Periodic Reporting for period 1 - E-COOL (A HOLISTIC APPROACH OF ELECTRIC MOTOR COOLING)
Periodo di rendicontazione: 2024-03-01 al 2025-02-28
According to the International Energy Agency (IEA), electric motors (e-motors) consume more than 40% of electricity produced globally. The EU has set bold targets (regulation 2019/1781, 2nd phase effective from 01.07.2023) for increasing their efficiency and aiming to save 110TWh by 2030; this corresponds to 40 Mt of CO2 emissions per year with today’s electricity mix, which is equivalent to the electricity consumption of the Netherlands. E-motors are also the driving force behind Electric Vehicles (EVs), currently leading the global efforts for decarbonisation of the transportation sector; e-motor efficiency is crucial in extending EV mileage. Unfortunately, electrification of commercial or heavy-duty vehicles, earth-moving machines and aircrafts have to overcome, among other significant limitations (such efficient energy storage and battery weight, safety, limited availability of carbon-free fuels, cost etc), the technological barrier of excess e-motor heat generated during power-demanding operations. The long-term vision of E-COOL is to address this challenge via the development of a breakthrough direct-contact, spray cooling system. The envisioned technology will provide unprecedented cooling rates at the local temperature hot spots and result to an average 20% increase in e-motor efficiency compared to today’s state-of-the-art. In an equally important manner, the developed technology will also allow utilisation of e-motors in power-demanding applications, currently prohibited by the excess temperatures generated. These applications expand over the whole range of transportation sectors and, thus, they will contribute to EU’s decarbonization strategy, via significant additional energy and CO2 savings relative to the ones currently envisioned from existing regulations. Moreover, E-COOL aspires to achieve this technological breakthrough at time-scales compatible to those required for industrial innovations to reach the market.
Work has progressed in three work packages.
In WP1 benchmark coolants have been synthesised and their critical rheological and transport properties (e.g. shear/elongational viscosity and equilibrium/dynamic surface tension) have been determined. Well-characterised coolants incorporating various types of additives in a range of concentrations have been developed (CITY). In a parallel activity, the test rigs for spray visualisation and impact on heated targets have been set-up and instrumented (LU). Appropriate nozzles, sensors, imaging and laser systems have been configured and tested. Temperature measurement techniques and sensors have been assessed relevant to the heated targets. Finally, one linear and one e-motor have been designed, assembled (ICCS) and instrumented with various temperature sensors. The first is used at ICCS as a benchmark case for conventional cooling, while the second incorporates the spray cooling system; it has optical access so it will allow spray visualisations.
In WP2, MD simulations (CITY) have been performed, complementing the experiments of WP1 for operating points for which measurements will be either time consuming or with low accuracy. Parallel to this activity, an ML platform for accelerating the stress tensor calculation is under development at (CITY) for the ML-enabled constitutive equation providing the mathematical relationship between the strain-rate history and the corresponding tensorial stress. On the spray modelling side, theoretical constitutive models, accounting for the contribution of polymeric chains in the stress tensor of the momentum conservation equation and embedding the ML term have been implemented in OenFOAM (OVGU). This CFD modelling framework has been utilised to simulate from ‘first-principles’ the size distribution of the structures (ligaments and satellite droplets) forming during near-nozzle atomisation/fragmentation processes for the examined cooling fluids. On-the-fly adaptive grid refinement has been employed to provide the required resolution near the interface of the atomising liquid. Additional transport equations for the liquid-air interface surface area and its generation rate during atomisation have been included for resolving structures smaller than the grid resolution. Sufficient number of numerical simulations will allow the formulation of a SGS viscoelastic-atomisation model for predicting jet/spray primary break-up, as function of the liquid properties and injector geometry.
In WP3 numerical simulations have been performed by partners AVL-AT and AVL-SLO who have developed the complex numerical grids required for considering the geometries of the motor windings. Analytical functions for surface parametrisation (e.g. surface approximation polynomials such as Bezier or Splines) required by the ML algorithm to facilitate continuous input values have been implemented. Their CFD solver AVL-FIRE is currently extended to incorporate the atomisation SGS model developed in the previous WP. Preliminary simulations for the conjugate heat transfer between the impacting liquid and the heated targets have been performed. These datasets will be used for the training of the ML-tool, which will be also developed in this WP (CITY).
In WP1 benchmark coolants have been synthesised and their critical rheological and transport properties (e.g. shear/elongational viscosity and equilibrium/dynamic surface tension) have been determined. Well-characterised coolants incorporating various types of additives in a range of concentrations have been developed (CITY). In a parallel activity, the test rigs for spray visualisation and impact on heated targets have been set-up and instrumented (LU). Appropriate nozzles, sensors, imaging and laser systems have been configured and tested. Temperature measurement techniques and sensors have been assessed relevant to the heated targets. Finally, one linear and one e-motor have been designed, assembled (ICCS) and instrumented with various temperature sensors. The first is used at ICCS as a benchmark case for conventional cooling, while the second incorporates the spray cooling system; it has optical access so it will allow spray visualisations.
In WP2, MD simulations (CITY) have been performed, complementing the experiments of WP1 for operating points for which measurements will be either time consuming or with low accuracy. Parallel to this activity, an ML platform for accelerating the stress tensor calculation is under development at (CITY) for the ML-enabled constitutive equation providing the mathematical relationship between the strain-rate history and the corresponding tensorial stress. On the spray modelling side, theoretical constitutive models, accounting for the contribution of polymeric chains in the stress tensor of the momentum conservation equation and embedding the ML term have been implemented in OenFOAM (OVGU). This CFD modelling framework has been utilised to simulate from ‘first-principles’ the size distribution of the structures (ligaments and satellite droplets) forming during near-nozzle atomisation/fragmentation processes for the examined cooling fluids. On-the-fly adaptive grid refinement has been employed to provide the required resolution near the interface of the atomising liquid. Additional transport equations for the liquid-air interface surface area and its generation rate during atomisation have been included for resolving structures smaller than the grid resolution. Sufficient number of numerical simulations will allow the formulation of a SGS viscoelastic-atomisation model for predicting jet/spray primary break-up, as function of the liquid properties and injector geometry.
In WP3 numerical simulations have been performed by partners AVL-AT and AVL-SLO who have developed the complex numerical grids required for considering the geometries of the motor windings. Analytical functions for surface parametrisation (e.g. surface approximation polynomials such as Bezier or Splines) required by the ML algorithm to facilitate continuous input values have been implemented. Their CFD solver AVL-FIRE is currently extended to incorporate the atomisation SGS model developed in the previous WP. Preliminary simulations for the conjugate heat transfer between the impacting liquid and the heated targets have been performed. These datasets will be used for the training of the ML-tool, which will be also developed in this WP (CITY).
All experiments and simulations performed so far are new in the relevant literature. Specifically: the coolants developed, characterised in terms of rheological, physical and thermal properties and tested experimentally have not been tested before. Spray measurements with such coolants are new and will be reported in the next reporting period. Numerical simulations both MD and sprays for these fluids are also new and will be validated against the obtained measurements. Finally, the e-motor designs (both the linear e-motor for fundamental studies and the full motor which is closer to a real application) have not been tested before with such cooling concept. The first publications will appear in conference in summer 2025.