Xionglin Luo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Xionglin Luo has a total of 104 co-authors. Here are the top co-authors by number of shared publications:
Feng Xu
17 shared publicationsLin Sun
7 shared publications- Feng Xu
China University of Petroleum Beijing
5 shared publications Jia-Jiang Lin
Department of Automation, China University of Petroleum Bejing, Beijing, China
5 shared publications- Jian-Wei Liu
Department of Automation, College of Artificial intelligence, China University of Petroleum, Beijing, China
5 shared publications Rui Wang
Department of Automation, China University of Petroleum, Beijing 102249, China
4 shared publications
Latest Publications
- Open Access
Modeling and Evaluation of Attention Mechanism Neural Network Based on Industrial Time Series Data
Chemical process control systems are complex, and modeling the controlled object is the first task in automatic control and optimal design. Most chemical process modeling experiments require test signals to be applied to the process, which may lead to production interruptions or cause safety accidents. Therefore, this paper proposes an improved transformer model based on a self-attention mechanism for modeling industrial processes. Then, an evaluation mechanism based on root mean square error (RMSE) and Kullback–Leibler divergence (KLD) metrics is designed to obtain more appropriate model parameters. The Variational Auto-Encoder (VAE) network is used to compute the associated KLD. Finally, a real nonlinear dynamic process in the petrochemical industry is modeled and evaluated using the proposed methodology to predict the time series data of the process. This study demonstrates the validity of the proposed transformer model and illustrates the versatility of using an integrated modeling, evaluation, and prediction scheme for nonlinear dynamic processes in process industries. The scheme is of great importance for the field of industrial soft measurements as well as for deep learning-based time series prediction. In addition, the issue of a suitable time domain for the prediction is discussed.
Open AccessModeling and Evaluation of Attention Mechanism Neural Network Based on Industrial Time Series Data
Chemical process control systems are complex, and modeling the controlled object is the first task in automatic control and optimal design. Most chemical process modeling experiments require test signals to be applied to the process, which may lead to production interruptions or cause safety accidents. Therefore, this paper proposes an improved transformer model based on a self-attention mechanism for modeling industrial processes. Then, an evaluation mechanism based on root mean square error (RMSE) and Kullback–Leibler divergence (KLD) metrics is designed to obtain more appropriate model parameters. The Variational Auto-Encoder (VAE) network is used to compute the associated KLD. Finally, a real nonlinear dynamic process in the petrochemical industry is modeled and evaluated using the proposed methodology to predict the time series data of the process. This study demonstrates the validity of the proposed transformer model and illustrates the versatility of using an integrated modeling, evaluation, and prediction scheme for nonlinear dynamic processes in process industries. The scheme is of great importance for the field of industrial soft measurements as well as for deep learning-based time series prediction. In addition, the issue of a suitable time domain for the prediction is discussed.
- Open Access
Process Scheduling Analysis and Dynamic Optimization Maintaining the Operation Margin for the Acetylene Hydrogenation Fixed-Bed Reactor
The full-cycle operation optimization of the acetylene hydrogenation reactor should strictly adhere to the operation optimization scheme within the operation cycle, regardless of scheduling changes. However, in actual industrial processes, in order to meet temporary process scheduling requirements, the acetylene hydrogenation reactor needs to adjust its operation strategy temporarily within the remaining operation cycle based on the results of dynamic optimization for a certain period. It brings additional challenges and a research gap to the operational optimization problem. To make up for this research gap, this paper focuses on researching a type of full-cycle dynamic optimization problem where the operation optimization scheme is temporarily adjusted during the operation cycle. The methods employed for changing the operation optimization scheme include modifying the operation cycle, maximizing economic benefits, and altering the optimization goal to maximize the operation cycle. A novelty full-cycle scheduling optimization framework based on surplus margin estimate is proposed to build a platform for these methods. The paper analyzes the impact of process scheduling changes on full-cycle optimization using a dynamic optimization model that maintains the operation margin. It establishes a full-cycle scheduling optimization model and obtains the optimal scheduling strategy by a novelty method NSGBD (non-convex sensitivity-based generalized Benders decomposition). In this process, an adaptive CVP (control vector parameterization) based on a decomposition optimization algorithm is proposed, which tackles the challenge of optimizing complex acetylene hydrogenation reactor models on a large time scale. Scheduling optimization can be realized as an annualized benefit of 1.56 × 106 and 1.57 × 106 ¥ separately within two scheduling optimization constraints, and the computational time required is much less than previous operational optimizations.
Open AccessProcess Scheduling Analysis and Dynamic Optimization Maintaining the Operation Margin for the Acetylene Hydrogenation Fixed-Bed Reactor
The full-cycle operation optimization of the acetylene hydrogenation reactor should strictly adhere to the operation optimization scheme within the operation cycle, regardless of scheduling changes. However, in actual industrial processes, in order to meet temporary process scheduling requirements, the acetylene hydrogenation reactor needs to adjust its operation strategy temporarily within the remaining operation cycle based on the results of dynamic optimization for a certain period. It brings additional challenges and a research gap to the operational optimization problem. To make up for this research gap, this paper focuses on researching a type of full-cycle dynamic optimization problem where the operation optimization scheme is temporarily adjusted during the operation cycle. The methods employed for changing the operation optimization scheme include modifying the operation cycle, maximizing economic benefits, and altering the optimization goal to maximize the operation cycle. A novelty full-cycle scheduling optimization framework based on surplus margin estimate is proposed to build a platform for these methods. The paper analyzes the impact of process scheduling changes on full-cycle optimization using a dynamic optimization model that maintains the operation margin. It establishes a full-cycle scheduling optimization model and obtains the optimal scheduling strategy by a novelty method NSGBD (non-convex sensitivity-based generalized Benders decomposition). In this process, an adaptive CVP (control vector parameterization) based on a decomposition optimization algorithm is proposed, which tackles the challenge of optimizing complex acetylene hydrogenation reactor models on a large time scale. Scheduling optimization can be realized as an annualized benefit of 1.56 × 106 and 1.57 × 106 ¥ separately within two scheduling optimization constraints, and the computational time required is much less than previous operational optimizations.
- Open Access
Online Dynamic Optimization of Multi-Rate Processes with the Case of a Fluid Catalytic Cracking Unit
Due to operational limitations in the industrial field, the operating variables of fluid catalytic cracking units (FCCU) are of multiple operating frequencies, which are CO combustion promoter amount, recycle slurry flow rate, combustion air flow rate, heat escape, and reaction temperature, from low frequency to high frequency. There are usually two schemes for operation optimization of FCCU. The former is called single-rate, single-window optimization, whose operating variables are optimized only once in the whole operation cycle, which is easy to achieve, but the optimization effect is poor. The latter is called single-rate multi-window optimization, whose operating variables are optimized repeatedly and whose operation cycle is discretized into multiple optimization periods with the same frequency, which costs a heavy calculation burden and cannot adapt to the optimization variables with multiple operating frequencies. So, a multi-rate, variable-window online dynamic optimization method is proposed. In an operation cycle, the high-frequency operating variable is optimized in a short optimization period, and the low-frequency operating variable is optimized in a long optimization period; each optimization period has integer multiples to the minimum optimization period. Each optimized result for each optimization period is put into use online immediately. The optimization model involves the time domain differential equations, integral cost objective function, and measured disturbances. The experimental results show that compared with the single-rate, single-window optimization method and single-rate multi-window optimization method, the optimization effect of multi-rate, variable-window online dynamic optimization is better than single-rate, single-window optimization but worse than single-rate multi-window optimization. However, the optimization results are consistent with the operation frequency of each optimization variable, which can be implemented in complex chemical processes and increase certain economic benefits.
Open AccessOnline Dynamic Optimization of Multi-Rate Processes with the Case of a Fluid Catalytic Cracking Unit
Due to operational limitations in the industrial field, the operating variables of fluid catalytic cracking units (FCCU) are of multiple operating frequencies, which are CO combustion promoter amount, recycle slurry flow rate, combustion air flow rate, heat escape, and reaction temperature, from low frequency to high frequency. There are usually two schemes for operation optimization of FCCU. The former is called single-rate, single-window optimization, whose operating variables are optimized only once in the whole operation cycle, which is easy to achieve, but the optimization effect is poor. The latter is called single-rate multi-window optimization, whose operating variables are optimized repeatedly and whose operation cycle is discretized into multiple optimization periods with the same frequency, which costs a heavy calculation burden and cannot adapt to the optimization variables with multiple operating frequencies. So, a multi-rate, variable-window online dynamic optimization method is proposed. In an operation cycle, the high-frequency operating variable is optimized in a short optimization period, and the low-frequency operating variable is optimized in a long optimization period; each optimization period has integer multiples to the minimum optimization period. Each optimized result for each optimization period is put into use online immediately. The optimization model involves the time domain differential equations, integral cost objective function, and measured disturbances. The experimental results show that compared with the single-rate, single-window optimization method and single-rate multi-window optimization method, the optimization effect of multi-rate, variable-window online dynamic optimization is better than single-rate, single-window optimization but worse than single-rate multi-window optimization. However, the optimization results are consistent with the operation frequency of each optimization variable, which can be implemented in complex chemical processes and increase certain economic benefits.
- Open Access
- 26 March 2023
- Published by Wiley
Full cycle on-line autonomous reconfiguration control for heat exchanger networks
This study presents a novel autonomous reconfiguration strategy to achieve full cycle optimal control of the heat exchanger networks. The proposed method aims to select the optimal control structure online to solve the problem of slow time-varying process transition caused by the inevitable fouling accumulation. Specifically, an auxiliary variable is introduced to increase the degree of control freedom, and a flexible control structure is studied to provide appropriate pairings dynamically between variables. Subsequently, the nonparametric Relative Energy Gain Array is considered as the switching mechanism to determine switching instants of control loops. Furthermore, the online optimization of switching regions ensures the optimal control performance. Finally, one typical experiment for the heat exchanger network in crude distillation unit was conducted and the results illustrated the effectiveness of the proposed control method. Compared with the conventional control system, the available service time was extended by nearly 18 months.
- 26 March 2023
- Published by Wiley
Open AccessFull cycle on-line autonomous reconfiguration control for heat exchanger networks
This study presents a novel autonomous reconfiguration strategy to achieve full cycle optimal control of the heat exchanger networks. The proposed method aims to select the optimal control structure online to solve the problem of slow time-varying process transition caused by the inevitable fouling accumulation. Specifically, an auxiliary variable is introduced to increase the degree of control freedom, and a flexible control structure is studied to provide appropriate pairings dynamically between variables. Subsequently, the nonparametric Relative Energy Gain Array is considered as the switching mechanism to determine switching instants of control loops. Furthermore, the online optimization of switching regions ensures the optimal control performance. Finally, one typical experiment for the heat exchanger network in crude distillation unit was conducted and the results illustrated the effectiveness of the proposed control method. Compared with the conventional control system, the available service time was extended by nearly 18 months.
- 8 February 2023
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- Vol. 28 (04) , 1594-1610
Improving Drilling Efficiency and Safety Based on Hydraulic Mechanical Specific Energy in the Framework of Economic Model Predictive Control
Summary: The economic value of real-time optimization of the drilling process has attracted extensive attention in the period of crude oil price fluctuation. Because of the complicated bottomhole conditions and coupling relationship among subsystems, it is difficult for the driller to continuously adjust the primary manipulation variables—weight on bit (WOB), revolutions per minute (RPM), and managed pressure drilling (MPD) equipment—which leads to unsatisfactory drilling effect and increased nonproductive time. Here, we propose an integrated control strategy based on economic model predictive control with zone tracking (ZoneEMPC), nonlinear model predictive control (NMPC), and Lyapunov-based model predictive control (LMPC) to optimize hydraulic mechanical specific energy (HMSE) while stabilizing bottomhole pressure (BHP). During normal drilling, ZoneEMPC is used to minimize HMSE through operating RPM, WOB, mud pump flow, and choke-valve opening as well as its tracking performance, which ensures that the BHP is maintained within the pressure window. Moreover, the controller is switched to NMPC, which adjusts the BHP to quickly reach the new setting zone. The reason for controller switching is that the tracking accuracy of ZoneEMPC could decline when encountering high-pressure formation with gas invasion. As the BHP reaches the new pressure range, the controller is switched to ZoneEMPC again and the HMSE is reoptimized. In addition, the transmission rate of mud pulse telemetry is not enough to meet the real-time control in practical engineering. Therefore, the LMPC controller is used to ensure the closed-loop stability of the system when the downhole measurement data (e.g., BHP) delayed. This study uses the drilling data of a vertical well located in Tarim, China, to verify the control strategy. The results indicated that the proposed integrated control strategy could improve the rate of penetration (ROP), especially in a formation with high rock strength. Meanwhile, the BHP could be adjusted efficiently and stably under different conditions. This work provides a unified framework integrating multiple system models for practical engineering, which has certain theoretical guiding significance for the automation and intelligent development of the oil and gas industry.
- 8 February 2023
- Published by Society of Petroleum Engineers (SPE) in SPE Journal
- Vol. 28 (04) , 1594-1610
Improving Drilling Efficiency and Safety Based on Hydraulic Mechanical Specific Energy in the Framework of Economic Model Predictive Control
Summary: The economic value of real-time optimization of the drilling process has attracted extensive attention in the period of crude oil price fluctuation. Because of the complicated bottomhole conditions and coupling relationship among subsystems, it is difficult for the driller to continuously adjust the primary manipulation variables—weight on bit (WOB), revolutions per minute (RPM), and managed pressure drilling (MPD) equipment—which leads to unsatisfactory drilling effect and increased nonproductive time. Here, we propose an integrated control strategy based on economic model predictive control with zone tracking (ZoneEMPC), nonlinear model predictive control (NMPC), and Lyapunov-based model predictive control (LMPC) to optimize hydraulic mechanical specific energy (HMSE) while stabilizing bottomhole pressure (BHP). During normal drilling, ZoneEMPC is used to minimize HMSE through operating RPM, WOB, mud pump flow, and choke-valve opening as well as its tracking performance, which ensures that the BHP is maintained within the pressure window. Moreover, the controller is switched to NMPC, which adjusts the BHP to quickly reach the new setting zone. The reason for controller switching is that the tracking accuracy of ZoneEMPC could decline when encountering high-pressure formation with gas invasion. As the BHP reaches the new pressure range, the controller is switched to ZoneEMPC again and the HMSE is reoptimized. In addition, the transmission rate of mud pulse telemetry is not enough to meet the real-time control in practical engineering. Therefore, the LMPC controller is used to ensure the closed-loop stability of the system when the downhole measurement data (e.g., BHP) delayed. This study uses the drilling data of a vertical well located in Tarim, China, to verify the control strategy. The results indicated that the proposed integrated control strategy could improve the rate of penetration (ROP), especially in a formation with high rock strength. Meanwhile, the BHP could be adjusted efficiently and stably under different conditions. This work provides a unified framework integrating multiple system models for practical engineering, which has certain theoretical guiding significance for the automation and intelligent development of the oil and gas industry.
- Open Access
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (00) , 47
Explaining deep residual networks predictions with symplectic adjoint method
Understanding deep residual networks (ResNets) decisions are receiving much attention as a way to ensure their security and reliability. Recent research, however, lacks theoretical analysis to guarantee the faithfulness of explanations and could produce an unreliable explanation. In order to explain ResNets predictions, we suggest a provably faithful explanation for ResNet using a surrogate explainable model, a neural ordinary differential equation network (Neural ODE). First, ResNets are proved to converge to a Neural ODE and the Neural ODE is regarded as a surrogate model to explain the decision-making attribution of the ResNets. And then the decision feature and the explanation map of inputs belonging to the target class for Neural ODE are generated via the symplectic adjoint method. Finally, we prove that the explanations of Neural ODE can be sufficiently approximate to ResNet. Experiments show that the proposed explanation method has higher faith fulness with lower computational cost than other explanation approaches and it is effective for troubleshooting and optimizing a model by the explanation.
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (00) , 47
Open AccessExplaining deep residual networks predictions with symplectic adjoint method
Understanding deep residual networks (ResNets) decisions are receiving much attention as a way to ensure their security and reliability. Recent research, however, lacks theoretical analysis to guarantee the faithfulness of explanations and could produce an unreliable explanation. In order to explain ResNets predictions, we suggest a provably faithful explanation for ResNet using a surrogate explainable model, a neural ordinary differential equation network (Neural ODE). First, ResNets are proved to converge to a Neural ODE and the Neural ODE is regarded as a surrogate model to explain the decision-making attribution of the ResNets. And then the decision feature and the explanation map of inputs belonging to the target class for Neural ODE are generated via the symplectic adjoint method. Finally, we prove that the explanations of Neural ODE can be sufficiently approximate to ResNet. Experiments show that the proposed explanation method has higher faith fulness with lower computational cost than other explanation approaches and it is effective for troubleshooting and optimizing a model by the explanation.
- 21 December 2022
- Published by Wiley in Asia-Pacific Journal of Chemical Engineering
- Vol. 18 (2)
Novel feedforward‐feedback‐based continuous process transition strategy for chemical industry
As a typical regulation means to realize a large‐scale transition of the system condition, the lack of feedback compensation in the transition strategy leads to the fact that it has no moderating effect on interference. In order to ensure the safe and smooth operation of the system, it is necessary to coordinate the transition and control, that is, to achieve the process transition through a closed‐loop control structure. The existing coordination scheme can meet this demand, but there is a defect of poor anti‐interference. To this end, a novel coordination scheme integrating transition with control is proposed in the paper, which alleviates the burden on the feedback control system by adding the manipulated variable reference trajectory. When interferences occur, the proposed coordination scheme can positively respond and compensate for the negative influence of disturbances, allowing the system to quickly return to the desired state. Besides, the corresponding analytical solution method is presented for the new coordination scheme, which avoids the complication of the conventional optimization solution method. Finally, the advantage of our coordination scheme has been effectively verified through the feed fluctuation ethylene column.
- 21 December 2022
- Published by Wiley in Asia-Pacific Journal of Chemical Engineering
- Vol. 18 (2)
Novel feedforward‐feedback‐based continuous process transition strategy for chemical industry
As a typical regulation means to realize a large‐scale transition of the system condition, the lack of feedback compensation in the transition strategy leads to the fact that it has no moderating effect on interference. In order to ensure the safe and smooth operation of the system, it is necessary to coordinate the transition and control, that is, to achieve the process transition through a closed‐loop control structure. The existing coordination scheme can meet this demand, but there is a defect of poor anti‐interference. To this end, a novel coordination scheme integrating transition with control is proposed in the paper, which alleviates the burden on the feedback control system by adding the manipulated variable reference trajectory. When interferences occur, the proposed coordination scheme can positively respond and compensate for the negative influence of disturbances, allowing the system to quickly return to the desired state. Besides, the corresponding analytical solution method is presented for the new coordination scheme, which avoids the complication of the conventional optimization solution method. Finally, the advantage of our coordination scheme has been effectively verified through the feed fluctuation ethylene column.
- 29 November 2022
- Published by Springer Nature in Journal of Global Optimization
- Vol. 86 (1) , 37-60
Nonconvex sensitivity-based generalized Benders decomposition
- 29 November 2022
- Published by Springer Nature in Journal of Global Optimization
- Vol. 86 (1) , 37-60
Nonconvex sensitivity-based generalized Benders decomposition