DETOCS

The aim of Ramón Bandak Ramírez's thesis is to develop a model to predict the mineralogical composition of an OPC cement clinker in a cement plant, considering its chemical composition, type of raw materials, the mineralogical composition combined with relevant process parameters.

The first step is to set up a consistent set of Gibbs energies for all clinker phases accounting for minor components. This database is the basis for the clinkerization model using the Factsage Gibbs Energy minimization software and its Python interface. This database will also be used in the KilnSimu software, which allows a mass/heat transfer steady state modeling of a rotary kiln for clinker production.

In the second phase, the clinker database will be combined with hydration models using the GEMS software from the Paul Scherrer Institute in Switzerland, within the Marie-Curie network, and benefit from collaboration with RWTH Aachen in Germany and Cimentos Argos in Colombia.

The thesis is structured into two parts: 18 months at the SIMaP laboratory (CNRS / Université Grenoble Alpes) in France for model development and experiments, and 18 months at FLSmidth in Denmark to facilitate a more detailed integration with the industry.

The PhD project, titled "Thermal Activation of SCMs: Process Models and Quality Prediction," under DETOCS framework, aims to develop advanced process models for the thermal activation of Supplementary Cementitious Materials (SCMs), focusing on calcined clay. This project, conducted by Wang Qun, will integrate thermodynamic and heat transfer models to optimize kiln and flash calciner operations. Supervised by Dr. Wilson Ricardo Leal da Silva at FLSmidth A/S, Dr. Alexander Pisch at the University of Grenoble Alpes, and Prof. Marcus Campbell Bannerman at the University of Aberdeen, the research will link process conditions to material properties using machine learning, enhancing low-carbon cement production.

The PhD project, titled "Chemical Activation of SCMs via Carbonation and Admixtures” under the DETOCS framework, focuses on recycled cement paste (RCP) and industrial slags, specifically on how their properties can be improved through carbonation. The research, conducted by Sruthi Sreeram, aims to explore the influence of carbonation on the reactivity and rheology of the SCMs in order to develop a model that links material properties to the performance of the carbonated material. This project is supervised by Dr. Tim Wangler from ETH Zürich and Dr. Wilson Ricardo Leal de Silva from FLSmidth Cement.

In DC4, we propose a novel data-driven, application-based testing regime designed for integration into cement plants, which aims to optimize mechanical processing parameters for supplementary cementitious materials (SCMs). The approach involves identifying key material properties that influence the reactivity of mechanically activated SCMs, along with determining process parameters crucial for achieving desired material quality. By utilizing data collected through experimental work and process parameters, our objective is to accurately predict the performance of SCMs and then validate the reactivity of SCMs using this application-based testing methodology. This integrated approach promises significant advancements in the efficiency and applicability of SCMs in the construction industry.

This project explores innovative methodologies to enhance the sustainability of concrete production by incorporating recycled cement paste (RCP) as a secondary additive. Focused on reducing the industry's carbon footprint, the research integrates cutting-edge modelling techniques to predict and optimize the performance of concrete mixes. These models will be designed to evaluate early strength development and long-term durability, crucial for achieving the dual objectives of sustainability and structural reliability, as well as implementing an environmental impact assessment.

DianaGomez started as a PhD researcher in the #DETOCS project under the supervision of Dr. Abraham Gebremariam on the variability of materials on concrete performance, with special emphasis on the variation of #aggregate types on #concrete performance and their synergy of #SCMs in concrete. The study aims to identify critical material characteristics that influence concrete performance and develop guidelines or strategies for mitigating the downsides of material variability (including #recycledAggregates and SCMs) in the production of concrete. Her ultimate goal will be assessing such variations and providing practical solutions for achieving a more reliable and #sustainable #concreteProduction process.

The purpose of this project is to develop a methodology to unravel long-term durability behaviour of concrete from the paste-scale tests. Various durability-based testing methods with respect to chloride ingress and carbonation will be tested on blended systems at both paste and concrete scales. In addition, microstructural characterization will be used to identify the key durability indicators at the paste scale for rapid assessment of concrete durability. Furthermore, the paste to concrete predictions will be validated based on lab tests and concrete plants data. Overall, this thesis will potentially probe the key aspect of predicting concrete properties from cement paste scale tests in a reliable and fast manner, which could pave the path for rapid adoption of novel low-carbon cements.

The purpose of this project is to build mathematical systems capable of increasing the level of control and interpretability of cement production processes, by developing soft-sensing and interpretable anomaly detection systems to be implemented in real life processes. Soft-sensors are software sensors that take data from the process and using their programming directives predict in real time a desired set of properties that are otherwise costly and time consuming to evaluate using laboratory techniques. The core of a soft-sensors is the predictive models embedded in it, models that the DC9 project aims to develop by identifying the best machine-learning techniques capable of accurately describing data typically collected in the cement industry. Using these models and the developed soft-sensors, the next step is to build monitoring and interpretable anomaly detection systems capable of real-time detection of anomalies in the cement production processes allowing for timely interventions on the process operators side that can implement solutions for the detected problems, hence preventing potential process unscheduled shutdowns and save product quantities that would otherwise not meet the specifications required by the industry. Thus, the DC9 project aims at using computational techniques, especially machine learning, to increase the level of control and operability of cement production processes and increase the degree of digitalisation of the cement industry.

Muhammad Imran Waris started his PhD on 01-05-2024, working in the Resource and Recycling group at TU Delft under the supervision of Dr. Francesco Di Maio. His research focuses on the traceability of concrete materials along the value chain. His objective is to develop and apply a material passport based on Radiofrequency Identification (RFID) to monitor and track the quality of concrete materials. Furthermore, he aims to develop an online database of the material’s properties and integrate it with the RFID to access the material’s information at any stage of the value chain. He is working in collaboration with Cementos Argos, C2CA, and Mannok Holdings.

Machine Learning (ML) is well-established for process monitoring in other industries, but data-usage in cement production remains challenging: The processing of solid material is intrinsically complex to describe, while both the variability of raw material characteristics as well as process conditions affect product quality. DC11 aims to develop a framework for ML in cement and concrete production, incorporating both process data and data of raw and intermediate materials to enhance product quality and formulations and maximize the use of SCM, while reducing energy consumption, emissions, costs and variability in the final product. Supervised ML will be applied on quality estimation; in particular, latent variable model inversion for product optimization and explainability for identification of main drivers for variability. Further techniques include Data Mining, ML-based prognostics and unsupervised ML for advanced diagnostics. The developed ML framework will combine knowledge from first-principles models, raw data and soft-sensed data.

Several non-technical issues constrain the end users' adoption of innovative low-carbon cement blends. Prescriptive cement standards are identified as one of the most important barriers, but industry- and firm-specific characteristics, such as end-user avoidance of technical and market risks and a high level of price sensitivity, also play significant roles. These factors, together with high investment and learning costs, are crucial considerations in decarbonizing the concrete industry. However, the specific barriers to adopting low-carbon emission cement blends are largely understudied.

The objective of this doctoral mission is to determine how industry regulations and standards affect the adoption of low-carbon cement blends. Additionally, it aims to assess how industry-specific characteristics influence the adoption of low-carbon cement blends and to identify which firm-level internal characteristics impact their adoption. This interdisciplinary work draws key concepts from political economy and sciences, evolutionary economics, sustainability transitions, and innovation management.

The DC12 operates within Work Package 4 on market acceptance of low-carbon cement. It is expected to deliver: 1) a comprehensive list of bottlenecks related to low-carbon cement market acceptance, 2) solid evidence of the environmental and economic impacts of concrete made with low-carbon cement, and 3) a report on “The Future of the Cement and Concrete Industry".

This project aims to explore the measurements of particle properties through image analysis and correlate these to the chemistry and composition of SCMs and their reactivity in low carbon cements. The project focus is to develop an image analysis method to capture SCM dissolution and integrate a hyperspectral imaging technique for characterizing the reactivity of SCMs. By correlating SCM composition and reactivity with optical data, the project seeks to provide deeper insights into SCM behavior and performance. The integration of advanced optical methods will pave the way for a rapid SCM characterization method, contributing to the ongoing efforts of reducing carbon emissions in the construction industry.

This project aims to develop digital and AI-driven technologies to optimize the production of high-quality secondary cementitious materials (SCMs) from construction and demolition waste which may include but not be limited to concrete. It involves evaluating input material quality using AI and sensor data, identifying physical correlations through literature review, numerical modeling, physical modeling, Systems identification, and experiments. Consequently, Integrating these insights into a process model for optimal control encompassing robust-adaptive techniques. The model will be tested and validated in C2CA Technology's flagship industrial plant. The project's outcome will facilitate the automatic evaluation of diverse input materials and enhance the production of SCMs, promoting sustainable concrete practices.

There are a multitude of potential alternative cementitious materials which can be used to reduce the clinker-to-cement ratio, which are at varying levels of industrial readiness. DC15 seeks to define a landscape of commercially competitive novel cement blends and processes which improve the emissions profile and environmental performance of the cement industry, particularly as cost-competitive alternatives are essential for decarbonisation in developing economies. Life Cycle Assessment (LCA) and Techno-Economic Analysis (TEA) will be performed based on detailed assessments of upstream processes for low carbon cement blends. Prospective LCA, an emerging approach based on process modelling, will be leveraged to generate system inventory data in lieu of full-scale operational data. In addition to CO2 emissions, other environmental impact categories (e.g., ecotoxicity, particulate emissions) will be studied to ensure a holistic view of the environmental impacts is given. A typical production site will be defined as the basis for economic analysis, which will include both CAPEX (e.g., new unit installation, additional machinery, equipment upgrades) and OPEX (e.g., staffing, transport, raw materials, disposal) cost considerations.