Journal of Process Control最新文献

Due to their predictive capabilities and computational efficiency, data-driven models are often employed in model predictive controller (MPC) design. These models offer precise predictions within their training domains, which aids in effective process control. However, real-world processes frequently experience operational changes, requiring control under new conditions that can lie beyond the training domains of existing data-driven models. Developing new models for these scenarios is challenging due to limited historical data. To address this limitation, we develop a novel data-driven control framework integrating an adaptive modeling approach called operable adaptive sparse identification of systems (OASIS) with the Luenberger observer. Firstly, we train the OASIS model and identify its domain of applicability (DA) using a support vector machine-based classifier. Subsequently, we formulate a Lyapunov-based MPC that relies on the OASIS model within the DA and the OASIS-based observer model beyond the DA. Additionally, we establish theoretical guarantees on the input-to-state stability of the observer, along with analyzing the stabilizability and recursive feasibility of the designed LMPC. The developed framework enhances the applicability of data-driven process control in diverse operating conditions. We highlighted its effectiveness using a chemical reactor example.

The conventional extended empirical wavelet transform (EEWT) proposed recently is intended to decompose multivariate signals with clear peaks in power spectra without considering the cases where the signals contain high noise levels. Even when dealing with signals with distinct peaks, the EEWT method can still encounter challenges in properly decomposing the signals. However, plant-wide data from industrial control loops, including controllers’ outputs, process variables, and manipulated variables, are commonly corrupted by high levels of noise, which can be introduced at various stages of data acquisition, transmission, and processing within the control system. To address these limitations and ensure the applicability of the EEWT to real-world industrial data with diverse and challenging characteristics, this paper presents an improved version called the improved extended empirical wavelet transform (IEEWT). The IEEWT incorporates noise-reduced power spectra and detrended fluctuation analysis techniques to enhance the decomposition. The proposed method demonstrates accurate multivariate data decomposition for both simulated and real data sets, surpassing the limitations associated with the conventional EEWT.

We study the problem of performance optimization of closed-loop control systems with unmodeled dynamics. Bayesian optimization (BO) has been demonstrated to be effective for improving closed-loop performance by automatically tuning controller gains or reference setpoints in a model-free manner. However, BO methods have rarely been tested on dynamical systems with unmodeled constraints and time-varying ambient conditions. In this paper, we propose a violation-aware contextual BO algorithm (VACBO) that optimizes closed-loop performance while simultaneously learning constraint-feasible solutions under time-varying ambient conditions. Unlike classical constrained BO methods which allow unlimited constraint violations, or ‘safe’ BO algorithms that are conservative and try to operate with near-zero violations, we allow budgeted constraint violations to improve constraint learning and accelerate optimization. We demonstrate the effectiveness of our proposed VACBO method for energy minimization of industrial vapor compression systems under time-varying ambient temperature and humidity.

Self-optimizing control is a strategy for selecting controlled variables, where the economic objective guides the selection and design of controlled variables, with the expectation that maintaining the controlled variables at constant values can achieve optimization effects, translating the process optimization problem into a process control problem. Currently, self-optimizing control is widely applied to steady-state optimization problems. However, the development of process systems exhibits a trend towards refinement, highlighting the importance of optimizing dynamic processes such as batch processes and grade transitions. This paper formally introduces the self-optimizing control problem for dynamic optimization, termed the dynamic self-optimizing control problem, extending the original definition of self-optimizing control. A novel concept, ”dynamic controlled variables” (DCVs), is proposed, and an implicit control policy is presented based on this concept. The paper theoretically analyzes the advantages and generality of DCVs compared to explicit control strategies and elucidates the relationship between DCVs and traditional controllers. Moreover, this paper puts forth a data-driven approach to designing self-optimizing DCVs, which considers DCV design as a mapping identification problem and employs deep neural networks to parameterize the variables. Three case studies validate the efficacy and superiority of DCVs in approximating multi-valued and discontinuous functions, as well as their application to dynamic optimization problems with non-fixed horizons, which traditional self-optimizing control methods are unable to address.

Modern manufacturing value chains require intelligent orchestration of processes across company borders in order to maximize profits while fostering social and environmental sustainability. However, the implementation of integrated, systems-level approaches for data-informed decision-making along value chains is currently hampered by privacy concerns associated with cross-organizational data exchange and integration. We here propose Privacy-Preserving Partial Least Squares (P3LS) regression, a novel federated learning technique that enables cross-organizational data integration and process modeling with privacy guarantees. P3LS involves a singular value decomposition (SVD) based PLS algorithm and employs removable, random masks generated by a trusted authority in order to protect the privacy of the data contributed by each data holder. We demonstrate the capability of P3LS to vertically integrate process data along a hypothetical value chain consisting of three parties and to improve the prediction performance on several process-related key performance indicators. Furthermore, we show the numerical equivalence of P3LS and PLS model components on both a synthetic and a real-world dataset and provide a thorough privacy analysis of the former. Moreover, we propose privacy-preserving explained X- and Y-block variance computations for determining the contribution of each data holder to the federated process model as a basis to incentivize data federation and fair profit-sharing.

This paper develops a data-driven approach for incipient fault diagnosis based on ANFIS and Takagi–Sugeno (TS) interval observers. First, the nonlinear bioreactor system is identified using an adaptive neuro-fuzzy inference system (ANFIS), which results in a set of polytopic TS models derived from measurement data. Second, a bank of TS interval observers is deployed to detect sensor and process faults using adaptive thresholds. Unlike other works that require training fault data, only fault-free data is considered for ANFIS learning in this work. Fault insolation is based on residual generation and evaluated on a fault signal matrix (FSM). Parametric uncertainty and measurement noise are considered to guarantee the method’s robustness. The effectiveness of the proposed method is tested on a well-known bioreactor Continuous stirred tank reactor system (CSTR) reference simulator.

This contribution develops the model predictive control for an unstable chemostat reactor. Initially, we analyze the system’s model — a nonlinear first-order hyperbolic partial integro-differential equation (PIDE) — and carry the model linearization around an unstable operating condition. Employing the structure-preserving Cayley–Tustin transformation, we obtain a discrete-time model representation of the continuous model. Subsequently, we solve the operator Ricatti equations in the discrete-time setting to derive a full state feedback controller that stabilizes the closed-loop and design a Luenberger observer for state reconstruction given the system output measures. Finally, we formulate a dual-mode MPC ensuring constraint satisfaction and optimality, integrating the gain-based unconstrained full-state feedback optimal control obtained from the Ricatti equation. This dual-mode strategy describes an optimization problem where the predictive controller acts only if constraints become active within the control horizon. Simulation studies validate the controller performance, where the MPC only takes action if the constraints are predicted to be active within the control horizon while also guaranteeing closed-loop stabilization under only output feedback. This type of controller can be easily implemented with other control strategies and significantly decreases the computational costs of solving the optimal control problems when compared to other MPC approaches.

Wastewater treatment plants (WWTPs) employ a series of complex chemical and biological processes, to transform an influent stream of contaminated water to an effluent suitable for return to the water cycle. To optimize the performance of WWTP control schemes, appropriate mathematical models capable of accurately simulating the plant dynamic behavior are essential. However, the development of reliable dynamic representations for these large-scale plants is challenging, mainly because of the complex biological reactions taking place and the significant fluctuations in the disturbances that affect the operation of WWTPs. First-principles models, such as the well-known benchmark simulation model no. 1 (BSM1), may be capable of capturing the highly nonlinear nature of WWTPs, but this comes at the cost of employing complex, high-order representations of the reactive units and settling processes. This complexity leads to highly complicated configurations that cannot be efficiently integrated in advanced process control schemes, like model predictive controllers (MPCs). Furthermore, the large number of unknown parameters in these models, along with the non-convex nature of the underlying functions, renders the use of conventional system identification techniques insufficient. To remedy these issues, in this work we introduce a reduced-order first-principles model for WWTPs, incorporating low order mathematical models for the chemical phenomena of the reactive units and the settling procedure. Furthermore, we present a novel system identification scheme, which is based on a customized cooperative particle swarm optimization approach; the scheme effectively handles the high-dimensionality and multimodality of the underlying nonlinear optimization problem, enabling accurate estimation of the model parameters. Comparison results between the dynamic behavior of the original BSM1 and the identified reduced-order model, indicate that the proposed approach is capable of accurately and robustly capturing the highly nonlinear nature of WWTPs, while being simple enough for incorporation in the design of MPC and other advanced control schemes. This represents a significant advancement over traditional models, offering a more practical and efficient approach for WWTP management and control.

This study presents a novel approach for controlling an industrial process that exhibits uncertainty and significant nonlinear features. The proposed method utilizes a virtual unmodeled dynamic and data-driven nonlinear robust predictive control strategy. The representation of a controlled object involves a composite state space model that combines both linear and high-order nonlinear elements. Moreover, a robust model predictive controller is developed using the linear component. In addition, the notion of one-step optimal feedforward is used in combination with a compensating controller to handle the high-order nonlinear factor specifically. Subsequently, a compensation controller with incremental characteristics is developed for a modified version of the high-order nonlinear term. Furthermore, the stability conditions of the closed-loop system are derived, and an analysis is conducted on the stability and convergence of the proposed approach. The TTS20 three-capacity water tank was utilized in both simulations and practical scenarios. The study demonstrated that the suggested approach successfully reduces system output variations and enhances overall performance in response to unpredictable changes in the process’s dynamic features.

In the aluminum electrolysis process, the accurate identification of anode effect (AE) can improve production efficiency. However, the existing methods fail to effectively capture the features of the anode current signal (ACS) due to its complex dynamic characteristics and temporal–spatial dependence. To address this issue, we propose a Process Knowledge-guided Deep Temporal–spatial Feature Learning Network (PKG-DTSFLN). We believe that knowledge and production data are complementary. Knowledge has potential to deduce beyond observational conditions. Data can be used to detect unexpected patterns. The combination of data and knowledge is potential to improve the performance. Specifically, knowledge is utilized to construct the adjacency matrix to represent the spatial structure of ACS. Then, a deep learning model is constructed by integrating the 1D-CNN and GAT, which is used to capture the temporal–spatial features of ACS. The experimental results on ACS dataset show that the accuracy is more than 99% with low computational cost.