Search

In recent years, fault detection and diagnosis for industrial processes have been rapidly developed to minimize costs and maximize efficiency by taking advantages of cheap sensors and microprocessors, data analysis and artificial intelligence methods. However, due to the nonlinear and dynamic characteristics of industrial process data, the accuracy and efficiency of fault detection and diagnosis methods have always been an urgent problem in industry and academia. Therefore, this study proposes an adaptive fault detection and root-cause analysis scheme for complex industrial processes using moving window kernel principle component analysis (KPCA) and information geometric causal inference (IGCI). The proposed scheme has three main contributions. Firstly, a research scheme combining moving window KPCA with adaptive threshold is presented to handle the nonlinear and dynamic characteristics of complex industrial processes. Then, the multiobjective evolutionary algorithm is employed to select the optimal hyperparameters for fault detection, which not only avoids the blindness of hyperparameters selection, but also maximize model accuracy. Finally, the IGCI-based fault root-cause analysis method can help field operators to take corrective measures in time to resume the normal process. The proposed scheme is tested by the Tennessee Eastman platform. Its results show that this scheme has a good performance in reducing the faulty false alarms and missed detection rates and locating fault root-cause.

To clarify the causality among process parameters is a core issue of data-driven production performance analysis and product quality optimization. The difficulty lies in accurately measuring and distinguishing direct and indirect associations of complex manufacturing systems. In this work, the nonparametric-copula-entropy and network deconvolution method is proposed for causal discovery in complex manufacturing systems. Firstly, based on copula theory and kernel density estimation method, the nonparametric-copula-entropy is introduced to improve the accuracy of association measurement between parameters, and its superiority is verified by comparing with the results of different association measurement methods. Then, the global association matrix is constructed by the nonparametric-copula-entropy, and network deconvolution method is employed to extract the direct information from the global association matrix. The proposed method is tested by using an open gene expression dataset. Finally, as an experimental application, the causal analysis for a diesel engine production line is carried out by the proposed method. The results show that the proposed method can reveal causal relationship between process parameters and quality parameters in the diesel engine production line well, which provide theoretical guidance and implementation approach for the optimal control of complex manufacturing system.

Uncontrollable dendrite growth hinders the direct use of lithium metal anode in batteries even though it has the highest energy density of all anode materials. Li and N atoms have strong interaction and could form Li–N bond, promising for regulating Li-ion flux during the plating/stripping process. Herein, we successfully prepared ultrathin graphitic carbon nitride (g-C3N4) nanosheets with a thickness of ~ 1 nm and formed a g-C3N4 thin layer over the lithium metal anode. The abundant nitrogen species within g-C3N4 nanosheets could form Li–N bonds to powerfully stabilize the lithium-ion flux and enhance the affinity of electrodes with electrolytes. On top of that, the thin layer could act as an artificial solid electrolyte interface (SEI) to suppress lithium dendrite growth and enable stable Li plating/stripping over 350 h at a high current density of 5 mA cm−2 with a low overpotential of about 50 mV. The reported work demonstrates a promising strategy of the functional artificial SEI design for Li metal anodes.

With the widespread application of complex network theory, identifying vital nodes is an important part of complex network analysis, which has been a key issue in analyzing the characteristics of network structure and functions. Although many centrality measures have been proposed to identify vital nodes, there is still a lack of an objective performance testing platform for evaluating vital nodes identification algorithms. This study introduces the complex scheduling network as an objective performance testing platform where the identification of vital nodes directly determines the decision-making in each decision scenario through the development of heuristic algorithms upon centrality measures, and therefore, the optimization goals of scheduling problems can be regarded as an objective index for evaluating vital node identification algorithms. Finally, the undirected network obtained by the open shop scheduling problem is taken as an example to analyze and evaluate the performance of various vital node identification algorithms.

Establishing an effective early warning mechanism for the rocket final assembly process (RFAP) is crucial for the timely delivery of rockets and the reduction of additional production costs. To solve the unsystematic design of warning indicators and warning levels in RFAP and address the problem of low warning accuracy caused by imbalanced data distribution, this paper redesigns the warning indicators and warning levels in a systematic way, and develops a novel multiclass imbalanced learning method based on dynamic sampling algorithm (DyS) and improved ensemble neural network (IENN). The DyS algorithm dynamically determines the training set after oversampling the minority class, while the IENN can effectively suppress the oscillation in the iterative process of the DyS algorithm and improve the overall classification accuracy by removing the redundant and ineffective networks from the ensemble neural network. The experiment results indicate that the proposed method outperforms other methods in terms of accuracy and stability for early warning of tardiness in RFAP.