The gas utilization ratio (GUR) in a blast furnace is directly linked to the efficiency, cost-effectiveness and environmental impact on the blast furnace ironmaking process. However, The stability of the published GUR...
详细信息
ISBN:
(纸本)9798350387780;9798350387797
The gas utilization ratio (GUR) in a blast furnace is directly linked to the efficiency, cost-effectiveness and environmental impact on the blast furnace ironmaking process. However, The stability of the published GUR time series prediction models need to be improved. This paper presents an improved particle swarm optimization (PSO) incorporating linearly decreasing inertia weights (LDIW) to optimize the kernel-based extreme learning machine (KELM) for single-step prediction. This paper uses singular spectrum analysis (SSA) to preprocess the data and extract the key components from the GUR time series to solve the problem of high volatility of the GUR time series. In addition, this paper introduces LDIW to improve the optimization ability of particle swarm optimization algorithm, which enhances the stability of a single-step prediction model. Then this paper uses the improved PSO algorithm to extract the optimal parameters of KELM, and establishes a single-step GUR prediction model based on the improved PSO-KELM. Finally, this paper uses the actual production processdata of blast furnace to verify the prediction model. The results show that the prediction accuracy of GUR and the overall stability of the model are significantly improved, providing important guidance for the blast furnace ironmaking process.
Modern communities are one of the most important energy consumers in the city energy system, offering substantial potential for active participation in demand response services. In this paper, we propose an energy sch...
详细信息
ISBN:
(纸本)9798350382662;9798350382655
Modern communities are one of the most important energy consumers in the city energy system, offering substantial potential for active participation in demand response services. In this paper, we propose an energy scheduling and control framework for grid-interactive communities, integrating a physically consistent deep learning (PCDL) model for thermal dynamic approximation. The construction of the PCDL model relies on open-loop simulation data, with its physical consistency assured through parameter constraints and a specialized two-layer model structure. Subsequently, the PCDL model is utilized as the prediction model in both the scheduling and control systems. The energy dispatch plan is derived through a day-ahead scheduling process and then integrated into the model predictive control (MPC) system to facilitate real-time energy management and thermal comfort regulation. A simulation case is presented in this paper to verify the performance of the proposed modeling and control methods. According to the simulation results, the PCDL model exhibits superior control-oriented generalization ability compared to the other data-driven models. Furthermore, the PCDL-based scheduling and control framework outperforms other controllers in maintaining indoor thermal comfort and achieves at least a 29.5% reduction in energy expenditure compared to the baseline controller.
Deep learning methods can extract reliable feature representations from massive processdata to build accurate soft sensor models. However, the data in actual industrial production is often nonlinear, dynamic, and eve...
详细信息
The design of active controllers for vibration attenuation in mechanical systems depends on an accurate identification of the system. This task becomes very challenging in nonlinear systems due to the complexity and v...
详细信息
Mathematical model is the core of the digital twin basin, which can simulate the whole elements of the physical basin and the whole process of water conservancy governance and management activities, and provide intell...
详细信息
For the safety of automatic industry processing, the prediction and analysis of abnormal conditions of field equipment fast and accurately is of great significance. The KPCA-RF fault diagnosis approach, which is based...
详细信息
ISBN:
(纸本)9798350334722
For the safety of automatic industry processing, the prediction and analysis of abnormal conditions of field equipment fast and accurately is of great significance. The KPCA-RF fault diagnosis approach, which is based on the kernel principal component analysis (KPCA) and the random forest (RF), is developed in this study to address the issues with the mainstream fault diagnosis method in the Tennessee-Eastman (TE) process. High-dimensional raw data must be feature extracted with KPCA before time series based nonlinear feature data can be obtained. The RF method is then used to determine the fault type of the feature data. The accuracy of this diagnosis method has been proved through comparative experiments.
The equipment digital twins (EDTs) for discrete manufacturing should be calibrated quickly to avoid irreversible physical damage to the equipment caused by biased control commands. Therefore, an online credibility ass...
详细信息
The equipment digital twins (EDTs) for discrete manufacturing should be calibrated quickly to avoid irreversible physical damage to the equipment caused by biased control commands. Therefore, an online credibility assessment method for EDTs is urgently needed. However, existing assessment approaches consume too much time, and thus could not reveal dynamic faults in time. In this paper, the dynamic relationship between online evolution and actual applications of EDTs is investigated. Then, two steps are proposed to accelerate the assessment process significantly. One involves pre-constructing a performance-deviation-agent (PDA), and the other involves dynamically fitting the application-time-window (ATW) probability distribution. The methodology is applicable to discrete manufacturing processes. The dynamic credibility of EDT evolution process can be updated after every iteration of the model evolution. Sorting manufacturing equipment was used as a case study to demonstrate the effectiveness of this method. The time consumption was reduced by 90% compared with traditional assessment methods in the case. Note to Practitioners-The application of digital twins (DTs) for discrete manufacturing equipment typically aims to enhance the control effectiveness of the physical equipment. To ensure the effectiveness of control signals derived from DT-based simulations, the credibility of the equipment DTs must be known and guaranteed. This paper addresses the quick updating of the credibility of equipment DTs as they evolve according to real-time data collected from the equipment. The proposed methods are currently applicable to individual pieces of equipment with straightforward workflows. By adapting the training method for the error prediction model to scenarios involving multiple pieces of equipment and more complex workflows, the overall control efficacy within a workshop can be significantly improved.
In recent years, modeling of nonlinear systems has increasingly involved machine learning (ML). Recurrent neural networks (RNNs), a type of supervised learning technique, have shown to be effective in modeling time se...
详细信息
ISBN:
(纸本)9798350315431
In recent years, modeling of nonlinear systems has increasingly involved machine learning (ML). Recurrent neural networks (RNNs), a type of supervised learning technique, have shown to be effective in modeling time series data. Particularly, it has been demonstrated in several works that physics-informed RNN models (where the network structure is informed by the pattern of interactions of physical process variables) are preferable to dense RNN models. Motivated by this, the present work focuses on the generalization error of partially-connected RNN models and its relationship to the corresponding error of fully-connected RNN models for the same training and testing data sets. The RNN models are subsequently used in model predictive control of nonlinear processes. It is found that the generalization error bounds for the partially-connected RNN models are lower than that of the fully-connected RNN models, and a comparison study using a chemical process example is conducted to demonstrate these results.
The connected car services are one of the most widely used services in the Internet of Things environment, and they provide numerous services to existing vehicles by connecting them through networks inside and outside...
详细信息
The connected car services are one of the most widely used services in the Internet of Things environment, and they provide numerous services to existing vehicles by connecting them through networks inside and outside the vehicle. However, although vehicle manufacturers are developing services considering the means to secure the connected car services, concerns about the security of the connected car services are growing due to the increasing number of attack cases. In this study, we reviewed the research related to the connected car services that have been announced so far, and we identified the threats that may exist in the connected car services through security threat modeling to improve the fundamental security level of the connected car services. As a result of performing the test to the applications for connected car services developed by four manufacturers, we found that all four companies' applications excessively requested unnecessary permissions for application operation, and the apps did not obfuscate the source code. Additionally, we found that there were still vulnerabilities in application items such as exposing error messages and debugging information.
Deep learning has been widely used in industrial processes, which automatically learns hidden knowledge from processdata and detects quality variables that are difficult to measure. Therefore, accurate extraction and...
详细信息
ISBN:
(纸本)9798350334722
Deep learning has been widely used in industrial processes, which automatically learns hidden knowledge from processdata and detects quality variables that are difficult to measure. Therefore, accurate extraction and utilization of effective features and elimination of useless features are still one of the most important research issues in soft sensor modeling. In this paper, an attention-based gated supervised encoder-decoder BiLSTM (AGSED-BiLSTM) is proposed. AGSED-BiLSTM first encodes input variables to obtain feature expressions of different abstract levels with high correlation with quality parameters, and links the outputs of each layer together through gating neurons, so as to make full use of the information of different hidden layers to obtain the final output, and improves learning efficiency through bidirectional architecture. Finally, the proposed model is applied to penicillin fermentation process, which proves the effectiveness and superiority of the proposed soft sensor model based on AGSED-BiLSTM network, and is superior to the most advanced and traditional soft sensor models based on deep learning.
暂无评论