Micro-pin-fin evaporators are a promising alternative to multi-microchannel heat sinks for two-phase cooling of high power-density devices. Within pin-fin evaporators, the refrigerant flows through arrays of obstacles in cross-flow and is not restricted by the walls of a channel. The dynamics of bubbles generated upon flow boiling and the associated heat transfer mechanisms are expected to be substantially different from those pertinent to microchannels; however, the fundamental aspects of two-phase flows evolving through micro-pin-fin arrays are still little understood. This article presents a systematic analysis of flow boiling within a micro-pin-fin evaporator, encompassing bubble, thin-film dynamics and heat transfer. The flow is studied by means of numerical simulations, performed using a customised boiling solver in OpenFOAM v2106, which adopts the built-in geometric Volume of Fluid method to capture the liquid–vapour interface dynamics. The numerical model of the evaporator includes in-line arrays of pin-fins of diameter of 50μm and height of 100μm, streamwise pitch of 91.7μm and cross-stream pitch of 150μm. The fluid utilised is refrigerant R236fa at a saturation temperature of 30 °C. The range of operating conditions simulated includes values of mass flux G=500–2000kg/(m²s), heat flux q=200kW/m², and inlet subcooling ΔT_sub=0–5K. This study shows that bubbles nucleated in a pin-fin evaporator tend to travel along the channels formed in between the pin-fin lines. Bubbles grow due to liquid evaporation and elongate in the direction of the flow, leaving thin liquid films that partially cover the pin-fins surface. The main contributions to heat transfer arise from the evaporation of this thin liquid film and from a cross-stream convective motion induced by the bubbles in the gap between the cylinders, which displace the hot fluid otherwise stagnant in the cylinders wakes. When the mass flow rate is increased, bubbles depart earlier from the nucleation sites and grow more slowly, which results in a reduction of the two-phase heat transfer. Higher inlet subcooling yields lower two-phase heat transfer coefficients because condensation becomes important when bubbles depart from the hot pin-fin surfaces and reach highly subcooled regions, thus reducing the two-phase heat transfer.

In annular downward flow, an annular liquid film flows at the perimeter of the channel pushed down by the gravity force and by the shear stress that the vapor core exerts on it. Depending on the working conditions, the vapor-liquid interface can be flat or rippled by waves. The knowledge of the liquid film thickness is very important for the study of annular flow condensation because the thermal resistance of the liquid is often the most important parameter controlling the heat transfer. A new approach for the simulation of annular flow is here proposed using an in-house developed transient solver based on the Volume of Fluid (VOF) adiabatic solver interIsoFoam available in OpenFOAM. With the VOF method, in addition to the standard set of equations (continuity and momentum), a transport equation related to the advection of the volume fraction scalar field has to be solved. The numerical setup consists of 2D axisymmetric domain. An adaptive mesh refinement (AMR) method is added to the solver to better capture the interface position. The k-ω SST model is used for turbulence modelling in both the liquid and vapor phases and a source term (whose magnitude is controlled by a model parameter named B) is included in the ω equation to damp the turbulence at the interface.

Background and objective:This paper focuses on nutritional recommendation systems (RS), i.e. AI-powered automatic systems providing users with suggestions about what to eat to pursue their weight/body shape goals. A trade-off among (potentially) conflictual requirements must be taken into account when designing these kinds of systems, there including: (i) adherence to experts’ prescriptions, (ii) adherence to users’ tastes and preferences, (iii) explainability of the whole recommendation process. Accordingly, in this paper we propose a novel approach to the engineering of nutritional RS, combining machine learning and symbolic knowledge extraction to profile users—hence harmonising the aforementioned requirements. MethodsOur contribution focuses on the data processing workflow. Stemming from neural networks (NN) trained to predict user preferences, we use CART Breiman et al.(1984) to extract symbolic rules in Prolog Körner et al.(2022) form, and we combine them with expert prescriptions brought in similar form. We can then query the resulting symbolic knowledge base via logic solvers, to draw explainable recommendations. ResultsExperiments are performed involving a publicly available dataset of 45,723 recipes, plus 12 synthetic datasets about as many imaginary users, and 6 experts’ prescriptions. Fully-connected 4-layered NN are trained on those datasets, reaching ∼86% test-set accuracy, on average. Extracted rules, in turn, have ∼80% fidelity w.r.t. those NN. The resulting recommendation system has a test-set precision of ∼74%. The symbolic approach makes it possible to devise how the system draws recommendations. ConclusionsThanks to our approach, intelligent agents may learn users’ preferences from data, convert them into symbolic form, and extend them with experts’ goal-directed prescriptions. The resulting recommendations are then simultaneously acceptable for the end user and adequate under a nutritional perspective, while the whole process of recommendation generation is made explainable.

Building on prior works on explanation negotiation protocols, this paper proposes a general-purpose protocol for multi-agent systems where recommender agents may need to provide explanations for their recommendations. The protocol specifies the roles and responsibilities of the explainee and the explainer agent and the types of information that should be exchanged between them to ensure a clear and effective explanation. However, it does not prescribe any particular sort of recommendation or explanation, hence remaining agnostic w.r.t. such notions. Novelty lays in the extended support for both ordinary and contrastive explanations, as well as for the situation where no explanation is needed as none is requested by the explainee. Accordingly, we formally present and analyse the protocol, motivating its design and discussing its generality. We also discuss the reification of the protocol into a re-usable software library, namely PyXMas, which is meant to support developers willing to build explainable MAS leveraging our protocol. Finally, we discuss how custom notions of recommendation and explanation can be easily plugged into PyXMas.