A novel approach for multivariate process monitoring using several intelligences

Nowadays, research and development for optimal setting control (OSC) of complex industrial processes is becoming a hot topic. However, there is no configurable software for OSC design. In this paper, a novel software platform is proposed. The major superiority is that it not only integrates intelligent and data-driven mainstream algorithms within a unified framework, but also provides a powerful controller design tool and employs a Petri net engine to schedule control units in a parallel computation environment. The overall framework and key technologies are introduced in detail. The platform is verified and validated through a case of application to the design and development of OSC system of grinding process in a hardware-in-the-loop simulation environment.

The operation control for complex industrial processes consists of two layer — loop control layer and operational control layer. The former aims at achieving the required loop control for each production unit along production line whilst the later optimizes the set-points to the control loops so that certain performance indexes (such as product quality and energy cost) is optimized when the closed loop controlled variables follow their set-points well. This paper presents a novel method that can be used to analyze the performance deterioration of the optimized operational control, where the impact of tracking errors of loop control to the optimized performance is quantitatively formulated when the tracking errors are small. It has been shown that loop tracking errors would generally deteriorate the optimized performance in a quadratic way — leading to the establishment of the square impact principle (SIP). Moreover, it has been shown that the production infrastructure will also affect the deterioration of the optimized performance. Formulation on the analysis on the flat robustness (FR) and randomness of the optimized performance indexes will also be made and several issues on the future studies are listed.

A central challenge in the control of complex industrial processes like heat exchangers is the persistent disconnection between global offline optimization and real-time online adaptation, leading to suboptimal performance under dynamic conditions. To address this, this paper proposes a coordinated hierarchical control strategy, CB-BPP, a three-layer framework integrating Chaotic Particle Swarm Optimization (CFPSO), a Backpropagation Neural Network (BPNN) identifier, and a Predictive Functional Control (PFC)-based PID controller. At the top layer, CFPSO performs offline global optimization of the BPNN's initial weights and establishes safe bounds for controller parameters. The middle layer employs the BPNN for online system identification and real-time adaptation of PID parameters. The bottom layer executes high-precision control. This tiered architecture decouples offline optimization from online adaptation, enabling synergistic control. Comprehensive simulations demonstrate the superiority of CB-BPP, reducing the Integral of Absolute Error (IAE) by up to 45% and improving recovery speed by 35% compared with advanced baseline methods. Furthermore, extensive robustness analysis against sensor noise and parameter drift, alongside a successful Hardware-in-the-Loop (HIL) implementation, validates its practical applicability. The results confirm the proposed method's high precision, strong robustness, and real-time feasibility, providing an advanced control solution for complex industrial processes.

Aiming at the defect traceability challenges caused by the complexity of modern industrial process operation mechanisms and high-dimensional parameter coupling, a quality anomaly diagnosis framework integrating interpretable machine learning and knowledge graph is proposed. The quantitative contributions of process parameters to quality indicators are assessed by constructing an architecture combining XGBoost prediction model and SHAP (SHapley Additive exPlanation) interpretable analysis. Structured storage of variable association patterns of historical data based on a knowledge graph. The K Nearest Neighbors (KNN) algorithm is further introduced to construct a data matching mechanism to achieve causal traceability of real-time anomalies through similarity-driven historical defect retrieval. The experimental results show that the method has a good ability to identify the root cause parameters of defects under multiple working condition scenarios, and provides a solution with both data-driven characteristics and knowledge interpretability for the quality control of complex industrial processes.

URLLC is the key feature of 5G enabling extensive V2X applications and control of complex industrial processes via wireless communication systems. In order to ensure the highest reliability requirements with low latency communication, the fluctuating characteristics of a wireless link must not become a limiting factor for the use case. Cooperative communication approaches eliminate this problem by exploiting additional links via relay nodes to transmit data redundantly on different paths in the network to the destination node. We present a suitable multi-path scheme for integration in 5G and B5G systems. The theoretical analysis shows nearly one order of magnitude improvement of reliability for each additional path with the proposed architecture. Furthermore, we demonstrate a significant gain in terms of reliability for the overall link even for highly disrupted relay paths. Using experimental vehicle platooning measurements in real-world traffic based on LTE and IEEE802.11p, we confirm the conceptual advantage of our multi-path approach.

Actually, there is a trend of integrating simulation based components in enterprise information and decision support systems. The available data in production planning and control software systems can be used to construct the models, the objective-oriented realisation of simulation experiments and the economic evaluation of the results. In this context, the conception and realisation of integrated systems is very important for the analysis, modelling, simulation and optimisation of business processes. The paper describes an approach of the production scheduling problem under consideration of high complex process conditions. The new control strategies are based on hierarchical concepts. Local and global mechanisms are used for the order scheduling. These methods are integrated in a software system for short time production planning and control.

Artificial intelligence algorithms demonstrate exceptional potential for decision optimization in adaptive control of complex industrial processes. Addressing nonlinear coupling, signal fluctuations, and operational uncertainties in sand and gravel crushing processes, this study employs reinforcement learning to construct a multimodal perception-driven adaptive control framework. The objective is to achieve dynamic adjustment of crushing parameters and energy consumption optimization through intelligent algorithms, thereby enhancing system stability and precision. Methodologically, Deep Reinforcement Learning (DRL) integrates signal and image features to construct the state space. Based on the DQN algorithm, state perception, reward functions, and policy network structures are designed. Adaptive learning rates and online update mechanisms are introduced to enable real-time iteration of control strategies. Results demonstrate that during experimental platform validation, this model significantly reduces crushing particle size error by 57.3%, lowers energy consumption by 29.3%, and shortens response time by 46.4% compared to traditional PID and fuzzy control methods, achieving rapid convergence and stable control under complex operating conditions. Conclusions indicate that the reinforcement learning-based adaptive control architecture effectively enhances the intelligence level of sand and gravel crushing systems, providing an extensible and transferable algorithmic pathway for industrial process control with broad engineering application prospects.

The purpose of this article is to present an approach allowing the fault detection of a multivariate process with a bayesian network. As a discriminant analysis is easily modeled with a bayesian network, we will show that we we can consider the multivariate T² and MEWMA control charts as particular cases of the discriminant analysis. So, we give the structure of the bayesian network as well as the parameters of the network in order to detect faults in the multivariate space in the same manners as if we used multivariate control charts. The resulting bayesian network, with a computed threshold, is similar to the multivariate control charts.

An alternative technique for developing intelligent process monitoring and control systems is expert systems (ESs). ESs use logic rules to carry out heuristic reasoning. Knowledge used to reach a conclusion is transparent because ESs have explanation capabilities. In this respect, ESs are superior to neural networks because the knowledge embedded in neural networks is opaque. It has long been recognised that a critical issue in developing knowledge based ESs is the bottleneck of knowledge acquisition. Traditionally knowledge acquisition has been dependent on consulting domain experts. A disadvantage of this approach is that experts are usually better in collecting and archiving cases than in expressing the experience and cases explicitly into production rules. There are more difficulties in compiling knowledge about process operations due to the large number of interacting variables.

In modern industrial field, some complex processes have multi subprocesses, which are homo-structural variable-parameter systems and have quite a few control parameters. When the subprocesses have nonlinear characteristics, it is difficult to turn these control parameters manually. To solve this problem, a distributed immune algorithm for controlling the complex industrial process is designed on the basis of theory of immune network, which is able to learn control knowledge of different operators online and optimize the existing control knowledge by immune evolution and learning in terms of control results. Experiments and simulations show that it has the ability of distributed learning and its control results are superior to that of the manual. The algorithm can solve the problem of intelligent control for homo-structural variable-parameters systems in complex industrial process.

In recent decades, multivariate statistical methods for monitoring and controlling of industrial processes have received increasing attention. Multivariate process control charts such as Hotelling's T2 and multivariate exponentially weighted moving average have gained acceptance as manufacturing process control tools especially for the semiconductor industry. We propose in this paper a novel multivariate nonparametric control chart based on the permutation and nonparametric combination (NPC) methodology. The proposed NPC chart is always an exact inferential procedure for every finite sample size and overcomes some limitations of traditional approaches such as the ability to include a large number of variables. Moreover, NPC chart offers several advantages because it is a robust solution with respect to the true underlying random error distribution, and it is not affected by the problem of the loss of degrees of freedom when keeping fixed the number of observations. Unlike traditional methods, when the number of informative variables increases its power monotonically increases as well (finite sample consistency property of NPC test). As confirmed by an extensive simulation study and by an application to some real case studies in the field of microelectronics, the NPC charts are certainly good alternatives with respect to the traditional multivariate control charts such as Hotelling's T2 especially in cases of asymmetric or heavy‐tailed error distributions. Finally, NPC chart represents a suitable solution for monitoring a multivariate process when the number of observed process variables may be larger than the number of observations. Copyright © 2013 John Wiley & Sons, Ltd.

A New Approach In Learning For Intelligent Multi Agent Systems

The Sixth International Symposium on Neural Networks (ISNN 2009)

Multiple intelligences in a MultiAgent System applied to telecontrol

Multivariate Control Chart is an effective tool in Statistical Process Control to identify either an out-of-control process or in-control process. Hotelling T2 control chart is a quite popular and widely used technique in this field. However, its performance is deteriorated when the underlying distribution of the quality characteristics is not following the multivariate normal distribution. Multivariate control chart usually recommends a procedure in the phase-in Hotelling T2 control chart although there is difficulty in interpreting the signals from multivariate control charts more work is needed on data reduction methods and graphics techniques. Basically, the multivariate control chart refers to the theory of prediction interval. Therefore, it’s called predictive multivariate control charts. The purpose of this research is to construct what so-called predictive multivariate control chart both classical in this part in the phase-in Hotelling T2 and Copula-based ones. We argue that appropriate joint distribution function may be well estimated by employing Copula. A numerical analysis is carried out to illustrate that an Application Copula-based Multivariate control chart outperforms than bivariate control chart and others.