Enhancing Generalizability of Predictive Models with Synergy of Data and Physics

Wind farm needs prediction models for predictive maintenance. There is a need to predict values of non-observable parameters beyond ranges reflected in available data. A prediction model developed for one machine many not perform well in another similar machine. This is usually due to lack of generalizability of data-driven models. To increase generalizability of predictive models, this research integrates the data mining with first-principle knowledge. Physics-based principles are combined with machine learning algorithms through feature engineering, strong rules and divide-and-conquer. The proposed synergy concept is illustrated with the wind turbine blade icing prediction and achieves significant prediction accuracy across different turbines. The proposed process is widely accepted by wind energy predictive maintenance practitioners because of its simplicity and efficiency. Furthermore, the testing scores of KNN, CART and DNN algorithm are increased by 44.78%, 32.72% and 9.13% with our proposed process. We demonstrated the importance of embedding physical principles within the machine learning process, and also highlight an important point that the need for more complex machine learning algorithms in industrial big data mining is often much less than it is in other applications, making it essential to incorporate physics and follow “Less is More” philosophy.

Handling Imbalanced Datasets for Robust Deep Neural Network-Based Fault Detection in Manufacturing Systems

Over the recent years, Industry 4.0 (I4.0) technologies such as the Industrial Internet of Things (IIoT), Artificial Intelligence (AI), and the presence of Industrial Big Data (IBD) have helped achieve intelligent Fault Detection (FD) in manufacturing. Notably, data-driven approaches in FD apply Deep Learning (DL) techniques to help generate insights required for monitoring complex manufacturing processes. However, due to the ratio of instances where actual faults occur, FD datasets tend to be imbalanced, leading to training challenges that result in inefficient DL-based FD models. In this paper, we propose Dual Logits Weights Perturbation (DLWP) loss, a method featuring weight vectors for improved dataset generalization in FD systems. The weight vectors act as hyperparameters adjusted on a case-by-case basis to regulate focus accorded to individual minority classes during training. In particular, our proposed method is suitable for imbalanced datasets from safety-related FD tasks as it generates DL models that minimize false negatives. Subsequently, we integrate human experts into the workflow as a strategy to help safeguard the system. A subset of the results, model predictions with uncertainties exceeding a preset threshold, are considered a preliminary output subject to cross-checking by human experts. We demonstrate that DLWP achieves improved Recall, AUC, F1 scores.

Design of Management Platform Architecture and Key Algorithm for Massive Monitoring Big Data

With the construction and development of industrial informatization, industrial big data has become a trend within the smart industry. To obtain valuable information on massive data, achieving the acquisition, storage, analysis, and mining is becoming an important area of research. Focusing on the application requirements for industrial fields, we propose a data acquisition and analysis system based on the NB-IoT for industrial applications. The system is an integrated system that includes sensor data acquisition, data transmission, data storage, and analysis mining. In this study, we mainly focused on the use of the NB-IoT network to collect and transmit real-time data for sensors. First, for the long time series (e.g., if we collect the data streams for one year for the sensor with a frequency of 1 Hz, the length of the series will reach 107). Then, we propose DSCS-LTS, a distributed storage and calculation model, and CCCA-LTS, an algorithm for the correlation coefficient of long time series in a distributed environment. Third, we propose a granularity selection algorithm and query process logic for visualization. We tested the platform in our laboratory and an automated production line for one year, and the experimental results using real data sets show that our approach is effective and scalable, can achieve efficient data management, and provide the basis for intelligent enterprise decision-making.

Industrial Big Data Analytics and Cyber-Physical Systems for Future Maintenance & Service Innovation

Based on Information and Communication Technologies (ICT) fast advancement and the integration of advanced analytics into manufacturing, products, and services, several industries face new opportunities and at the identical time challenges of maintaining their ability and market desires. Such integration, that is termed Cyber-physical Systems (CPS), is remodeling the industry into a future level. CPS facilitates the systematic conversion of big data into information that reveals invisible patterns of deterioration and inefficiency and leads to better decision-making. This project focuses on existing trends within the development of industrial huge information analytics and CPS. Then it, in brief, discusses a system architecture for applying CPS in manufacturing referred to as 5C. The 5C architecture, comprises necessary steps to totally integrate cyber-physical systems within the manufacturing industry.

The Conundrum of Talent Training of Industrial Big DataAnalysisin China's E-Commerce Talent Cultivating System

With the development of information technology, Industrial Big Data Analysis (BDA) has become the most important link in China's E-commerce, but the development of BDA in China's E-commerce has been greatly restricted by talents. The main reason is that the specialty of Industrial BDA and the practicality of E-commerce are difficult to be well integrated in China's existing education system, which is due to the very special development history of E-commerce in China. The training of talents for E-commerce BDA has become a structural problem in China.

Several Typical Paradigms of Industrial Big data Application

Industrial big data is an important part of big data family, which has important application value for industrial production scheduling, risk perception, state identification, safety monitoring and quality control, etc. Due to the particularity of the industrial field, some concepts in the existing big data research field are unable to reflect accurately the characteristics of industrial big data, such as what is industrial big data, how to measure industrial big data, how to apply industrial big data, and so on. In order to overcome the limitation that the existing definition of big data is not suitable for industrial big data, this paper intuitively proposes the concept of big data cloud and the 3M (Multi-source, Multi-dimension, Multi-span in time) definition of cloud-based big data. Based on big data cloud and 3M definition, three typical paradigms of industrial big data applications are built, including the fusion calculation paradigm, the model correction paradigm and the information compensation paradigm. These results are helpful for grasping systematically the methods and approaches of industrial big data applications.

An Analytics Environment Architecture for Industrial Cyber-Physical Systems Big Data Solutions

The architecture design of industrial data analytics system addresses industrial process challenges and the design phase of the industrial Big Data management drivers that consider the novel paradigm in integrating Big Data technologies into industrial cyber-physical systems (iCPS). The goal of this paper is to support the design of analytics Big Data solutions for iCPS for the modeling of data elements, predictive analysis, inference of the key performance indicators, and real-time analytics, through the proposal of an architecture that will support the integration from IIoT environment, communications, and the cloud in the iCPS. An attribute driven design (ADD) approach has been adopted for architectural design gathering requirements from smart production planning, manufacturing process monitoring, and active preventive maintenance, repair, and overhaul (MRO) scenarios. Data management drivers presented consider new Big Data modeling analytics techniques that show data is an invaluable asset in iCPS. An architectural design reference for a Big Data analytics architecture is proposed. The before-mentioned architecture supports the Industrial Internet of Things (IIoT) environment, communications, and the cloud in the iCPS context. A fault diagnosis case study illustrates how the reference architecture is applied to meet the functional and quality requirements for Big Data analytics in iCPS.

A Semi-Supervised Multiclass Anomaly Detection Approach for Partially Labeled In-Process Measurement Data

Machine learning and other data-driven methods have developed at a prolific rate for industrial applications due to the advent of industrial big data. However, industrial datasets may not be especially well-suited to supervised learning approaches that require extensive domain knowledge in the complete and accurate labeling of datasets. To address these challenges, a semi-supervised learning approach is proposed that makes use of partially labeled subsets. The proposed methodology is applied to high-dimensional in-process measurement data, utilizing a convolutional autoencoder for unsupervised feature extraction. A multiclass extension for semi-supervised anomaly diagnosis is proposed that utilizes principal component analysis as the basis for anomaly scoring, and the proposed approach intersects the results of targeted one-against-all phases on partially labeled sets to classify faults. Experiments in a case study on semiconductor manufacturing measurement data are performed to explore the relationship between latent features extracted and anomaly detection performance. The application of the proposed algorithm achieves a true positive detection rate of over 90% with false positive rate under 9% for both local and global anomaly types, with these results accomplished while reducing over 99% of the original input data dimensions. In addition, the approach also allows for positive samples to be identified that were previously undetected by human experts. These results are promising for the application of the proposed semi-supervised methodology in real industrial settings.