which leads to enhancing its availability and improving productivity. Future research considers adopting ML algorithms for predicting machine conditions and failure types of other machines.

Production efficiency analysis based on the RFID-collected manufacturing big data

Radio Frequency Identification (RFID) technology is widely used for production data collection in manufacturing shop-floors. The RFID-collected manufacturing big data reflects detailed statuses of production processes and manufactured products, which, in turn, can be used to support a comprehensive production analysis. This paper introduces an analytical approach to conduct production efficiency analysis based on the RFID-collected manufacturing big data. The proposed method involves four key steps: data cleansing, data processing, key performance indicator (KPI) estimation, and data analytics. Specifically, speed and quality aspects of shopfloor manufacturing are investigated collectively as production efficiency to support the comprehensive production analysis. The findings highlight the influence of various factors, such as operation date, working hour, and machine failure, on production efficiency.

PERANCANGAN SISTEM PERAWATAN PULVERIZER MODEL MVM31F MENGGUNAKAN METODE FAILURE MODE AND EFFECT ANALYSIS (FMEA) DI PT. XYZ

The continuous increase in population growth leads to a rising demand for electrical energy. PT. XYZ as a supplier of electrical energy, relies heavily on the reliability of its components to meet this demand. The frequency of failures in pulverizer machines is one of the factors that reduce production capacity. This study uses the FMEA method combined with LTA to determine the failure consequences caused by failure modes, resulting in appropriate maintenance actions for critical components. From the FMEA analysis, 10 failure modes were identified along with their causes and effects if failures occur. Based on the RPN value calculations, a critical value of 69.2 was obtained. Components with RPN values above the critical value that need prioritized maintenance are the classifier and grinding roller. In the maintenance planning stage, 7 failure modes are placed under time-directed (TD) maintenance, 2 failure modes under condition-directed (CD) maintenance, and 1 failure mode under finding failure (FF) maintenance.

Dynamic job allocation method of multiple agricultural machinery cooperation based on improved ant colony algorithm

Contemporary large-scale and systematic agricultural operations demand the collaborative efforts of multiple agricultural machines with distinct functionalities. However, the failure of a single agricultural machine during collaborative operations jeopardizes the entire undertaking. To address this challenge, this paper proposes a multi-machine collaborative dynamic job allocation method based on the improved ant colony algorithm. Initially, the improved ant colony algorithm is employed to determine the optimal solution for harvester scheduling. This solution is then fed into the data conversion algorithm to acquire the necessary unloading point information for transport vehicle scheduling. Subsequently, the improved ant colony algorithm is once again utilized to optimize the transport vehicle scheduling. In cases of agricultural machinery failure or changes in the operating environment, two distinct methods are employed based on the situation. The first involves double-layer rescheduling of both harvester and transport vehicles, while the second employs single-layer rescheduling exclusively for the transport vehicles, yielding the respective rescheduling results. The outcomes demonstrate that the proposed solution method effectively identifies the current optimal scheduling plan for both the harvester and transport vehicle in the event of malfunctions. Moreover, under the premise that the unproductive waiting time of the harvester is reduced to zero, and the number of transport vehicles is minimized, it achieves the minimization of operating time cost and transportation cost. This method exhibits significant potential for seamless integration into the practical application of unmanned farms, providing a foundation for addressing scheduling and management challenges in multi-agricultural machinery collaborative operations within complex farmland operating environments.

Implementation of machine learning methods in predicting failures in electrical submersible pump machines

The Electric Submersible Pump (ESP) is one of the artificial lift tools widely used in oil and gas wells. Approximately 25% of oil and gas wells have been equipped with ESP. The ESP unit excels in lifting fluid rates much larger than most other types of artificial lift. The purpose of this research is to develop a machine learning model that can predict the likelihood of ESP installation failure. By making predictions at the early stage, necessary actions can be prepared to prevent and anticipate these issues. Thus, the reduction in the amount of produced fluid can be minimized. The model in this study uses 8 input parameters, namely, motor Ampere, Frequency, Pump Intake Pressure, Temperature Motor, Output Volt, Pump Discharge Pressure, Input Voltage, and Motor Horse Power (HP). There are also 8 failure classifications for which the model will be created. The initial stage involves collecting and cleaning raw data and processing it to detect anomalies. Once the data is processed, it proceeds to the model formation and model validation stage. The method used for the model is based on the comparison of two methods, namely, decision tree and k-nearest neighbor. The output of the research is a dashboard that can display visualizations of data and the prediction results from the created model. Millions of data points are used to create a database consisting of 77 wells over the past 2 years. The model developed generates several types of failures that can be read. These failures have their own characteristic parameters. The model obtained has an accuracy rate above 90%. The output of this model includes the most likely failure prediction based on the latest input data, thus effectively addressing well problems, warning of impending failures, reducing failures, and assisting in scheduling ESP repairs and maintenance.

Multi-objective rescheduling of job shop scheduling problems in manufacturing using Quantum Annealing

Unforeseen events in manufacturing necessitate rescheduling to minimize lead times and maintain continuous production. Reducing rescheduling effort is also crucial otherwise additional setup times are incurred. This paper extends a dynamic rescheduling workflow, previously demonstrating success with Quantum Annealing through a multicriterial optimization. It introduces a second objective function considering rescheduling effort alongside makespan. The study addresses job arrivals and machine failures, employing a binary quadratic model for optimization. Results highlight the superiority of Quantum Annealing for larger problems compared to classical metaheuristics. Nevertheless, parameter selection complexities urge further exploration such as exploring deep learning techniques to tackle this challenge.

Data Envelopment Analysis-based Scenario Selection for Sequencing Pattern in a Simulated Robotic Cell

In this study, the performance of suggested scenarios for part input sequences in a 3-machine robotic cell producing different parts is determined through the application of data envelopment analysis (DEA) and the Banker–Charnes–Cooper model. A single gripper robot supports the manufacturing process by loading and unloading products and moving them inside the system. This study addresses random machine failures and repairs to minimize cycle time based on two robot move cycles in a three-machine robotic cell and overall production costs. Here, simulation assists in the modeling of uncertainty and a simulation-based optimization approach is applied to find the best scenarios for sequencing patterns in the cell through several numerical examples using DEA. The results displayed that, efficient scenarios satisfying minimum time and cost, are those, in which the percentages of operations assigned to the machines are close to each other. This enables decision-makers in manufacturing systems to make precise selections of the optimal part sequencing pattern with the lowest production cost and cycle time for robotic cells.

Enhancing machine failure prediction with a hybrid model approach

<span lang="EN-US">The industrial sector is undergoing a substantial transformation by embracing predictive maintenance approaches, aiming to minimize downtime and reduce operational expenses. This transformative shift involves the incorporation of machine learning techniques to refine the accuracy of predicting machinery failures. In this article, we delve into an in-depth exploration of machine failure prediction, employing a hybrid model amalgamating long short-term memory (LSTM) and support vector machine (SVM). Our comprehensive study meticulously assesses the hybrid model’s performance, comparing it with standalone LSTM and SVM models across three distinct datasets. The results showcase that the hybrid model outperformed, providing the modest dependable, and highest F1-score values in our evaluation.</span>

Review of imbalanced fault diagnosis technology based on generative adversarial networks

Abstract In the field of industrial production, machine failures not only negatively affect productivity and product quality, but also lead to safety accidents, so it is crucial to accurately diagnose machine failures in time and take appropriate measures. However, machines cannot operate with faults for extended periods, and the diversity of fault modes results in limited data collection, posing challenges to building accurate fault prediction models. Despite recent advancements, intelligent fault diagnosis methods based on traditional sampling and machine learning have shown notable progress. Nonetheless, these methods heavily rely on human expertise, making it challenging to extract comprehensive feature information. To address these challenges, numerous imbalance fault diagnosis methods based on generative adversarial networks (GANs) have emerged, GANs can generate realistic samples that conform to the distribution of the original data, showing promising results in diagnosing imbalances in critical components such as bearings and gears, despite their great potential, GAN methods also face challenges, including difficulties in training and generating abnormal samples. However, whether it is GAN-based resampling technology or traditional sampling technology, there are fewer reviews on noise-containing imbalance, intra- and inter-class dual imbalance, multi-class imbalance, time series imbalance and other problems in small samples, and there is a lack of a more comprehensive summary of the solutions to the above imbalance problems. Therefore, the purpose of this paper is to deeply explore the imbalance problems under various failure modes, and review and analyze the research methods and results based on GANs on this basis. By suggesting future research directions, this paper aims to provide guidance and reference for research in the field of industrial production maintenance.

Integrated two-stage multi-factory assembly scheduling with maintenance considerations

ABSTRACT In the dynamic landscape of contemporary industry, integrating maintenance practices with production scheduling is essential for sustaining operational efficiency and competitiveness. This study addresses a two-stage multi-factory assembly scheduling problem, introducing an innovative approach that incorporates maintenance practices to enhance system reliability in the face of unexpected machine failures. For deterministic scheduling, a tailored mixed-integer programming model is presented to minimise the makespan. This model is extended to formulate a stochastic schedule, accommodating unforeseen machine breakdowns through stochastic distributions. The extension includes the integration of preventive and corrective maintenance activities into an integrated two-stage multi-factory assembly scheduling problem. To solve the resulting optimisation problem, a decomposition algorithm utilising an exact solver is proposed. This approach breaks down the main model into smaller models, addressing computational challenges. Comparative analyses against the widely adopted CPLEX software across various instances validate the effectiveness of our approach. A significant finding from this comparison is that our decomposition algorithm outperforms the exact solver, achieving the optimal solution at a faster rate. In sensitivity analyses, the results underscore the superior solution-finding capability of our integrated stochastic model compared to maintenance heuristics from existing literature.

Mathematical Framework for Enhancing Machine Failure Prediction in Aviation and Beyond: Leveraging Deep Convolutional Neural Networks and Visual Analytic

Mathematical Framework for Enhancing Machine Failure Prediction in Aviation and Beyond: Leveraging Deep Convolutional Neural Networks and Visual Analytic