KSEE 2026: From Intent to Impact - Systems of Systems Engineering with Digital Ways of Working

Title: Agentic AI: Orchestrating the Industrial System of Systems" (see note below *)

Abstract: This keynote explores the potential of Agentic AI as an orchestration layer for complex industrial “systems of systems.” Drawing on Aker BP’s AI-native strategy, we outline the proposed framework and building blocks—such as common context, secure tool access, and governance—necessary to move from simple interoperability to agentic cooperation. Through conceptual and early-stage use cases and development examples, this keynote illustrates how this emerging layer can help independent systems reason and act together safely in safety-critical environments. 

Title: The Systems Architect: the Realistic Visionary or the Visionary Realist?

Title: SFI CELECT and Systems Engineering

Title: Accelerating the Architecture Gradient: Human-in-the-Loop Agentic Frameworks for Agile UAF Modeling

Abstract: Traditional Model-Based Systems Engineering (MBSE) often struggles with an inherent agility bottleneck: the time and effort required to build, maintain, and translate complex models frequently outpaces the rapid decision-making needs of project stakeholders. This challenges the systems architect’s ability to efficiently bridge the gap between high-level Conceptual Visions and highly formalized System Architectures.

This presentation introduces a novel method implemented at Kongsberg Maritime that utilizes LLM-driven Agentic frameworks to dramatically accelerate the knowledge management pipeline. By deploying semi-autonomous, interactive agents, we have automated the mundane tasks of capturing unstructured information (e.g., whiteboards, documents, and sketches) and normalizing it directly into a structured Unified Architecture Framework (UAF) core.

Crucially, this method maintains a strict human-in-the-loop paradigm. Systems Engineers retain full control, validating and refining the agents' structural interpretations while dedicating their time to high-value analysis rather than manual data entry. Furthermore, the framework operates bidirectionally, allowing agents to instantly generate tailored, unstructured views back out to address individual stakeholder concerns across the engineering gradient.

The application of this method has collapsed traditional analysis timelines from weeks to hours. This drastic reduction in lifecycle friction unlocks a critical "fail-fast" mentality, empowering Systems Engineers to manage and analyze highly complex systems-of-systems interactions that were previously cost- and time-prohibitive.

Title: Intelligent Mechatronics for System of Systems Engineering

Abstract: A system of systems is only as capable as the machines inside it. Passive constituents push the burden of coordination onto integration; constituents intelligent enough to handle more themselves make the whole more tractable. Engineering such machines is itself starting to change. In software, AI already writes much of the code, and the engineer's effort shifts to deciding what to build; physical engineering is starting to follow. One reason it lags, the one this talk takes up, is that the knowledge is not in a form an AI can use: the models sit in tools made for people, and the reasoning behind them stays in the engineers' heads. The way forward is co-intelligence: with that knowledge made usable, the engineer becomes a systems thinker, framing the problem and judging the result while the AI does the detailed design. One such engineer can then span the disciplines a complex machine needs, where it once took a team. Mechatronics, where hardware and software meet, is where this is hardest and most valuable, and where the machines we build most shape what a system of systems can do.

Title: Behind the Dashboards: Asset Modelling, Contextualization and Stakeholder Alignment in an Industrial Asset Condition Monitoring System

Abstract: Industrial plants generate vast volumes of sensor data, yet it is rarely ready for the machine learning and AI organizations want to apply to it. Real-time, maintenance, and historic data often sit in siloed systems with inconsistent naming and little context. The layer that organizes and contextualizes this data is often missing. This layer is also what separates reactive maintenance from the condition-based and predictive approaches that industry is increasingly moving towards.

This presentation looks behind the dashboards of an Asset Condition Monitoring System (ACMS) for rotating equipment. The system was built on an industrial data platform during the Green Zinc Odda expansion at Boliden Odda, a Norwegian zinc plant, and covers roughly 250 machines and 800 sensors. The ACMS was the first system developed on this platform at Boliden. The platform itself is a first-of-its-kind combination of services and integrations, and the development of the ACMS played a central role in shaping the platform's design.

In the asset condition monitoring project, Knightec Group sat at the intersection of these efforts. In this presentation we discuss the engineering work users rarely see: modelling equipment, breaking the plant down into a standardized structure and naming scheme, and managing risks and dependencies across many connected systems. Much of it comes down to stakeholder alignment, and even small technical decisions require negotiation to reach a shared understanding.

Using the ACMS as a case, we ask where the real value lies: in the platform, or in the engineering and alignment that turn scattered data into a single source of truth that people, and eventually AI, can rely on?

Title: Systems Engineering, Kongsberg kommue

Title: Aluminium Post Consumer Scrap Sorting and Recycling to Develop a Circular Value Chain 
in the Automotive Supplier Industry

Abstract: Aluminium recycling is a key requirement for reducing environmental impact and promoting sustainable manufacturing in the automotive industry. Traditionally, recycling of aluminium has been focused on producing cast alloys, which limits the potential for high-performance applications. However, the automotive sector requires wrought aluminium alloys (5XXX, 6XXX, and 7XXX series) for structural components due to their strength, corrosion resistance, and lightweight properties. 
Wrought aluminium manufacturing is reliant on primary aluminium due to strict alloy purity requirements, which is energy-intensive and contributes to global carbon emissions. With growing demand and the shift toward electric vehicles, the automotive industry faces challenges in sourcing high-quality wrought aluminium alloys from post-consumer scrap (PCS), due to strict alloy composition requirements.

This PhD research addresses these challenges by investigating Laser Induced Breakdown Spectroscopy (LIBS), as an advanced sorting technology capable of separating aluminium scrap into specific wrought alloy families based on their elemental composition. An industrial-scale LIBS machine at Metallco Fredrikstad is being evaluated for its ability to sort wrought aluminium alloys from PCS aluminium. In parallel, a Life Cycle Assessment (LCA) is conducted to compare existing open-loop recycling practices  with the potential of closed-loop recycling using PCS aluminium. This analysis follows ISO standards to quantify environmental benefits, including reduced energy consumption and CO₂ emissions. This research also aims to develop a techno-economic and environmental assessment framework, integrating technical performance, cost modeling, and sustainability metrics. This framework will support decision - making for recyclers and manufacturers considering investment in advanced sorting technologies and circular production models.

Through experimental validation, environmental modeling, and academic dissemination the research  will contribute to both scientific knowledge and industrial practice. Ultimately, it seeks to establish an adaptable model for circular aluminium use in the automotive sector by enhancing sustainability, resource efficiency, and innovation across the supply chain.

Project Partners: Benteler Automotive, Metallco Aluminium AS, Norsk Hydro

Title: Cooperative Autonomous Drones for Preparedness: Enabling Operational Applications from Research.

Abstract:  This presentation introduces a PhD project within the INRESCOS research school focused on cooperative autonomous drone fleets for preparedness and critical operations. The work addresses the operational challenge of rapidly and efficiently covering large areas to improve detection probability while reducing response time and resource usage. These challenges are also relevant to search and rescue, surveillance, inspection, offshore incidents, environmental emergencies, and other large-area operations.

The research develops a coordination framework that enables multiple fixed-wing unmanned aerial vehicles (UAVs) to operate collaboratively with minimal human intervention. Using predictive decision-making and probability-based search maps, each drone continuously adapts its trajectory to focus effort where it is most valuable while avoiding redundant coverage. The approach aims to support scalable operations under realistic constraints such as limited communication, energy endurance, environmental uncertainty, and real-time decision requirements. The work currently focuses on high-fidelity simulation and system integration, with intellectual property under development.

Title: Modelling and Control of Self-Hoisting Crane for Offshore Wind Turbine Maintenance.

Abstract: The number of offshore wind turbines (OWTs) has significantly been increased in recent years due to the global demand for clean energy. However, replacing major components in these turbines is challenging and often requires jack-up vessels and heavy-lift cranes, which are expensive and complex to operate. A self-hoisting crane (SHC) is a type of crane designed to lift itself up along wires towards the top of a wind turbine tower, without the need for external lifting equipment or auxiliary cranes. SHC offers a simpler and more cost-effective alternative, enabling efficient installation and maintenance. However, SHCs are currently limited to onshore wind turbines, as their operation relies on fixed structures to provide stable guide wires for climbing the turbine. Adapting SHCs to offshore wind turbines remains an unexplored area of research, with challenges posed by dynamic marine environments.

The main objective of this work is to model and control the dynamic behaviour of SHC system when operated with floating vessels for major repair tasks on both bottom-fixed and floating OWTs. Floating vessels with floating wind turbines introduce unique complexities, including motion induced by environmental forces such as wind, waves, and currents. These factors significantly affect guide wire tension and create load-handling challenges between the wind turbine with the SHC mounted on top and the vessel. This work will further focus on developing advanced control strategies, such as Active Heave Compensation (AHC) and Dynamic Positioning (DP), to ensure stable, efficient, and reliable operations under dynamic environmental conditions.

Title: Biomethanation for Biogas Upgrading and Industrial CO₂ Conversion. Experimental development of energy-efficient trickle-bed reactors and integrated Power-to-Gas concept.

Abstract: The reduction carbon dioxide (CO2) into methane (CH4) offers a two-fold solution: producing value added product while preventing release of CO2 into the atmosphere. Conventional methanation processes are energy-intensive and economically challenging. Bio-methanation technology, operated at milder conditions, offers a promising alternative process to reduce energy demand. This renewable pathway can reduce fossil fuel dependency and mitigate emissions, aligning global goals for energy storage, grid injection, and industrial decarbonization.

In this study, three 1.2 L lab scale trickle bed reactors (TBRs) packed with six alternating layers of LECA and plastic media were operated for 30 days. H₂ utilization rose from 16.83–18.99 mmol L⁻¹ d⁻¹ (days 1–10) to 25.53–27.55 mmol L⁻¹ d⁻¹ (days 11–30) with high H₂ conversion (96.3–97%), while CH₄ production reached 7.72–8.27 mmol L⁻¹ d⁻¹. 

TBRs for biomethanation are currently at TRL level of 3-4 at laboratory scale and approaching TRL 6-7 level with further pilot scale demonstration. TBR shows promising aspects for future industrial solutions, including containment of toxic industrial flue gases such as H2S, CO and into value added products. Future work will focus on integration of TBR with other systems (electrochemical, anaerobic digestion), to further enhance methane yield.

Title: