Search for: All records

Accurate prediction of repair durations is a challenge in product maintenance due to its implications for resource allocation, customer satisfaction, and operational performance. This study aims to develop a deep learning framework to help fleet repair shops accurately categorize repair time given product historical data. The study uses an automobile repair and maintenance dataset and creates an end-to-end predictive framework by employing a multi-head attention network designed for tabular data. The developed framework combines categorical information, transformed through embeddings and attention mechanisms, with numerical historical data to facilitate integration and learning from diverse data features. A weighted loss function is introduced to overcome class imbalance issues in large datasets. Moreover, an online learning strategy is used for continuous incremental model updates to maintain predictive accuracy in evolving operational environments. Our empirical findings demonstrate that the multi-head attention mechanism extracts meaningful interactions between vehicle identifiers and repair types compared to a feed-forward neural network. Also, combining historical maintenance data with an online learning strategy facilitates real-time adjustments to changing patterns and increases the model’s predictive performance on new data. The model is tested on real-world repair data spanning 2013 to 2020 and achieves an accuracy of 78%, with attention weight analyses illustrating feature interactions. 

When a consumer is finished using an electronic device (End-of- First-Use), they might recycle, resell, donate/give away, trade-in or throw it in the trash. There are security threats if a hostile party obtains the device and extracts data. Data wiping at End- of-First-Use is thus an important security behavior, one that has received scant analytical attention. To explore consumer behavior and reasoning behind data wiping practices, we undertake a survey of the U.S. population. One key result is that 31% of the population did not wipe data when dispositioning a device. When asked why not, 44% replied that they did not find data wiping important or that it did not occur to them. 33% replied the device was broken and data could not be wiped, 12% reported difficulty in wiping and 11% could not find a way to wipe. The 44% who thought data wiping was not important showed lower awareness of the security threat, 23% had heard that data can be recovered from discarded devices, versus 44% for the general population. The most prevalent device types for which data wiping was reported as unimportant are smart TVs, kitchen appliances, streaming, and gaming devices, suggesting that consumers may not be aware that private information is being stored on these devices. To inform future interventions that aim to raise awareness, we queried respondents where they obtained security knowledge. 47% replied that they learned about security threats from a single venue; social media was this single venue 43% of the time. This suggests that social media is a key channel for security education 

The goal of this study is to find patterns in how consumers disposition electronic devices at End-of-First-Use, i.e. store, recycle, resell, trade in, donate/give or and throw in the trash. K-means clustering was used on survey data from 3,747 U.S. respondents across 10 device categories to divide the population into three clusters of consumers based on stated attitudes and knowledge of data privacy, environmental benefits, convenience and other aspects of End-of-First-Use options. We then measure the reported intended disposition of devices for each cluster and compare with the general population. Cluster 1 has higher data security concerns when recycling, reselling or donating, and less knowledge and trust in End-of-First-Use options overall. The intended behavior of cluster 1 shows higher than average uncertainty in what to do at End-of-First-Use and more intent to store (lower values for other options - recycling, reselling and donating). Cluster 2 shows higher knowledge and trust in recycling, reselling, and donation, and slightly higher than average concern about data security of these options. The intended behavior of cluster 2 shows higher intent to resell, trade-in or donate, and lower levels of being uncertain of what to do and of storing. Cluster 3 expresses much less concern about data security, and lower utility of a stored device. Their intended behavior shows less storage and higher levels of other End-of- First-Use" options. While cluster analysis does not yield causal connections, the groups show consistent trends in stated knowledge and attitudes towards different End-of-First-Use options and corresponding planned behaviors. These results indicate there are subgroups of the general population with similar reported attitudes, knowledge and behaviors. The three subgroups do not have distinct demographic characteristics, i.e. knowledge and attitudes regarding disposition of electronics does not depend strongly on age, education level, income and similar factors. Understanding segmentation is useful to investigate more effective interventions to influence behavior for better sustainability outcomes. 

Products often experience different failure and repair needs during their lifespan. Prediction of the type of failure is crucial to the maintenance team for various reasons, such as realizing the device performance, creating standard strategies for repair, and analyzing the trade-off between cost and profit of repair. This study aims to apply machine learning tools to forecast failure types of medical devices and help the maintenance team properly decides on repair strategies based on a limited dataset. Two types of medical devices are used as the case study. The main challenge resides in using the limited attributes of the dataset to forecast product failure type. First, a multilayer perceptron (MLP) algorithm is used as a regression model to forecast three attributes, including the time of next failure, repair time, and repair time z-scores. Then, eight classification models, including Naïve Bayes with Bernoulli (NB-Bernoulli), Gaussian (NB-Gaussian), Multinomial (NB-Multinomial) model, Support Vector Machine with linear (SVM-Linear), polynomial (SVM-Poly), sigmoid (SVM-Sigmoid), and radical basis (SVM-RBF) function, and K-Nearest Neighbors (KNN) are used to forecast the failure type. Finally, Gaussian Mixture Model (GMM) is used to identify maintenance conditions for each product. The results reveal that the classification models could forecast failure type with similar performance, although the attributes of the dataset were limited. 

Accurate prediction of product failures and the need for repair services become critical for various reasons, including understanding the warranty performance of manufacturers, defining cost-efficient repair strategies, and compliance with safety standards. The purpose of this study is to use machine learning tools to analyze several parameters crucial for achieving a robust repair service system, including the number of repairs, the time of the next repair ticket or product failure, and the time to repair. A large dataset of over 530,000 repairs and maintenance of medical devices has been investigated by employing the Support Vector Machine (SVM) tool. SVM with four kernel functions is used to forecast the timing of the next failure or repair request in the system for two different products and two different failure types, namely random failure and physical damage. A frequency analysis is also conducted to explore the product quality level based on product failure and the time to repair it. Besides, the best probability distributions are fitted for the number of failures, the time between failures, and the time to repair. The results reveal the value of data analytics and machine learning tools in analyzing post-market product performance and the cost of repair and maintenance operations. 

Abstract Accurate prediction of product failures and the need for repair services become critical for various reasons, including understanding the warranty performance of manufacturers, defining cost-efficient repair strategies, and compliance with safety standards. The purpose of this study is to use machine learning tools to analyze several parameters crucial for achieving a robust repair service system, including the number of repairs, the time of the next repair ticket or product failure, and the time to repair. A large data set of over 530,000 repairs and maintenance of medical devices has been investigated by employing the Support Vector Machine (SVM) tool. SVM with four kernel functions is used to forecast the timing of the next failure or repair request in the system for two different products and two different failure types, namely, random failure and physical damage. Frequency analysis is also conducted to explore the product quality level based on product failure and the time to repair it. Besides, the best probability distributions are fitted for the failure count, the time between failures, and the time to repair. The results reveal the value of data analytics and machine learning tools in analyzing post-market product performance and the cost of repair and maintenance operations. 

Since its emergence, the cloud manufacturing concept has been transforming the manufacturing and remanufacturing industry into a big data and service-oriented environment. The aggressive push toward data collection in cloud-based and cyber-physical systems provides both challenges and opportunities for predictive analytics. One of the key applications of predictive analytics in such domains is predictive quality management that aims to fully exploit the potentials provided by the enormous data collected via cloud-based systems. As a case study, a data set of hard disk drives’ Self-Monitoring, Analysis and Reporting Technology (SMART) attributes from a cloud-storage service provider has been analyzed to derive some insights about the challenges and opportunities of using product lifecycle data. An analysis of time-to-failure monitoring of hard disk drives in real-time has been carried out and the corresponding challenges have been discussed.