Time Series Classification using Deep Learning for Process Planning: A Case from the Process Industry

Mehdiyev, Nijat; Lahann, Johannes; Emrich, Andreas; Enke, David; Fettke, Peter; Loos, Peter

doi:10.1016/j.procs.2017.09.066

Cited by 92 publications

(43 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…• Quality inspection [6,51] • Fault diagnostic (detection, identification, estimation of magnitudes) •…”

Section: Deep Learningmentioning

confidence: 99%

“…Predictive analytics for defect prognosis [52,53] • Condition monitoring [54,55] • Service or operation planning [56,57] For example, Nijat et al [6] implemented stacked LSTM to extract features from time series data, which switches the intensive hand-crafted feature extraction way to a more automatic and intelligent way. Their network can also be used in process planning, monitoring the semi-finished products quality and determining the next process step.…”

Section: Deep Learningmentioning

confidence: 99%

“…Smart manufacturing brings such a profound impact to various industry sectors with broad applications including integrated circuit manufacturing, workload rescheduling, and production optimization. For example, modern computer processing techniques, such as deep learning, utilizes big data to characterize, model, and control production processes, leading to enhanced product quality and lowered risk of accidents [6][7][8][9][10][11]. Smart sensor techniques, like radio-frequency identification, enables tracking and identifying the position of a large number of objects accurately [12].…”

mentioning

confidence: 99%

See 2 more Smart Citations

Enhancing Sustainability and Energy Efficiency in Smart Factories: A Review

et al. 2018

View full text Add to dashboard Cite

With the rapid development of sensing, communication, computing technologies, and analytics techniques, today’s manufacturing is marching towards a new generation of sustainability, digitalization, and intelligence. Even though the significance of both sustainability and intelligence is well recognized by academia, industry, as well as governments, and substantial efforts are devoted to both areas, the intersection of the two has not been fully exploited. Conventionally, studies in sustainable manufacturing and smart manufacturing have different objectives and employ different tools. Nevertheless, in the design and implementation of smart factories, sustainability, and energy efficiency are supposed to be important goals. Moreover, big data based decision-making techniques that are developed and applied for smart manufacturing have great potential in promoting the sustainability of manufacturing. In this paper, the state-of-the-art of sustainable and smart manufacturing is first reviewed based on the PRISMA framework, with a focus on how they interact and benefit each other. Key problems in both fields are then identified and discussed. Specially, different technologies emerging in the 4th industrial revolution and their dedications on sustainability are discussed. In addition, the impacts of smart manufacturing technologies on sustainable energy industry are analyzed. Finally, opportunities and challenges in the intersection of the two are identified for future investigation. The scope examined in this paper will be interesting to researchers, engineers, business owners, and policymakers in the manufacturing community, and could serve as a fundamental guideline for future studies in these areas.

show abstract

“…• Quality inspection [6,51] • Fault diagnostic (detection, identification, estimation of magnitudes) •…”

Section: Deep Learningmentioning

confidence: 99%

Section: Deep Learningmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Sustainability and Energy Efficiency in Smart Factories: A Review

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Faced with the correlated relationship in a long time-step span, the standard RNNs structure is awkward with regards to the gradient problem. There are some important branches of RNNs that appear to solve the long-term problem, such as LSTM [5,20,21] and Gated Recurrent Unit (GRU) [32] structures. However, these redesigned complex structures require specific training algorithms.…”

Section: Gradient Problemmentioning

confidence: 99%

“…Actually, the original process is very brittle because it has little feasible access to further information over a long period of time. In recent years, some intricate techniques have led to impressive results, such as gradient clipping [18], modifying activation function [19], long short-term memory (LSTM) [20,21], and identity initialization [22]. However, few research efforts have attempted flexible RNNs designs in the area of time series.…”

Section: Introductionmentioning

confidence: 99%

Reconstruct Recurrent Neural Networks via Flexible Sub-Models for Time Series Classification

Chang

Lü

et al. 2018

Applied Sciences

View full text Add to dashboard Cite

Recurrent neural networks (RNNs) remain challenging, and there is still a lack of long-term memory or learning ability in sequential data classification and prediction. In this paper, we propose a flexible recurrent model, BIdirectional COnvolutional RaNdom RNNs (BICORN-RNNs), incorporating a series of sub-models: random projection, convolutional operation, and bidirectional transmission. These subcategories advance classification accuracy, which was limited by the gradient vanishing and the exploding problem. Experiments on public time series datasets demonstrate that our proposed method substantially outperforms a variety of existing models. Furthermore, the coordination of the accuracy and efficiency concerning a variety of factors, including SNR, length, data missing, and overlapping, is also discussed.

show abstract

Machine Learning in Solid‐State Chemistry: Heusler Compounds

Gzyl

Mar

Oliynyk

2021

Encyclopedia of Inorganic and Bioinorganic Chemistry

View full text Add to dashboard Cite

Machine learning attempts to find underlying trends in data and offer predictions of outcomes. When machine learning is applied to materials science, in a discipline called materials informatics, the complex relationships between composition, structure, and properties can be unraveled even when the quantity of data is limited. To illustrate this application, the large class of materials known as Heusler compounds are modeled through machine learning, enabling new candidates to be predicted or existing compounds to be screened for potentially interesting properties. Data, algorithms, and preprocessing techniques are important components of a successful machine‐learning model. Efforts to predict structures and properties of Heusler compounds are reviewed, and other machine‐learning approaches to discover materials in general are discussed. Ultimately, a machine‐learning model is only valuable if its predictions are validated by experimental results. Thus, perspectives are offered to guide experimentalists on how machine learning can be useful for targeting new Heusler compounds.

show abstract

Time Series Classification using Deep Learning for Process Planning: A Case from the Process Industry

Cited by 92 publications

References 16 publications

Enhancing Sustainability and Energy Efficiency in Smart Factories: A Review

Enhancing Sustainability and Energy Efficiency in Smart Factories: A Review

Reconstruct Recurrent Neural Networks via Flexible Sub-Models for Time Series Classification

Machine Learning in Solid‐State Chemistry: Heusler Compounds

Contact Info

Product

Resources

About