Search

Thermal error modeling (TEM) stands as a pivotal factor in maintaining the machining accuracy of electric spindles. Deep learning (DL) techniques have shown promising potential in this area, however, they face formidable challenges including domain shift issue between training and testing datasets, as well as scarce samples due to the dynamic nature of variable working conditions. Deep transfer learning (DTL) emerges as an encouraging tool, leveraging knowledge obtained from the source domain to elevate learning outcomes within the target domain, notably addressing complicated cross-domain modeling predicaments. Nonetheless, prevailing research largely focuses on enhancing the prediction accuracy within the specific working condition, inadvertently sidestepping the problems of domain shift and scarce sample modeling. To address these issues, a novel weakly supervised adversarial network (WSAN) is proposed for cross-domain TEM with scarce samples. Firstly, the multi-scale convolutional neural network (MSCNN) is constructed to adeptly extract predictive information from multi-sensor data, effectively capturing the time-series patterns embedded within. Furthermore, the adversarial training technique is adopted to address distribution discrepancies that often manifest between distinct domains. To overcome the challenge of scarce labeled samples in the target domain, a weakly supervised learning strategy is employed to deftly control the gradient within the target domain, significantly enhancing the efficacy of positive transfer. The effectiveness and superiority of the proposed method are assessed through extensive experiments using datasets from variable working conditions. The experimental results demonstrate that the proposed method achieves satisfactory performance and outperforms state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

As an essential element in the overall designing of a service system, service planning plays an important role in improving final service quality and user experience. Service planning is a customer-oriented approach that facilitates the translation of customer requirements into activity process features or engineering characteristics. The trend is towards learning decision-support knowledge from massive service case data to facilitate service planning. However, when learning from imbalanced data, existing methods have poor predictive ability to identify minority patterns. To improve the quality and efficiency of the service planning, an innovative approach based on the combination of granular computing and deep learning are presented. It employs an inductive paradigm, clustering examples in a best granularity, sampling and refining into a more balanced example set, and feeding them into a deep learning model. The proposed approach can mine the planning patterns between the customer requirements and service process features, thereby facilitating knowledge transferring in service planning under the social manufacturing context. [ABSTRACT FROM AUTHOR]

There is an increasing unstructured text data produced in cross-enterprise social interaction media, forming a social interaction context that contains massive manufacturing relationships , which can be potentially used as decision support information for cross-enterprise manufacturing demand-capability matchmaking. How to enable decision-makers to capture these relationships remains a challenge. The text-based context contains high levels of noise and irrelevant information, causing both highly complexity and sparsity. Under this circumstance, instead of exploiting man-made features carefully optimized for the relationship extraction task, a deep learning model based on an improved stacked denoising auto-encoder on sentence-level features is proposed to extract manufacturing relationships among various named entities ( e.g. , enterprises, products, demands, and capabilities) underlying the text-based context. Experiment results show that the proposed approach can achieve a comparable performance with the state-of-the-art learning models, as well as a good practicality of its’ web-based implementation in social manufacturing interaction context. The ultimate goal of this study is to facilitate knowledge transferring and sharing in the context of enterprise social interaction, thereby supporting the integration of the resources and capabilities among different enterprises. [ABSTRACT FROM AUTHOR]

Printed Circuit Board (PCB) manufacturing is a kind of energy-intensive and pollution-intensive industries. With the increment of individualized requirements, PCB manufacturers face massive customized orders with a variety of specifications. The individualized customization on orders results in large differentials of the profit, energy consumption, and environmental pollution. Making energy-efficient order acceptance decisions can reduce carbon consumption and improve material utilization during the whole manufacturing process. An order acceptance decision model is established based on a loosely-coupled integration of deep learning and reinforcement learning techniques. Firstly, different from a simple assumption of the linear cost function in a small-scale manufacturing system, the deep learning algorithm is presented for accurately predicting the production cost, makespan, and carbon consumption of incoming PCB orders in the large-scale manufacturing system. Secondly, these predicted cleaner production indicators are combined with original order features to perform a reinforcement learning-based order acceptance decision. The proposed loosely-coupled deep reinforcement learning approach is verified with a dataset built based on data collected from a PCB manufacturer in China. This research is expected to provide an environment-friendly order acceptance decision-making approach for sustainable manufacturing in the Industry 4.0 context. [ABSTRACT FROM AUTHOR]

First-article inspection of flexographic printing labels (FPLs) is significant before the mass production of FPLs, which is manually implemented by quality check (QC) workers in real industries. In this paper, an adaptive coarse-to-fine framework is proposed for automatic first-article inspection of FPLs based on design drawings, consisting of coarse matching, fine matching, and defect detection. Specifically, a deep learning method integrating SuperPoint and SuperGlue is used for coarse matching between the design drawings and the corresponding FPL image to reveal their inherent characteristics. To deal with local deviations of the characters in the FPL image, a character-wise normalized cross-correlation (NCC) method is proposed to finely align each character in the design drawings with that in the FPL image, which utilizes different searching windows for different characters. Next, a constrained clustering method is proposed to accurately extract the contents of interest (COIs) in the FPL image. Finally, an edge distance measure with an adaptive threshold scheme is proposed to evaluate the COIs for inspection. Comparison experiments show that the proposed framework achieves better inspection performance for first-article inspection of FPLs than other fabric inspection methods. Specifically, it achieves the inspection performance of 1.3% missing rate, 2.0% false rate, and 0.951 Dice at an acceptable inspection speed of 2.735 s/pcs, which demonstrates that it is applicable to facilitate the QC task in real industries. [ABSTRACT FROM AUTHOR]