Amir M. Horr

Artificial Intelligence (AI) and Machine Learning (ML) are increasingly driving the digital transformation of manufacturing systems, enabling the transition from conventional process operation toward intelligent, adaptive, and data-centric production environments. This work presents the development of AI-enabled advisory systems for casting processes, integrating singular value decomposition (SVD)-based reduced-order models with a Variational Autoencoder with Arbitrary Conditioning (AC-VAE) and hybrid simulation frameworks to support real-time process prediction and optimization. The proposed approach leverages manufacturing data to establish predictive models capable of rapidly evaluating process conditions, optimizing operating parameters, and enhancing product quality while reducing material waste, energy consumption, and production costs. By combining physics-based understanding with AI-driven analytics, the framework facilitates real-time decision support, adaptive process control, and continuous performance improvement within modern manufacturing ecosystems. These capabilities contribute to the broader objectives of Industry 4.0 and emerging Industry 5.0 paradigms, including automation, connectivity, operational resilience, sustainability, and human-centered manufacturing. A representative Horizontal Direct Chill (HDC) continuous casting case study is presented to demonstrate the practical implementation of the framework, encompassing database generation, model training, validation, and deployment of predictive advisory tools for real-time manufacturing applications.

Artificial Intelligence (AI) and Machine Learning (ML) are increasingly driving the digital transformation of manufacturing systems, enabling the transition from conventional process operation toward intelligent, adaptive, and data-centric production environments. This work presents the development of AI-enabled advisory systems for casting processes, integrating advanced data-driven models, reduced-order modeling techniques, and hybrid simulation frameworks to support real-time process prediction and optimization. The proposed approach leverages manufacturing data to establish predictive models capable of rapidly evaluating process conditions, optimizing operating parameters, and enhancing product quality while reducing material waste, energy consumption, and production costs. By combining physics-based understanding with AI-driven analytics, the framework facilitates real-time decision support, adaptive process control, and continuous performance improvement within modern manufacturing ecosystems. These capabilities contribute to the broader objectives of Industry 4.0 and emerging Industry 5.0 paradigms, including automation, connectivity, operational resilience, sustainability, and human-centered manufacturing. A representative industrial casting case study is presented to demonstrate the practical implementation of the framework, encompassing database generation, model training, validation, and deployment of predictive advisory tools for real-time manufacturing applications.

Data science methodologies are playing an increasingly important role in advancing manufacturing systems, enabling improvements in efficiency, energy usage, cost reduction, product quality, and predictive maintenance capabilities. This raises a fundamental question: to what extent can data models reshape manufacturing processes, and what limitations prevent their full-scale adoption? Recent developments show a growing integration of data models within digital twin and digital shadow architectures, facilitating real-time monitoring and decision-making. Nonetheless, the complexity of industrial processes and the scarcity of high-quality, well-structured datasets pose significant challenges, particularly in terms of model robustness, interpretability, and scalability. Importantly, the effectiveness of such models depends more on data quality and representativeness than on data quantity alone. This review presents a structured analysis of data modeling techniques for manufacturing applications, with emphasis on data generation, sampling, preprocessing, and modeling approaches across diverse operational regimes, including steady, transient, and generative processes.

Data science techniques are increasingly employed to enhance process efficiency, reduce energy consumption and operational costs, enable active process control, ensure consistent product quality, and support predictive maintenance in modern manufacturing systems. A central question arising from recent developments is: How can data models fundamentally transform manufacturing processes, and what are the primary barriers to their widespread adoption? Contemporary manufacturing sectors are progressively integrating data models within digital twin and digital shadow frameworks to enable real-time process optimization and data-informed decision-making. However, the inherent complexity of manufacturing processes—combined with the frequent scarcity of high-quality, balanced datasets—often limits the generalizability and interpretability of purely data-driven models. In practice, quality, contextual relevance, representativeness, and richness of data are significantly more critical than its sheer volume when developing robust and reliable models. This paper provides a comprehensive overview of the application of data modeling in dynamic manufacturing environments. It examines key aspects such as data generation, sampling strategies, data preprocessing and handling, and model development methodologies across steady-state, transient, and generative process regimes.

The digital transformation of manufacturing is accelerating through the integration of artificial intelligence (AI), particularly via real-time predictive models. These models enable manufacturers to transition from reactive to proactive strategies, intelligent optimization and decision-making. Within the frameworks of Industry 4.0 and Industry 5.0, which emphasize technologies such as cyber-physical systems, cloud computing, and human-centric innovation, AI-driven data models are pivotal for achieving smart, adaptive, and sustainable production systems. This paper investigates the impact of AI-based predictive modeling on manufacturing digitalization and its future potential. It examines how these models contribute to advanced frameworks such as online process advisory systems, digital shadows, and digital twins, while addressing their limitations and implementation challenges. Furthermore, the study reviews current practices in real-time data modeling across manufacturing processes—including direct-chill casting—supported by real-world case studies. These examples illustrate both the practical benefits and technical hurdles of deploying AI in dynamic industrial environments.

The digital transformation of manufacturing is accelerating through the integration of artificial intelligence (AI), particularly via real-time predictive models. These models enable manufacturers to transition from reactive strategies to proactive, intelligent optimization and decision-making. Within the frameworks of Industry 4.0 and Industry 5.0, which emphasize technologies such as cyber-physical systems, cloud computing, and human-centric innovation, AI-driven data models are pivotal for achieving smart, adaptive, and sustainable production systems. This paper investigates the impact of AI-based predictive modeling on manufacturing digitalization and its future potential. It examines how these models contribute to advanced frameworks such as online pro-cess advisory systems, digital shadows, and digital twins, while addressing their limitations and implementation challenges. Furthermore, the study reviews current practices in real-time data modeling across manufacturing processes—including direct-chill casting—supported by real-world case studies. These examples illustrate both the practical benefits and technical hurdles of deploying AI in dynamic industrial environments.

Cyber and virtual representations of material processes have been promoted within the frameworks of digital twinning and shadowing for improved process monitoring and control. Real-time modelling practices based on sound data science techniques can be integrated into these digitalization frameworks to create efficient predictive and corrective analytics. However, modelling material processes involves multi-physical and multi-scale aspects that need to be conveyed for any optimization of product manufacturing. In this research work, a set of data solvers, interpolators, and trainers has been developed to model the macro-micro evolutions during material processes. Data search spaces were defined based on the process parameters. Snapshot matrices were created, and an appropriately sized database was generated to build real-time process models. Since the macro-micro data ranges for material processes can span over 20 orders of magnitude, a partitioned database technique has been employed to create accurate real-time models. The application of these combined data solvers, interpolators, and trainers has been developed for a real-world extrusion case study. The evolutions of macro and micro properties were predicted for material particles along predefined streamlines. These streamlines cover the entire distance from the back of the extrusion billet to the end of the final profile. In this way, the complete evolution of macro-micro material properties during the extrusion processes can be modelled.

As the use of data models and data science techniques in industrial processes grows exponentially, the question arises: to what extent can these techniques impact the future of manufacturing processes? This article examines the potential future impacts of these models based on an assessment of existing trends and practices. The drive towards digital-oriented manufacturing and cyber-based process optimization and control has brought many opportunities and challenges. On one hand, issues of data acquisition, handling, and quality for proper database building have become important subjects. On the other hand, the reliable utilization of this available data for optimization and control has inspired much research. This research work discusses the fundamental question of how far these models can help design and/or improve existing processes, highlighting their limitations and challenges. Furthermore, it reviews state-of-the-art practices and their successes and failures in material process applications, including casting, extrusion, and additive manufacturing (AM), and presents some quantitative indications.

As the use of data models and data science techniques in industrial processes grows exponentially, the question arises: to what extent can these techniques impact the future of manufacturing processes? This article examines the potential future impacts of these models based on an assessment of existing trends and practices. The drive towards digital-oriented manufacturing and cyber-based process optimization and control has brought many opportunities and challenges. On one hand, issues of data acquisition, handling, and quality for proper database building have become important subjects. On the other hand, the reliable utilization of this available data for optimization and control has inspired much research. This research work discusses the fundamental question of how far these models can help design and/or improve existing processes, highlighting their limitations and challenges. Furthermore, it reviews state-of-the-art practices and their successes and failures in material process applications, including casting, extrusion, and additive manufacturing (AM), and presents some quantitative indications.