Politecnico di Milano

Digital Twins (DTs) are enablers for the fast optimisation processes required in the Industry 4.0 context. Declarative equation-based modelling languages, in turn, enable the creation of large-scale simulation-based DTs, as they relieve the analyst from creating the solution code. However, most industrial assets are Cyber-Physical Systems (CPSs), the Cyber part being their digital controls. With the available technology, a precise representation of modulating and logic controls conflicts with DT simulation performance. The result is a barrier to using DTs for system-level optimisation. We analyse the problem, propose a modelling paradigm to solve it and suggest how to integrate that paradigm into equation-based language compilers. We support our proposal by presenting a Modelica/C++ library, that we release as free software, built according to the said paradigm.

Run-time resource management is fundamental for efficient execution of workloads on Chip Multiprocessors. Application- and system-level requirements (e.g. on performance vs. power vs. lifetime reliability) are generally conflicting each other, and any decision on resource assignment, such as core allocation or frequency tuning, may positively affect some of them while penalizing some others. Resource assignment decisions can be perceived in few instants of time on performance and power consumption, but not on lifetime reliability. In fact, this latter changes very slowly based on the accumulation of effects of various decisions over a long time horizon. Moreover, aging mechanisms are various and have different causes; most of them, such as Electromigration (EM), are subject to temperature levels, while Thermal Cycling (TC) is caused mainly by temperature variations (both amplitude and frequency). Mitigating only EM may negatively affect TC and vice versa. We propose a resource orchestration strategy to balance the performance and power consumption constraints in the short-term and EM and TC aging in the long-term. Experimental results show that the proposed approach improves the average Mean Time To Failure at least by 17% and 20% w.r.t. EM and TC, respectively, while providing same performance level of the nominal counterpart and guaranteeing the power budget.

Micro-Electro-Mechanical Systems (MEMS) accelerometers are entering high-end applications thanks to their improved performance and low costs. Biaxial sensors able to measure two in-plane components of the external accelerations by exploiting a single proof-mass have been recently proposed and optimized both mechanically and electronically in order to minimize cross-axis sensitivity, while preserving a good symmetry between the two axes and high performance. To the authors best knowledge, only very few commercial high-performance xz - or yz - biaxial MEMS accelerometers are available so far. In this work, we propose an innovative design strategy for xz -biaxial MEMS capacitive accelerometers immune from electrostatic nonlinearities and pull-in instabilities usually related to out-of-plane readout schemes. In particular, thanks to the proposed motion conversion mechanism realizable through the Thelma-Double fabrication process of STMicroelectronics, we predict a sensitivity bigger than 30fF/g on both axes, a cross-axis sensitivity smaller than 0.04% and a nonlinearity lower than 1% at 50g.

The work presents a sensor including a piezoresistive, nano-gauge-based, microelectromechanical system (MEMS) gyroscope with high-performance, coupled to an integrated circuit designed in a 130-nm CMOS process. The system works in amplitude-modulated mode-split conditions and features current-feedback instrumentation amplifiers as low-noise front-end stages. The circuit also embeds an automatic quadrature compensation loop and a newly designed programmable demodulation delay chain, with 16 ns steps, to compensate for phase lags introduced in the electro-mechanical and electronics domains in this operation mode. The circuit draws 9 mA from a 3.6 V supply, with no need for voltage boosting. Coupled to the sensor, it yields 250 μdps/√Hz noise with sub-0.2% linearity error over 225 dps.

The paper presents the process underpinning the development of a framework for architectural teaching and learning aimed to facilitating the implementation of the sustainable development goals (SDGs) in architectural teaching, with particular focus on design courses. It aims to offer a methodology of evaluation and self-assessment for architecture educators in higher education institutions. This study has been conducted in accordance with the principles set out at the international level by UNESCO and at a European level by the Green Deal and the New European Bauhaus initiative. The theoretical apparatus presented herewith in has been (and continues to be), formalised based on the experience gained teaching the course “Sustainability and the Built Environment” at the Politecnico di Milano (Mantua Campus) since 2018. The course has provided the academic framework to experiment with a multi-system approach to Education for Sustainable Development (ESD). The feedback collected and the results obtained have enabled the identification of three different perspectives to be considered when setting up a teaching programme, namely: the relationship between the teaching institution, the educators, and the students; the relationship between the students and their own agency within the learning context; and the teaching content. The general principles deduced from this experience, which is still ongoing, are transferable beyond the outlined disciplinary limits and are intended to provide students with valuable alternative perspectives to approach vital an ecological transition.

In the age of industry 4.0, digital platforms are becoming highly influential with their ability to facilitate the exchange, cooperation and coordination among enterprises. In spite of the significant role of digital platforms, only few research works directly explain their development and applications. This paper aims to provide an overview of digital platforms and their impact on agility using a systematic literature review. The results provide insights into digital platforms’ capabilities, challenges, and synergies with agility in the manufacturing sector.

Robotic therapies are receiving growing interest in the autism field, especially for the improvement of social skills of children, enhancing traditional human interventions. In this work, we conduct a scoping review of the literature in robotics for autism, providing the largest review on this field from the last five years. Our work underlines the need to better characterize participants and to increase the sample size. It is also important to develop homogeneous training protocols to analyse and compare the results. Nevertheless, 7 out of the 10 Randomized control trials reported a significant impact of robotic therapy. Overall, robot autonomy, adaptability and personalization as well as more standardized outcome measures were pointed as the most critical issues to address in future research.

Sustainable development and intergenerational responsibility entail the prudent use of natural resources. Water availability can constrain agriculture, a key sector in terms of resources consumed and goods and services provided. The sustainability of its intensification and expansion has been studied, often with a particular focus on water. Agricultural strategies have been based on local water availability, and some downstream effects have been evaluated. However, a method to identify and quantify hydrologically sustainable land use and crop use changes directly accounting for downstream effects is yet to be defined. We propose a framework to design land‐use and crop‐use changes preventing local and downstream effects. We apply it on of coffee plantations expansion in Kenya, a sector that is growing and planned to grow, given its agricultural, economic and social development potential, not without risks associated to hydroclimatic change. We use crop‐ and land‐use specific hydrological modeling to simulate water scarcity impacts of coffee plantation expansion onto available suitable areas, and use the results to iteratively identify and filter out expansion areas increasing water scarcity locally or downstream. This assessment proves effective in preserving water availability, identifying 10% of the suitable and available areas as hydrologically sustainable. Total water footprints are similar in these expansion areas and in currently used areas, but expansion areas have higher precipitation‐generated water availability. The proposed methodology locates and quantifies areas in a physically robust way, maintaining flexibility to the selected expansion scenario. Thus, the methodology is replicable for planning hydrologically agricultural development. This article is protected by copyright. All rights reserved.

Cast-in fasteners, like headed-studs, are important elements widely used in safety-critical applications of the building industry. They allow the connection of structural components through transfer of stresses from load-bearing elements to concrete. Their tensile strength in an unconfined configuration with wide supports depends on the concrete’s mechanical properties. Despite numerous studies performed on the short-term behaviour of cast-in anchors, little information is available on their sustained-load behaviour and the effect of the loading rate on their load capacity. The present research aims at studying these two effects by performing an experimental investigation consisting of sustained load tensile tests on cast-in headed studs. Firstly, short-term tests at different loading rates were performed. Secondly, long-term sustained load tests were performed at different load levels with respect to the ultimate load capacity. Two sets of anchors installed in two different concretes were tested. The first set consisted of 10 anchors tested to failure at different loading rates while sustained load tests were executed on additional 11 anchors. The second consisted of 14 anchors, 3 of which were tested to failure to determine their ultimate capacity and the remaining anchors were subjected to sustained loads at different load levels. The corresponding displacements and time-to-failure were continuously measured throughout the long-term tests. The results were used to construct time-to-failure curves where the load level is plotted against the time-to-failure in a semi logarithmic scale. The lifetime prediction of the anchors was assessed by applying a new model based on a sigmoid function. The predicted sustained load values for a 50-year service life are noticeably lower than the short-term capacity but remain larger than the characteristic load calculated according to standards for the design of cast-in anchors.

Policy-based algorithms are among the most widely adopted techniques in model-free RL, thanks to their strong theoretical groundings and good properties in continuous action spaces. Unfortunately, these methods require precise and problem-specific hyperparameter tuning to achieve good performance, and tend to struggle when asked to accomplish a series of heterogeneous tasks. In particular, the selection of the step size has a crucial impact on their ability to learn a highly performing policy, affecting the speed and the stability of the training process, and often being the main culprit for poor results. In this paper, we tackle these issues with a Meta Reinforcement Learning approach, by introducing a new formulation, known as meta-MDP, that can be used to solve any hyperparameter selection problem in RL with contextual processes. After providing a theoretical Lipschitz bound to the difference of performance in different tasks, we adopt the proposed framework to train a batch RL algorithm to dynamically recommend the most adequate step size for different policies and tasks. In conclusion, we present an experimental campaign to show the advantages of selecting an adaptive learning rate in heterogeneous environments.

This paper investigates the influence of tiltrotor blade twist on whirl-flutter stability boundaries. Preliminary evaluations indicate that the whirl-flutter speed can be increased if the blade twist slope is reduced. This positive effect results from the shift in the overall thrust toward the blade tip, increasing the flapwise bending moment at the blade root and the trim coning angle. This, in turn, generates a positive pitch-lag coupling, increasing the whirl-flutter speed. However, the shift of high sectional thrust forces toward the blade tip sections returns a higher induced drag, showing the tendency to increase the power required. The paper shows that, by using blade twist laws based on piecewise linear functions and adding the wing airfoil thickness as a second design parameter, it is possible to identify aircraft configurations that improve the whirl-flutter stability boundaries without penalizing the power required in airplane and helicopter mode flight. This is possible because the blade twist and the wing airfoil thickness have an impact on both power required and whirl-flutter speed, so a simple optimization algorithm can identify good tradeoffs. A detailed tiltrotor model representative of the Bell XV-15 is used to display the effectiveness of the proposed approach. The examples show that increases up to 21% on the whirl-flutter speed are achievable without penalties in the aircraft power required and with the additional benefit of a benign impact on rotor pitch link loads.

The rapid increase in digital transactions has led to a consequential surge in financial fraud, requiring an automatic way of defending effectively from such a threat. The past few years experienced a rise in the design and use by financial institutions of different machine learning-based fraud detection systems. However, these solutions may suffer severe drawbacks if a malevolent adversary adapts their behavior over time, making the selection of the existing fraud detectors difficult. In this paper, we study the application of online learning techniques to respond effectively to adaptive attackers. More specifically, the proposed approach takes as input a set of classifiers employed for fraud detection tasks and selects, based on the performances experienced in the past, the one to apply to analyze the next transaction. The use of an online learning approach guarantees to keep at a pace the loss due to the adaptive behavior of the attacker over a given learning period. To validate our methodology, we perform an extensive experimental evaluation using real-world banking data augmented with distinct fraudulent campaigns based on real-world attackers’ models. Our results demonstrate that the proposed approach allows prompt updates to detection models as new patterns and behaviors are occurring, leading to a more robust and effective fraud detection system.

We analyze markets for cryptoassets (cryptocurrencies and stablecoins), investigating market impact and efficiency through the lens of the market order flow. We provide evidence that markets where cryptoassets are exchanged between themselves play a central role on price formation and are more efficient than markets where cryptocurrencies are exchanged with the US dollar. For the first set of markets we observe some evidence of the presence of insiders/contrarians, instead in the latter we observe the predominance of herding and trend-followers.

In this study, a 3D physics-based numerical approach, based on the spectral element numerical code SPEED, is used to simulate seismic wave propagation due to a local earthquake in the Mexico City area. The availability of detailed geological, geophysical, geotechnical, and seismological data allowed for the creation of a large-scale (60 km × 60 km in plan, 10 km in depth) heterogeneous 3D numerical model of the Mexico City area, dimensioned to accurately propagate frequencies up to about 1.3 Hz. The results of numerical simulations are validated against the ground motion recordings of the July 17, 2019, Mw3.2 earthquake, with peak ground acceleration exceeding 0.3 g about 1 km away from the epicenter. A good agreement with records is found, quantitatively evaluated through goodness-of-fit checks. Furthermore, for the lake zone, the simulated decay trend of the peak ground velocity with epicentral distance is reasonably close to the observations, for both horizontal and vertical components. In spite of some limitations, the simulations are successful to provide a realistic picture of seismic wave propagation both in the hill and in the lake zones of Mexico City, including the onset of long-duration quasi-monochromatic ground motion in the basin, with strong amplification at low frequencies (between 0.4 and 0.7 Hz). The numerical results also suggest that surface waves, with predominant prograde particle motion at the ground surface and large ellipticities, dominate the wavefield in the lake zone. Based on these positive outcomes, we conclude that this numerical model may be useful for both a better insight into the seismic response of the Valley of Mexico and the simulation of ground motions during larger-magnitude earthquakes, to generate improved seismic damage scenarios in Mexico City.

The construction industry plays a critical role in tackling the challenges of climate change, carbon emissions, and resource consumption. To achieve a low-emission built environment, urgent action is required to reduce the carbon emissions associated with steel production and construction processes. Reusing structural steel elements could make a significant impact in this direction, but there are five key challenges to overcome: limited material availability, maximizing different reusable materials from demolition, lack of adequate design rules and standards, high upfront costs and overlooked carbon impact of the demolition prior to construction, and the need to engage and coordinate the complete construction ecosystem. This article described these barriers and proposed solutions to them by leveraging the digital technologies and artificial intelligence. The proposed solutions aim to promote reuse practices, facilitate the development of certification and regulation for reuse, and minimize the environmental impact of steel construction. The solutions explored here can also be extended to other construction materials.

Ladle furnace slags are characterized by volumetric expansions associated with the transition of dicalcium silicate (C 2 S) from β to γ phase, which generates fine dust during cooling, causing handling and storage issues that further reduce their recycling opportunities. The present work focuses on the effect of slag basicity on dusting and the role of sulfur on slag stability. Seven synthetic ladle slag precursors were made by mixing lime, magnesia, quartz and alumina in different proportions to match effective industrial compositions, increasing the binary basicity and keeping the ternary and quaternary indexes unchanged. Samples were heated to 1500 °C for 15 min and monitored during air cooling (< 5 °C/s) through thermocouples and camera to characterize the behavior, temperature, and time interval of dusting. The cooled samples were characterized chemically, mineralogically and morphologically. Starting from the chemistry of a self-stabilized slag, five additional slag precursors, characterized by increasing amounts of S, were created and analyzed using the same procedures. Experimental evidence showed the presence of three different dusting behaviors (stable, partial and complete) and stabilization of the slag once an optical basicity of 0.748 or higher was reached. In addition, mayenite was identified as the main phase capable of suppressing the β to γ transition by exerting hydrostatic pressure on C 2 S. Finally, although S can stabilize the β phase when dissolved in it, after saturation it precipitates as CaS, which can react with mayenite, locally decreasing the optical basicity and allowing dusting. Graphical Abstract