Industrial Process Control Research Articles

Learn-to-supervise: Causal reinforcement learning for high-level control in industrial processes

Possessing efficient supervisory control systems is crucial for maintaining the desired operational performance of complex industrial processes. Several challenges face the developers of these systems, such as requiring accurate physical models, dealing with the variability and uncertainty of process operating conditions and coordinating between local controllers to reach desired global performance. This paper proposes an intelligent supervisory control approach based on causal reinforcement learning (CRL) to effectively manipulate the controllers’ setpoints of the process in a way that optimizes its key performance indicators (KPIs), thereby improving the energy efficiency of the process. The approach adopts deep reinforcement learning (DRL) to develop an efficient control policy through interaction with a process simulation. The DRL training history is then exploited using interpretable machine learning and process mining to build a discrete event system (DES) model, in the form of a state-event graph. The DES model identifies causal relationships between events and provides interpretability to the control policy developed by the DRL method. The DES discovered is exploited as a Markov decision process to apply the Q-learning algorithm as a CRL supervisor. The supervisor incorporates causal knowledge into its training process, thus improving the DRL control policy developed and identifying the event paths that optimize the process’s KPIs. The proposed approach is validated using two heat recovery systems in a pulp & paper mill. It successfully achieves a control policy that reduces energy consumption by up to 15.6% for the first system and 5.02% for the second, compared to the expert’s baseline methods.

Automatic decomposition of large-scale industrial processes for distributed MPC on the Shell–Yokogawa Platform for Advanced Control and Estimation (PACE)

The kernel of industrial advanced process control (APC) lies in the formulation and solution of model predictive control (MPC) problems, which specify the controller moves according to the solution of an optimal control problem at each sampling time. A significant challenge is the online computation for large-scale industrial systems. As the state-of-the-art APC technology, the Shell–Yokogawa Platform for Advanced Control and Estimation (PACE) has adopted a systematic framework of handling dynamic optimization of large-scale systems, where an automatic decomposition procedure generates subsystems for distributed MPC. The decomposition is implemented on network representations of the MPC models that capture interactions among process variables, with community detection used to maximize the statistical significance of the subnetworks with preferred internal interconnections. This paper introduces the fundamentals of such a decomposition approach and this functionality in PACE, followed by a case study on a crude distillation process to showcase its industrial application.

State‐of‐the‐Art Review on the Applications of Nonlinear and Artificial Intelligence‐Based Controllers in Petrochemical Processes

AbstractPetrochemical industries are facing severe challenges in controlling nonlinear processes which are to be analyzed and monitored to develop efficient control models for improved efficiency and productivity. This review focuses on the control theory of petrochemical processes along with linear and nonlinear controllers with their limitations as well as on the application of artificial intelligence (AI)‐based controllers in multiple sectors of petrochemical industries. The findings and results from the review prove that the AI control technique is a promising one in the future for uncertain dynamic systems. The paper also covers a review of the vinyl chloride monomer (VCM) process and a recent control study on the plant. Being a nonlinear system, VCM requires advanced control techniques to control the uncertain dynamics of the process. The study shows a gap in the development of control strategies applied to VCM in terms of sustainability, efficiency, and environment‐concerned issues. Valuable insights for researchers, practitioners, and engineers engaged in process control and automation in the petrochemical industry are provided.

Intelligent Risk-Based Cybersecurity Protection for Industrial Systems Control—A Feasibility Study

Summary Intelligent automated industrial process control, enabled by IR 4.0 technology, requires increased system integration and connectedness. With such development comes a heightened risk of cyberattack for operational technology (OT) systems, including industrial control systems (ICS) and industrial automation and control systems (IACS), historically shielded from cyberspace. Cyberattacks on OT systems can also impact the physical world. The recent attack on the US Colonial Pipeline interrupted delivery of oil along the US West coast for several days, resulting in a declaration of a state of emergency, while the Ukraine power grid attacks interrupted electrical power supply in Kyiv, impacting thousands of consumers. In 2021, an Equinor-operated platform on the Norwegian Continental shelf was hit by malware. The incident had the potential to affect the drilling control system (DCS). For ICS, the concern is ensuring a safe process, as opposed to the more traditional concern with data protection for information technology (IT) system management. To ensure safe implementation of advanced industrial systems control, while avoiding unnecessary operational downtime, there is a need for cybersecurity solutions that account for risk for the whole cyberphysical system (CPS), including the ICS, the work environment, the product, and the physical surroundings. Diagnostic models and methods must therefore be developed that cover the functionality of the whole CPS. Further, to account for process dynamics and evolving cyberattack threats, adaptability is required with respect to both process state and new types of attacks. For this purpose, knowledge-based awareness monitoring combined with existing cyberattack detection tools for ICS, known as ICS intrusion detection systems (IDS), is proposed, providing an element of artificial intelligence. Results from studies in a DCS environment indicate that existing monitoring applications can be used to detect and discern between different types of cyberattacks on CPSs. This indicates feasibility with respect to monitoring the control and IT components of the ICS system for building risk-based cybersecurity decision support solutions. The key challenge and novelty of the proposed approach is to extend the capability of cybermonitoring in systems control to automated evaluation of process risk, including risk to the physical environment. Such capability will enable appropriate decision support with respect to both process risk and operational downtime.

Gas Sensor Array Drift in an E-Nose System: A Dataset for Machine Learning Applications

Gas sensor arrays are widely used in various applications such as environmental monitoring, industrial process control, and medical diagnosis. However, one of the main challenges in using gas sensor arrays is their tendency to drift over time, which can significantly affect their accuracy and reliability. In this research paper, we present a gas sensor array drift dataset that can be used to evaluate and develop drift compensation techniques. The dataset consists of measurements from an array of eight metal oxide gas sensors exposed to six different target gases at varying concentrations over several months. The paper also describes the experimental setup, data acquisition process, and preliminary dataset analysis. Our results show that the sensor array exhibits significant drift over time and that the drift patterns vary depending on the target gas and concentration. This dataset can provide a valuable resource for researchers and engineers working on gas sensor array applications and can help advance the development of more robust and accurate gas sensing systems.

Enhancing long-term stability of photoacoustic gas sensor using an extremum-seeking control algorithm

Smart sensor systems have gained increasing importance in various fields, including healthcare, environmental monitoring, industrial automation, and security. Photoacoustic gas sensors are an emerging type of optical sensor used in various applications due to its enhanced performance characteristics. However, the accuracy and reliability of gas concentration measurements from photoacoustic gas sensors may be impacted by several known limitations, including drift of the gas cell resonant frequency over extended periods of time. Researchers have proposed various solutions, including optimization methods and signal processing algorithms, to address this and others issues. In this paper, we propose a novel solution using an extremum-seeking control algorithm to manage the laser modulation frequency of photoacoustic gas sensors. By tracking the changing resonant frequency of the gas cell, long-term stability can be achieved, making it suitable for environmental monitoring, petroleum exploration, and industrial process control. Our approach has the potential to improve the accuracy and reliability of long-term measurements obtained from photoacoustic gas sensors, providing a stable and reliable method for gas concentration estimation.

LSTM-CNN Network-Based State-Dependent ARX Modeling and Predictive Control with Application to Water Tank System

Industrial process control systems commonly exhibit features of time-varying behavior, strong coupling, and strong nonlinearity. Obtaining accurate mathematical models of these nonlinear systems and achieving satisfactory control performance is still a challenging task. In this paper, data-driven modeling techniques and deep learning methods are used to accurately capture a category of a smooth nonlinear system’s spatiotemporal features. The operating point of these systems may change over time, and their nonlinear characteristics can be locally linearized. We use a fusion of the long short-term memory (LSTM) network and convolutional neural network (CNN) to fit the coefficients of the state-dependent AutoRegressive with the eXogenous variable (ARX) model to establish the LSTM-CNN-ARX model. Compared to other models, the hybrid LSTM-CNN-ARX model is more effective in capturing the nonlinear system’s spatiotemporal characteristics due to its incorporation of the strengths of LSTM for learning temporal characteristics and CNN for capturing spatial characteristics. The model-based predictive control (MPC) strategy, namely LSTM-CNN-ARX-MPC, is developed by utilizing the model’s local linear and global nonlinear features. The control comparison experiments conducted on a water tank system show the effectiveness of the developed models and MPC methods.

High-Order Modeling, Zeroing Dynamics Control, and Perturbations Rejection for Non-Linear Double-Holding Water Tank

The double-holding water tank system is a common non-linear control system that plays a crucial role in process control in the chemical industry. It consists of two cylindrical glass containers: the preset tank and the main tank. The main challenge in controlling this system is adjusting the main control valve to ensure that the actual liquid level of the main tank tracks the desired liquid level. This paper explores the zeroing dynamics (ZD) method and its application in tracking control. A non-linear model is developed for the double-holding water tank system, and the ZD method is used to design an effective controller (called the ZD controller) for tracking control. Additionally, the robustness of the double-holding water tank system in the presence of time-varying perturbations is investigated. In order to substantiate the effectiveness and robustness of the ZD controller, simulation experiments on four different tracking trajectories corresponding to four different practical situations, as well as an extra simulation experiment considering time-varying perturbations, are conducted. Furthermore, a comparative simulation experiment based on the backstepping method is conducted. The presented results successfully illustrate the feasibility and effectiveness of the ZD method for the tracking control of double-holding water tank systems.

Ordered Mesoporous Electrodes for Sensing Applications.

Electrochemical sensors have become increasingly relevant in fields such as medicine, environmental monitoring, and industrial process control. Selectivity, specificity, sensitivity, signal reproducibility, and robustness are among the most important challenges for their development, especially when the target compound is present in low concentrations or in complex analytical matrices. In this context, electrode modification with Mesoporous Thin Films (MTFs) has aroused great interest in the past years. MTFs present high surface area, uniform pore distribution, and tunable pore size. Furthermore, they offer a wide variety of electrochemical signal modulation possibilities through molecular sieving, electrostatic or steric exclusion, and preconcentration effects which are due to mesopore confinement and surface functionalization. In order to fully exploit these advantages, it is central to develop reproducible routes for sensitive, selective, and robust MTF-modified electrodes. In addition, it is necessary to understand the complex mass and charge transport processes that take place through the film (particularly in the mesopores, pore surfaces, and interfaces) and on the electrode in order to design future intelligent and adaptive sensors. We present here an overview of MTFs applied to electrochemical sensing, in which we address their fabrication methods and the transport processes that are critical to the electrode response. We also summarize the current applications in biosensing and electroanalysis, as well as the challenges and opportunities brought by integrating MTF synthesis with electrode microfabrication, which is critical when moving from laboratory work to in situ sensing in the field of interest.

State estimation approach to dual-modal imaging of two-phase flow based on electromagnetic flow tomography and electrical tomography

Monitoring, control and design of industrial processes involving multiphase flows often call for analysis of data from multiple sensors which give information on different quantities of the flowing materials. An example of such case is the problem of monitoring the flow of oil–water mixture: the phase fractions of oil and water, their velocities and volumetric flow rates cannot be retrieved from measurements given by a single sensing/imaging modality. For this reason, multi-modal tomographic imaging systems have been developed. In multi-phase flows, the quantities retrieved from different tomographic instruments are often interconnected—for example, the evolutions of the phase fractions depend on their velocities and vice versa. However, the analysis of data from different tomographic modalities is usually done separately—without taking into account physics that link the quantities of interest. In this paper, we propose a novel approach to image reconstruction in dual-modal tomography of multiphase flows. The governing idea is to combine the two modalities via Bayesian state estimation, that is, we write models that approximate connections between different quantities involved in the process and use sequential measurements from both modalities to jointly estimate these temporally evolving quantities. As an example case, we consider a dual-modal system comprising the electromagnetic flow tomography (EMFT) and electrical tomography (ET). While the EMFT is sensitive to the velocity field but also depends on the phase fractions of fluids, ET measurements are directly linked to phase fractions only. We study the performance of state estimation in EMFT-ET tomography with a set of numerical simulations. The results demonstrate that it outperforms the conventional stationary reconstruction approach, and also provides means for uncertainty quantification in multiphase flow imaging.

Learning-based near-optimal tracking control for industrial processes with slow and fast modes

This paper devotes to solving the optimal tracking control (OTC) problem of singular perturbation systems in industrial processes under the framework of reinforcement learning (RL) technology. The encountered challenges include the different time scales in system operations and an unknown slow process. The immeasurability of slow process states especially increases the difficulty of finding the optimal tracking controller. To overcome these challenges, a novel off-policy ridge RL method is developed after decomposing the singular perturbed systems using the singular perturbation (SP) theory and replacing unmeasured states using important mathematical manipulations. Theoretical analysis of approximate equivalence of the sum of solutions of subproblems to the solution of the OTC problem is presented. Finally, a mixed separation thickening process (MSTP) and a numerical example are used to verify the effectiveness.

TC-GATN: Temporal Causal Graph Attention Networks With Nonlinear Paradigm for Multivariate Time-Series Forecasting in Industrial Processes

Multivariate time series (MTS) forecasting plays an important role in industrial process monitoring, control and optimizations. Usually, hierarchical interactive behaviors among industrial MTS have formed complex nonlinear causal characteristics, which greatly hinders the applications of existing predictive models. It is found that graph attention networks (GAT) provide technical ideas to meet this challenge. However, the unknown directed graph and linear conversions of node information make conventional GAT less popular for the industrial fields. In this paper, we propose a novel prediction model termed as temporal causal graph attention networks with nonlinear paradigms (TC-GATN) to adequately capture inherent dependencies on industrial MTS. Specifically, the graph learning algorithm concerning the granger causality (GC) is exploited to extract potential relationships among multiple variables for guiding directional edge connections of the hierarchy. Then, parallel GRU encoders located in the graph neighborhood space are introduced to perform the nonlinear interaction of node features, which accomplishes the adaptive transformation and transmission. The self-attention mechanism is further employed to aggregate encoder hidden states across all stages. Finally, a temporal module is supplemented to process information from the graph layer, achieving satisfactory predictions. The feasibility and effectiveness of the TC-GATN are validated by two actual datasets from the methanol production and the chlorosilane distillation.

Design of Two-Degree-of-Freedom Fractional-Order Internal Model Control Algorithm for Pneumatic Control Valves

In response to the problems of the inaccurate pneumatic control valve model, the slow valve position control, and the low precision in the industrial control process, some improvement methods are proposed. Firstly, the fractional-order concept is introduced based on the first-order inertia model and IBBO (improved biogeography-based optimization) is used for iteration to obtain a specific transfer function model. Secondly, a fractional-order and two-degree-of-freedom combined internal model control algorithm is proposed. Finally, semi-physical experiments are carried out on a semi-physical experimental platform. The results show that in the field of pneumatic regulating valves, the fractional-order model has good adaptability and effectiveness, and the two-degree-of-freedom fractional-order internal model control algorithm also effectively improves the accuracy, speed, and robustness of the valve position control.

GRU-based model-free adaptive control for industrial processes

GRU-based model-free adaptive control for industrial processes

Validation of High-Availability Model for Edge Devices and IIoT

Competitiveness in industry requires smooth, efficient, and high-quality operation. For some industrial applications or process control and monitoring applications, it is necessary to achieve high availability and reliability because, for example, the failure of availability in industrial production can have serious consequences for the operation and profitability of the company, as well as for the safety of employees and the surrounding environment. At present, many new technologies that use data obtained from various sensors for evaluation or decision-making require the minimization of data processing latency to meet the needs of real-time applications. Cloud/Fog and Edge computing technologies have been proposed to overcome latency issues and to increase computing power. However, industrial applications also require the high availability and reliability of devices and systems. The potential malfunction of Edge devices can cause a failure of applications, and the unavailability of Edge computing results can have a significant impact on manufacturing processes. Therefore, our article deals with the creation and validation of an enhanced Edge device model, which in contrast to the current solutions, is aimed not only at the integration of various sensors within manufacturing solutions, but also brings the required redundancy to enable the high availability of Edge devices. In the model, we use Edge computing, which performs the recording of sensed data from various types of sensors, synchronizes them, and makes them available for decision making by applications in the Cloud. We focus on creating a suitable Edge device model that works with the redundancy, by using either mirroring or duplexing via a secondary Edge device. This enables high Edge device availability and rapid system recovery in the event of a failure of the primary Edge device. The created model of high availability is based on the mirroring and duplexing of the Edge devices, which support two protocols: OPC UA and MQTT. The models were implemented in the Node-Red software, tested, and subsequently validated and compared to confirm the required recovery time and 100% redundancy of the Edge device. In the contrast to the currently available Edge solutions, our proposed extended model based on Edge mirroring is able to address most of the critical cases, where fast recovery is required, and no adjustments are needed for critical applications. The maturity level of Edge high availability can be further extended by applying Edge duplexing for process control.

Plasma characteristics of a novel coaxial laser-plasma hybrid welding of Ti alloy

A novel coaxial laser-plasma hybrid welding method defined as “laser-soft plasma coaxial hybrid welding” was developed and studied. The plasma characteristics evolution, including plasma morphology, static electrical characteristic, and physical information of this welding method, was studied with the based material of TC4 alloy. Laser welding and plasma welding were carried out and used as comparative experiments. The plasma morphology of hybrid welding combines the characteristics of laser-induced plasma and arc plasma. The static electrical characteristic indicates that laser and plasma interact and influence each other but with different mechanisms under different current conditions. Physical information such as plasma electron temperature and ion density show that composite plasma can rapidly increase the heat source energy and ionization rate. To sum up, the coaxial plasma arc can increase the laser effect and redistribute the heat of the compound heat source, leading to a widened bead weld and better welding gap adaptability. This work provides an effective method and reference for hybrid welding process design and control in the practical industry.

Microservice-Oriented Architecture for Industry 4.0

Industry 4.0 (I4.0) is characterized by the integration of digital technologies into manufacturing processes and highlights new requirements for industrial systems such as greater interoperability, decentralization, modularization, and independence. The traditional hierarchical architecture of Industrial Automation Systems (IAS) does not fulfill these requirements and is evolving to incorporate information technologies in order to support I4.0 applications. The integration among these technologies, equipment, and systems at different industry levels requires a migration from the legacy vertical architecture to a flat architecture based on services. Service-oriented architecture (SOA) and, more recently, microservices play a critical role in I4.0 by providing a framework for integrating complex systems and meeting those requirements. This paper presents the development of a Microservice-Oriented Architecture for Industry 4.0 (MOAI), initially focused on evolving IAS to the I4.0. The objective is to describe the development, deployment, and testing of an IAS architecture based on microservices prepared for I4.0 applications. On the contrary to developing the whole software for the industrial SOA, the MOAI was developed on top of the Moleculer framework, which allowed focusing on creating services and applications for the automation and process control industry context. The development of several microservices and security mechanisms for the MOAI is presented, as is the deployment of IAS applications as services such as process control, SCADA, discrete automation, among others. The MOAI was implemented in a process control pilot plant for experimentation. Experimental results of the MOAI for IAS applications are investigated, the microservice communication performance is evaluated, and the pros and cons of microservices for I4.0 are discussed.

Towards Real-Time Analysis of Gas-Liquid Pipe Flow: A Wire-Mesh Sensor for Industrial Applications.

Real-time monitoring of gas-liquid pipe flow is highly demanded in industrial processes in the chemical and power engineering sectors. Therefore, the present contribution describes the novel design of a robust wire-mesh sensor with an integrated data processing unit. The developed device features a sensor body for industrial conditions of up to 400 °C and 135 bar as well as real-time processing of measured data, including phase fraction calculation, temperature compensation and flow pattern identification. Furthermore, user interfaces are included via a display and 4…20 mA connectivity for the integration into industrial process control systems. In the second part of the contribution, we describe the experimental verification of the main functionalities of the developed system. Firstly, the calculation of cross-sectionally averaged phase fractions along with temperature compensation was tested. Considering temperature drifts of up to 55 K, an average deviation of 3.9% across the full range of the phase fraction was found by comparison against image references from camera recordings. Secondly, the automatic flow pattern identification was tested in an air-water two-phase flow loop. The results reveal reasonable agreement with well-established flow pattern maps for both horizontal and vertical pipe orientations. The present results indicate that all prerequisites for an application in industrial environments in the near future are fulfilled.

A blockchain-enabled deep residual architecture for accountable, in-situ quality control in industry 4.0 with minimal latency

Real-time and vision-based quality control for industrial processes has drawn great interest from both scientists and practitioners, particularly following the transition to Zero Defect Manufacturing (ZDM) and Industry 4.0. Despite considerable progress, most ZDM approaches focus on the accuracy of the inspection process, often neglecting critical factors for application in the shop floor. On one hand, near real-time methods are needed for early defect detection and containment. On the other hand, data scarcity is an issue causing AI methods to overfit. Another concern is the accountability of AI results, since even if an AI pipeline is successfully deployed, its predictions are not verifiable in the long term. In this work, we explore a real-time solution based on lightweight Deep Residual Networks and Blockchain technology to address these issues. Concretely, we propose a two-phase training strategy to boost the performance of baseline classifiers while maintaining low inference times. The performance of the proposed methodology is presented in two different industrial use cases with strict timing requirements, one concerning battery assembly line and the other antenna manufacturing. We validate the proposed method for defect detection and compare the results with common training strategies demonstrating an improvement of 3% and 10% in F1-score and accuracy on the two cases respectively, while lowering inference time by 2.2× compared to existing light architectures. Contributing to the accountability of AI results, we present an IoT framework using Blockchain deployed in Private Ethereum.

A Framework and Operational Procedures for Metaverses-Based Industrial Foundation Models

Industrial processes are typical cyber–physical–social systems (CPSSs), where the effective management of employees and the efficient control of machines play important roles. Traditional industries heavily rely on human labor and neglect the development of collection–utilization–transmission integrated information loops, thereby leading to high costs and low efficiency in operational procedures. To facilitate the natural interactions and smart operations for humans and machines, industrial foundation models (IFMs) based on metaverses are proposed in this article, serving as the operating systems of industrial parallel machines that provide sustainable data resources and scenarios for management and control experiments. On this basis, IFM comprised of vision foundation models, language foundation models, as well as operational foundation models, are constructed to manage resources in industrial parallel machines and provides comprehensive services for industrial procedures. On the one hand, IFM can efficiently manage various resources including computing power, digital assets, enterprise resources, and platform I/O via the proposed CPSS-based competing, sharing, scheduling, monitoring, allocating, and recovering mechanisms. On the other hand, imaginative intelligence, linguistic intelligence, and algorithmic intelligence can be achieved through vivid visualization of vision foundation models, natural conversations of language foundation models, and smart manipulation of operational foundation models. With the proposed IFM, cyber–physical–social intelligence (CPSI) can be achieved to enhance the efficient management and control of industrial processes.