Journal of Process Control最新文献

Momentum wheel bearing is a critical component within satellite systems, and its condition monitoring not only extends the operational lifespan of the satellite but also ensures the seamless fulfillment of its mission objectives. Various data-driven techniques have been introduced to assimilate health-related information. However, these techniques neglect the significant challenges posed by robust disturbance and volatility of degradation process, resulting in suboptimal evaluation performance. To address these issues comprehensively, this paper proposes a novel approach named canonical variable fluctuation analysis (CVFA) to facilitate precise health monitoring of momentum wheel bearings by concurrent analysis of static deviation and dynamic oscillation. Firstly, three quantifiable standards of consistency, accuracy and sensitivity are defined to select the degradation trend-related indices from multi-domain features, which provides an automatic and objective feature selection method. Subsequently, CVFA is developed to realize feature reduction and extracts the dynamic information from the features with strong disturbance and high fluctuation. Two Fluctuation (F) statistics are defined to characterize the health degradation trend by integrating both static deviation and dynamic volatility within a sliding window. Afterwards, autoregressive moving average (ARMA) model is constructed on the basis of F statistics for short-term prognostication, which enables proactive detection of degradation trends. Lastly, by integrating two F statistics, a health degree (HD), which is independent of parameter adjustments, is defined to intuitively represent bearing health status. The efficacy and superiority of the proposed method are substantiated through validation and analysis conducted using accelerated life tests of bearings.

Predicting quality-relevant process variables is of paramount importance in optimizing and controlling chemical processes. Probabilistic Slow Feature Analysis (PSFA), a potent data-driven technique, plays a pivotal role in deducing quality indices by abstracting gradual variations in processes distinctly characterized by pronounced inertia. Nevertheless, PSFA’s predictive efficacy encounters a substantial bottleneck due to the assumption of a single operating condition, compromising its accuracy, particularly in industries represented by switching operating conditions. To surmount this limitation, this study proposes an innovative approach that enriches PSFA with multi-operating condition process data and limited labels within a Bayesian framework, effectively combining continuous and discrete first-order Markov chains to capture the processes’ inertia and dynamic shifts. The proposed method updates latent posterior distributions and model parameters iteratively via the Expectation–Maximization algorithm. The effectiveness of the proposed methodology is verified through a numerical case and industrial hydrocracking process data.

Microalgae cultivation for energy production is a promising avenue for converting solar light into sustainable biofuel. Solar processes are however subjected to the permanent fluctuations of light and medium temperature. Accurate temperature prediction of the culture medium turns out to be critical for optimising growth conditions. In this study, we introduce a reduced-model approach derived from existing models, turning the complex heat transfer modelling problem into an identification problem. The resulting generic model, called the Simplified Auto Tuning Heat Exchange (SATHE) model, has a clear and simple structure, offering a balance between accuracy and computational complexity. The SATHE model is versatile and contains the necessary terms to catch a large variety of heat transfer problems, while the parameters can be identified from experimental data. We first prove the parameter identifiability and then propose an identification strategy, based on the gradient computation, to identify the model’s underlying parameters. We further validate the SATHE model performance in two distinct reactors across various seasons. Finally, we discuss the potential of online applications with a continuous self-tuning strategy to keep optimal predictive performances. This work lays the foundation for enhanced control strategies in large-scale cultivation systems.

This work presents an algorithm for estimation-based model predictive control with objective prioritization such that distinct objectives may be defined for mutually exclusive operational regions. The objective prioritization algorithm is built by using logical conditions that define regions of operation which are incorporated into the objective function, thus allowing smooth transitions between a bank of objectives. The control objective prioritization is cast in the framework of model predictive control that is coupled with an extended Kalman filter for estimation of critical yet unmeasured state variables. The algorithm is applied to the challenging control problem of an industrial superheater (SH)-reheater (RH) system of a natural gas combined cycle plant under load following operation where smooth transitions among various control objectives is desired – operation under nominal conditions, avoidance of spraying to saturation at the inlet of the SH and RH systems, and avoidance of main steam temperature excursions. The results from the estimator framework are compared with the industrial data from an operating power plant. The control algorithm is evaluated by simulating a servo control problem and disturbance rejection scenarios as expected under load-following operation of the power plant. This algorithm is generic and can be applied to accomplish local control policies for safety, economics, quality control, state constraints, and others.

Topical Heading

Process Systems Engineering

In narrow gap gas-shielded arc welding (NG-GMAW) for pipelines, maintaining a stable welding process and ensuring weld quality necessitates controlling the extension length of the welding wire (WWEL) within a specific range. However, when dealing with three-dimensional weld workpieces featuring height variations, welding defects are prone to occur due to changes in welding wire extension length. Therefore, real-time adjustment of the distance between the contact tip and workpiece (CTWD) is crucial during the welding process. To address this challenge, this paper proposes a welding torch height (WTH) control method based on passive vision sensing. The proposed method utilizes a wide dynamic range (WDR) camera to acquire distinct real-time welding images. An adaptive region of interest extraction method for the welding wire is then proposed based on the position relationship between the welding wire and arc. To address false edge issues in the welding wire profile, a cellular neural network (CNN) edge detection algorithm, optimized by particle swarm optimization, is employed to eliminate false edges. The extended length of the welding wire is subsequently extracted using an adaptive mask kernel morphology and corner detection method. Accordingly, a model predictive control (MPC) technique is developed to govern the height of the welding torch with the WWEL as input. The proposed MPC algorithm's tracking performance and robustness are validated through feedback control experiments. The results indicate that the tracking error of the WTH trajectory can be controlled within±0.41 mm, meeting the requirements of NG-GMAW welding torch height control for welding robots.

This paper introduces a new approach to designing a disturbance observer called a modified extended state observer (ESO). The existing ESO technique ensures that the trajectories of estimation error dynamics globally asymptotically converge to zero in the absence of time varying disturbances. However, for perturbed systems, where time varying disturbances affect system behavior, these trajectories never reach zero but rather remain bounded around the origin within a constant value. Consequently, this discrepancy leads to challenges in accurately estimating state trajectories and discerning information about disturbances. This, in turn, complicates the precise estimation of state dynamics and disturbances and poses difficulties in designing control laws for stability analysis of the system. Unlike existing ESO methods, the distinguishing characteristic of this modified ESO is its capability to achieve global and asymptotic convergence of observation error in the presence of unknown bounded time varying disturbances. This unique property enables the exact estimation of state trajectories. Information about the bounded time varying disturbances is obtained and significantly attenuate more efficiently compared to existing ESO techniques. Based on the estimated disturbances, any classical controller can be designed for the system to achieve set-point tracking subject to time varying disturbances. To validate the performance of the proposed modified ESO, the model of a coupled-tank setup is simulated for the stabilization problem and its experimental setup is demonstrated for the tracking problem. For carrying out the experiment on a real-time hardware setup, an input of step change at every 60 s with water level variation of $\pm$2 cm from initial set value of 15 cm to achieve the set-point tracking of the water levels in the both the tanks along with good transient performance in the presence of time-varying external disturbances.

In industrial processes, stiction in control valves is a common cause of degradation in control loop performance. It is important to identify and establish reliable stiction models to enhance control loop performance. This study focuses on constructing a precise model of the pneumatic control valve using the LuGre friction model. To improve the model, a smooth function based on probability density is introduced to alleviate the LuGre friction model. The maximum error resulting from this smoothing process is also analyzed. To estimate the valve stiction parameters, the direct transcription method is utilized to convert the problem from systems of ordinary differential equations into a nonlinear programming problem. The interior point method is then used to solve this problem. Furthermore, the estimability of the parameters is analyzed based on the reduced Hessian matrix before estimation. Numerical results demonstrate that the proposed approach in this study effectively estimates the stiction parameters of pneumatic control valves.

Accurately predicting the burn-through point (BTP) is crucial for achieving stable control of the sintering process. However, accurately measuring the raw BTP is difficult due to the harsh production environment and poor thermocouple measurement accuracy of the temperature of exhaust gas in bellows. This paper proposes a prediction model of the BTP with data correction based on the feature matching of cross-section frame at discharge end. Firstly, a feature extraction method of cross-section frames at discharge end is designed. Next, the cross-section frame at discharge end features matching method is used to correct the raw BTP, and this method corrects anomalous data resulting from sensor failures. Finally, the temporal convolutional neural network and gated recurrent unit are used to predict the corrected BTP. The prediction model considers the cross-section frame feature at discharge end and state parameters as inputs, and it can achieve accurate prediction of the corrected BTP. A series of comparative experiments are conducted to verify the feasibility and effectiveness of the proposed model. At the same time, this paper also designs industrial implementation plan,and use actual operation data to verify the feasibility of the designed industrial implementation plan.

The tunable approximated explicit model predictive control (MPC) comes with the benefits of real-time tunability without the necessity of solving the optimization problem online. This paper provides a novel self-tunable control policy that does not require any interventions of the control engineer during operation in order to retune the controller subject to the changed working conditions. Based on the current operating conditions, the autonomous tuning parameter scales the control input using linear interpolation between the boundary optimal control actions. The adjustment of the tuning parameter depends on the current reference value, which makes this strategy suitable for reference tracking problems. Furthermore, a novel technique for scaling the tuning parameter is proposed. This extension provides to exploit different ranges of the tuning parameter assigned to specified operating conditions. The self-tunable explicit MPC was implemented on a laboratory heat exchanger with nonlinear and asymmetric behavior. The asymmetric behavior of the plant was compensated by tuning the controller’s aggressiveness, as the negative or positive sign of reference change was considered in the tuning procedure. The designed self-tunable controller improved control performance by decreasing sum-of-squared control error, maximal overshoots/undershoots, and settling time compared to the conventional control strategy based on a single (non-tunable) controller.

This work focuses on the design of an optimal adaptive control system for temperature regulation in a catalytic flow reversal reactor (CFRR), utilizing a reinforcement learning (RL) approach. First, a policy iteration algorithm is introduced to learn the optimal solution of the associated linear-quadratic control problem online. It should be mentioned that this approach is not reliant on the internal dynamics of the CFRR system, which is a complex process and is most effectively modeled using Partial Differential Equations (PDEs). The convergence of the iteration algorithm is established, assuming the initial policy is stabilizing. Additionally, a second algorithm is presented to enhance the implementability of the reinforcement learning algorithm from a practical perspective. Numerical simulations are carried out to illustrate the efficacy of the proposed algorithm.