{"title":"ECCBO: An inherently safe Bayesian optimization with embedded constraint control for real-time process optimization","authors":"Dinesh Krishnamoorthy","doi":"10.1016/j.jprocont.2025.103467","DOIUrl":"10.1016/j.jprocont.2025.103467","url":null,"abstract":"<div><div>This paper presents a model-free real-time optimization (RTO) framework that leverages unconstrained Bayesian optimization (BO) embedded with constraint control to achieve optimal steady-state operation of process systems without the need for detailed models. Leveraging the vertical decomposition of information flow with timescale separation, this paper proposes two approaches to BO with embedded constraint controllers that simplifies model-free RTO with unknown cost and constraints, while ensuring steady-state constraint feasibility. The first approach employs constraint controllers that controls the constraints to some feasible setpoint in the fast timescale, and an unconstrained BO finds the optimal setpoints to these controllers in the slower timescale. The second approach uses constraint controllers as safety filters, where BO searchers over the RTO degrees of freedom, which can be overridden by the constraint controller when necessary to ensure steady-state constraint feasibility. By embedding constraint controllers with Bayesian optimization, both approaches ensure zero cumulative constraint violation without depending on specific assumptions about the Gaussian process model used in Bayesian optimization, making it inherently safe. The proposed scheme is demonstrated on several illustrative benchmark examples.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103467"},"PeriodicalIF":3.3,"publicationDate":"2025-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144502844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hanwen Zhang , Qingqing Liu , Jianxun Zhang , Jun Shang
{"title":"Robust recursive transformed component statistical analysis for incipient industrial fault detection with missing data","authors":"Hanwen Zhang , Qingqing Liu , Jianxun Zhang , Jun Shang","doi":"10.1016/j.jprocont.2025.103470","DOIUrl":"10.1016/j.jprocont.2025.103470","url":null,"abstract":"<div><div>In practical industrial processes, data integrity is often compromised by sensor malfunctions or issues in data management. Furthermore, incipient faults, which can escalate into severe accidents, are typically challenging to detect due to their subtle nature. This paper introduces a robust recursive transformed component statistical analysis method for detecting incipient faults in industrial processes with missing data. Within a sliding window, missing data are restored by minimizing the detection index in a recursive way, and the converged statistical model is then used for fault detection. The detectability of the proposed method is analyzed theoretically in scenarios with incomplete data. To validate the effectiveness of the proposed method, experiments are conducted on both a numerical case study and the Tennessee Eastman process. The results demonstrate robust performance under incomplete training and testing data, enabling accurate detection of incipient faults in industrial settings. Furthermore, compared to existing methods, the proposed approach achieves significant improvements in fault detection under missing-data conditions, attaining a detection rate close to 100% for most fault scenarios while maintaining a near-zero false alarm rate.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103470"},"PeriodicalIF":3.3,"publicationDate":"2025-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144481118","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Big data-driven predictive control for nonlinear systems—A trajectory cluster-based contraction approach","authors":"Shuangyu Han , Yitao Yan , Jie Bao , Biao Huang","doi":"10.1016/j.jprocont.2025.103474","DOIUrl":"10.1016/j.jprocont.2025.103474","url":null,"abstract":"<div><div>This article presents a novel contraction-based big data-driven predictive control (CBDPC) approach for nonlinear systems using the behavioural systems framework. The nonlinear behavioural space is partitioned into linear sub-behavioural spaces, represented by connected trajectory clusters. The controller drives the process to travel through multiple linear sub-behavioural spaces to reach the setpoint. By introducing the concepts of data-based contraction and differential dissipativity, a trajectory cluster-based control contraction metric and contraction condition are developed to guarantee incremental exponential stability of the controlled nonlinear system behaviour and attenuate the effect of linear sub-behaviour approximation errors on controlled output. Connected trajectory clusters are obtained via multi-view fuzzy clustering, which partitions nonlinear system behaviour (i.e., a set of input–output data trajectories) into connected linear sub-behaviours (i.e., trajectory subsets with intersections). Based on the above contraction and dissipativity conditions, an online data-driven predictive control approach using Hankel matrices is developed. The proposed approach is illustrated using a case study on control of an aluminium smelting process, which demonstrates the control performance achieved by the CBDPC approach.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103474"},"PeriodicalIF":3.3,"publicationDate":"2025-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144481163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Halil Arici , Fatir A. Qureshi , Jay Mulmule , Juergen Hahn
{"title":"Multiple hypothesis testing for pre-screening of factor selection for classification of high-dimensional data","authors":"Halil Arici , Fatir A. Qureshi , Jay Mulmule , Juergen Hahn","doi":"10.1016/j.jprocont.2025.103469","DOIUrl":"10.1016/j.jprocont.2025.103469","url":null,"abstract":"<div><div>Modern instrumentation, such as mass spectrometry, enables the measurement of concentrations of hundreds or even thousands of compounds in individual samples. These measurements are often used in process data analytics to build classification models for determining whether a process is operating satisfactorily, if a product meets specifications, or to diagnose specific health conditions in patients. A common challenge associated with these applications is that the number of measured compounds far exceeds the number of available samples, increasing the risk of overfitting. Typically, it is advisable to have 10–20 samples per input factor of the classification model, thereby requiring the selection of only a handful of concentrations from potentially thousands. However, identifying the best combination of compounds from such a large pool by an exhaustive search is computationally infeasible.</div><div>A common approach to address this issue is pre-screening the compounds for statistically significant differences between groups, then limiting model inputs to only those identified as significant. The simplest form of pre-screening involves a student’s t-test, however, with a commonly-used <span><math><mi>p</mi></math></span>-value threshold of 0.05, one expects 5% of the compounds to be false positives, even when no true differences exist. Multiple hypothesis testing techniques, such as the Bonferroni correction and the Benjamini–Hochberg procedure, can reduce the number of compounds considered by accounting for these false positives. However, these methods often make assumptions about the data that are not valid in practice, leading to overly conservative results and potentially missing important compounds.</div><div>In this work, we present a screening procedure that computes the false discovery rate of p-values using a Leave-n-Out approach. By omitting <span><math><mi>n</mi></math></span> samples at a time and repeatedly calculating the p-values, we assess the robustness of statistical significance against small changes in the dataset. We compare this technique to the Bonferroni correction and Benjamini–Hochberg procedure using both synthetic examples and two experimental datasets from the life sciences. Our results demonstrate that while the proposed approach is more conservative than a simple t-test, it identifies compounds that lead to better-performing models compared to those selected using existing multiple hypothesis testing methods.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103469"},"PeriodicalIF":3.3,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144470983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Zeyuan Xu , Wei Xing Zheng , Yujia Wang , Danwei Wang , Zhe Wu
{"title":"Federated learning-based distributed model predictive control","authors":"Zeyuan Xu , Wei Xing Zheng , Yujia Wang , Danwei Wang , Zhe Wu","doi":"10.1016/j.jprocont.2025.103472","DOIUrl":"10.1016/j.jprocont.2025.103472","url":null,"abstract":"<div><div>This work develops a federated learning-based distributed model predictive control (FL-DMPC) method for nonlinear networked systems to improve data privacy in the development of machine learning models. Specifically, a novel framework of FL method with personalized optimization (Fedpo) is proposed to obtain the global FL model for the entire networked system by updating and aggregating the personalized models for subsystems. This new FL framework significantly reduces the complexity of the learning algorithm and improves computational efficiency compared to existing FL methods. Additionally, it addresses system heterogeneity due to various dynamics of subsystems. Subsequently, the convergence of the Fedpo framework is proved by deriving an upper bound for its generalization and personalization errors, followed by theoretical analysis of the closed-loop stability of nonlinear networked systems under FL-DMPC. Finally, a chemical process network is adopted to demonstrate the effectiveness of the proposed Fedpo modeling and FL-DMPC method.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103472"},"PeriodicalIF":3.3,"publicationDate":"2025-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144471077","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Development of an intelligent moving horizon estimator integrated with fault diagnosis for automated model maintenance","authors":"Giriraj Bagla, Sachin C. Patwardhan, Mani Bhushan","doi":"10.1016/j.jprocont.2025.103468","DOIUrl":"10.1016/j.jprocont.2025.103468","url":null,"abstract":"<div><div>Fidelity of the dynamic or steady-state model used for making economic decisions in real-time optimization (RTO) or economic nonlinear model predictive control (ENMPC) is critical in achieving the desired economic benefits in the face of fast changing and uncertain market conditions. Since the model parameters/ unmeasured disturbances keep changing with changes in the operating regime, online updating of the model parameters using recent operating data is essential for accruing the benefits of RTO or ENMPC over a long period of time. In practice, a large number of model parameters/ unmeasured disturbances are susceptible to change and adjusting all of them without discrimination can result in over-fitting and/or erroneous parameter estimates. Automating the task of finding the “active subset of parameters/ disturbances” that need to be adjusted while carrying out the online model update can eliminate the need for an expert intervention for online maintenance of a dynamic/ steady-state model. This can be achieved by developing an automated decision-making system that performs the active subset selection task by diagnosing the root cause(s) of departures from the normal behavior by analyzing transient data. In this work, fault tolerant moving horizon estimator (MHE) approaches have been proposed that integrate fault diagnosis and identification (FDI) with the conventional MHE formulation for carrying out automated online model maintenance. Diagnosis and compensation for bias and drift-type faults in sensors, actuators, model parameters, and unmeasured disturbances have been considered in the development. Statistical properties of decision variables of the unconstrained MHE formulation for linear systems are derived and further used for fault detection and estimation of the time of occurrence of a fault. Subsequently, fault identification step is derived using the generalized likelihood ratio framework. Since the magnitude of the isolated fault may drift with time, the fault magnitude estimates are refined by including the isolated fault magnitude as an additional parameter in the MHE decision variable set. A hypothesis test is developed to stop the magnitude refinement when the fault magnitudes converge. The model used in MHE is subsequently modified to accommodate persistent faults so that multiple faults occurring sequentially in time can be diagnosed. Further, to facilitate the application of the proposed approach to systems exhibiting nonlinear dynamics, trajectory linearization-based and nonlinear MHE-based approaches are developed for carrying out FDI. The efficacy of the proposed approaches is demonstrated by conducting stochastic simulations using the benchmark Williams–Otto reactor system. Analysis of the simulation results reveals that the proposed MHE-based FDI approaches outperform the Kalman filter and extended Kalman filter-based FDI approaches in terms of diagnostic performance. Moreover, the proposed MHE-FDI approaches are","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103468"},"PeriodicalIF":3.3,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144365198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Tracking and estimation of bottom-hole unmeasured pressure states using a fuzzy observer","authors":"Saeed Amiri , Mohsen Mohammadpour , Roza Abbasi","doi":"10.1016/j.jprocont.2025.103477","DOIUrl":"10.1016/j.jprocont.2025.103477","url":null,"abstract":"<div><div>This paper focuses on sustaining a constant bottomhole pressure, a widely used approach in Managed Pressure Drilling (MPD). MPD is a multifaceted, nonlinear process that uses various control methods for estimation and stabilization of the pressure at the bottom of the well. Applying the T–S fuzzy modeling method and the concept of parallel distributed compensation (PDC), an observer-based fuzzy controller is developed to ensure that the bottom-hole pressure tracks the desired target value without exceeding the pressure window, even in the presence of uncertainties and disturbances, such as pore and fracture pressures, while concurrently addressing cases where the system states are inaccessible. To accomplish this, an observer is developed using estimated antecedent variables to address practical challenges. To begin, the wellbore model is introduced, followed by the use of the T–S fuzzy model (TSFM) approach to represent the precise dynamics of the wellbore’s nonlinear system. Next, a tracking controller based on PDC is employed to achieve the control objectives. The Lyapunov method, together with the <span><math><msub><mrow><mi>H</mi></mrow><mrow><mi>∞</mi></mrow></msub></math></span> criterion, ensures the stability of the controlled wellbore system, with sufficient criteria established through an LMI optimization process. The results of the simulation reveal the improvement and merits of the proposed observer-based controller.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103477"},"PeriodicalIF":3.3,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144470982","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A novel hybrid neural network for modeling dynamic systems using physics-informed regularization","authors":"Devavrat Thosar , Abhijit Bhakte , Zukui Li , Rajagopalan Srinivasan , Vinay Prasad","doi":"10.1016/j.jprocont.2025.103473","DOIUrl":"10.1016/j.jprocont.2025.103473","url":null,"abstract":"<div><div>Physics-Informed Neural Networks (PINNs) are very popular due to their ability to incorporate first-principles knowledge in traditional neural network models. However, many applications of traditional PINNs in chemical process modeling treat time as an explicit input, rendering them incompatible with a process control framework. In contrast, more advanced approaches for modeling dynamic systems with process control in mind, such as Physics-Informed Recurrent Neural Networks (PI-RNNs), demand high computational resources for both training and implementation. As a solution, we propose a hybrid Physics-Informed Nonlinear Auto-Regressive with eXogenous inputs (PI-NARX) model that is accurate, computationally efficient, and inherits the desired properties of hybrid models. We demonstrate the effectiveness of this approach with a case study based on a Continuous Stirred Tank Reactor. The proposed hybrid model reduces the Mean Absolute Error by 17% for interpolation and 19.5% for extrapolation over the traditional data-driven NARX model. Additionally, we demonstrate the enhanced performance of PI-NARX over NARX in cases of practical importance, such as when limited data or limited process knowledge is available, and in the presence of noisy measurements, indicating the practicality and effectiveness of hybrid machine learning for real-world systems. We also benchmark the performance of the PI-NARX model against that of a PI-RNN, and demonstrate that the PI-NARX model outperforms the PI-RNN in terms of computational efficiency and prediction accuracy.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103473"},"PeriodicalIF":3.3,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144365196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ulrich Knechtelsdorfer , Martin Stefan Baumann , Andreas Steinboeck , Andreas Kugi
{"title":"On model predictive control of the lateral strip motion in tandem hot rolling mills","authors":"Ulrich Knechtelsdorfer , Martin Stefan Baumann , Andreas Steinboeck , Andreas Kugi","doi":"10.1016/j.jprocont.2025.103463","DOIUrl":"10.1016/j.jprocont.2025.103463","url":null,"abstract":"<div><div>Almost all flat steel production involves processing in a tandem hot strip finishing mill. The mass flow in the hot strip finishing mill influences final product’s quality and the stability of the process. The mass flow can be categorized into longitudinal and lateral depending on its main direction. Uncontrolled lateral strip movement may lead to plant damage; therefore, the lateral strip motion must be controlled. An existing model for the lateral strip motion is extended to serve as a basis for a model predictive controller. The controller systematically respects the limits of the manipulated variable, its time derivative, and the lateral position of the strip and shows good convergence behavior. The proposed control strategy works on top of the control structure currently applied at the plant. A state observer is proposed for the angle and position of the strip in the roll gap, which are not directly measurable but are required by the model predictive controller. The observer also estimates the camber of the strip in the finishing mill, which is valuable information for the operators. The efficacy of the proposed control structure is evaluated by simulation studies utilizing a mathematical plant model validated by measurements.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103463"},"PeriodicalIF":3.3,"publicationDate":"2025-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144364746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shuaishuai Han , Chenyang Fan , Mingzhang Wang , Kejin Huang , Yang Yuan , Xing Qian , Haisheng Chen , Wei Qin
{"title":"Improving controllability of reactive double dividing-wall distillation columns through an overdesign strategy","authors":"Shuaishuai Han , Chenyang Fan , Mingzhang Wang , Kejin Huang , Yang Yuan , Xing Qian , Haisheng Chen , Wei Qin","doi":"10.1016/j.jprocont.2025.103489","DOIUrl":"10.1016/j.jprocont.2025.103489","url":null,"abstract":"<div><div>In our previous work, reactive double dividing-wall distillation columns (R-DDWDCs) have been demonstrated to have significant potential for energy savings in the separation of reacting mixtures with the most unfavorable ranking of relative volatilities (A + B ⇌ C + D, with α<sub>A</sub> > α<sub>C</sub> > α<sub>D</sub> > α<sub>B</sub>). However, if its design process focuses solely on economic considerations, the resulting design is likely to suffer from poor controllability. This is primarily because the economically optimal design tends to use a relatively low reaction conversion rate to enhance contact between reactants. As a result, this leads to the excessive accumulation of unconverted reactants in the top and bottom product streams, thereby narrowing the control range for these product streams and decreasing control flexibility. To address this issue, an overdesign strategy was proposed. The specific measure of this strategy is to adjust the liquid or vapor split ratio near the reactive section on the basis of the economically optimal design to moderately increase the reaction conversion rate, thereby reducing the accumulation of unconverted reactants in the top and bottom product streams. Based on the designs of two typical R-DDWDC configurations for an ideal quaternary endothermic reversible reaction and the exothermic esterification system of methanol + lactic acid ⇌ water + methyl lactate, the feasibility and effectiveness of the proposed overdesign strategy were evaluated. The results showed that the overdesigned R-DDWDC exhibits a significant improvement in controllability, with an acceptable decrease in steady-state performance. This achievement offers insights into integrating steady-state and dynamic design, facilitating the practical application of R-DDWDCs in industry.</div></div>","PeriodicalId":50079,"journal":{"name":"Journal of Process Control","volume":"152 ","pages":"Article 103489"},"PeriodicalIF":3.3,"publicationDate":"2025-06-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144365195","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}