| With the increase of populations and improvement of living standards,the accelerated urbanization process and increasing infrastructures give rise to more energy consumption in buildings,which is mainly attributed to heating,ventilation and lighting.Two-thirds of the energy consumption in buildings is used to air conditioning in order to maintain the indoor temperature within the range of the thermal comfort of humans.In order to encounter with the energy crises and challenges,many countries announced the carbon neutral plans,and China also put forward two goals of carbon peaking and carbon neutrality.Based on the above problems,the subjects were studied in current study:The thermo-physical properties of low eutectic hydrated salts（EHS）synthesized from sodium carbonate decahydrate（Na2CO3·10H2O,SCD）and disodium hydrogen phosphate dodecahydrate（Na2HPO4·12H2O,DHPD）were studied,and the supercooling and anti-leakage were improved;Then the EHS were combined with hollow bricks of building materials to form phase change walls by macro-encapsulation method;The heat storage and retrieval processes of phase change walls were studied experimentally and numerically,and the phase change walls were extended to building level by systematic simulation;Finally,the temperature control effects were predicted by machine learning,which provided the guideline for experiments.SCD and DHPD were chosen to synthesize binary low EHS based on the low eutectic principle,different mass fractions of which were mixed,and it was found that when the mass ratio of SCD:DHPD was 4∶6,the EHS had good co-melting characteristics and no phase separation was appeared.The corresponding phase change temperature was 26.05℃,which was in the range of 22-28℃for thermal comfort of humans,and the latent heat was 197.31 k J/kg.The appropriate phase change temperature and large latent heat met the requirements of building energy storage.In order to solve the supercooling of hydrated salts,several nanoparticles were selected as nucleating agents to make the improvements.It was found that when the mass fraction of added BN was 5%,the material of EHS/5 wt.%BN had the best improvement effects in lowering subcooling,which was reduced from 10.72℃to 5.47℃.Solid-liquid phase change process is accompanied by leakage,which limits the application of phase change materials（PCMs）.Expanded graphite（EG）with porous structure was used to avoid the leakage,and it was found that the leakage problem of composite PCM was solved when the mass fraction of EG was 7.5%.The composite PCM（EHS/BN/EG）was subjected to morphological characterization,FTIR and thermal conductivity tests.The structural characteristics showed that the hydrated salts composite PCM was completely filled into the pores of EG with the addition of 7.5 wt.%EG.The FTIR tests indicated that the composite PCM was physically mixed and no new functional groups were formed among the internal components.The thermal conductivities of all the composite PCMs were tested with the steady-state method.The thermal conductivity of the composite PCM with the addition of 7.5 wt.%EG was 2.360 W/（m·K）,which was 330%larger than that of pure EHS.With the addition of 10 wt.%EG,the improvement in thermal conductivity was the most significant,since the thermal conductivity reached 4.087 W/（m·K）,which was 472%larger than that of EHS.EHS/5 wt.%BN and EHS/5 wt.%BN/7.5 wt.%EG were filled into hollow brick holes by macro-encapsulation method to prepare phase change walls,which were heated by infrared radiator used for simulating solar radiation.Simulataneously,pure brick walls were built as the control group.Three types of walls were tested for heat storing/releasing performance to study the temperature control effects on indoor side temperature.Under different heating temperatures and distances,the phase change walls were found to keep indoor temperature within a comfortable thermal range with more time,which was longer than that of pure brick walls,and the temperature rise on the indoor side was slower.At the heating temperature of 50℃and heating distance of 30cm,phase change walls had the best temperature control effects.Pure brick walls could only maintain the indoor thermal comfort for 5690 s while phase change walls filled with EHS/5 wt.%BN and EHS/5 wt.%BN/7.5 wt.%EG could maintain the indoor thermal comfort for 13210 s and7740 s respectively.The two types of walls extended the time-durations of thermal comfort by132%and 53%,respectively.The DSC and step-cooling tests were repeated after the system walls had been in place for nearly one year,proving the long service life.With the software COMSOL Multiphysics 6.0,a 1:1 model in size was built to obtain the internal temperature changes of the walls more intuitively.The numerical results were compared with the experimental data,and the temperature measurement points on the interior side of the walls agreed well with the experimental results during the stage of heat storage.A typical building composed of phase change walls was simulated in Shanghai as an example.The ordinary walls can only maintain the time-duration of indoor thermal comfort for 4.25 h while the phase change walls filled with EHS/5 wt.%BN can maintain the time-duration of indoor thermal comfort to 6.56 h and the phase change walls filled with EHS/5 wt.%BN/7.5 wt.%EG can maintain the time-duration of indoor thermal comfort to7.50 h,extended by 54%and 76%respectively.Beijing was also selected for a case study.The time-duration of indoor thermal comfort could be maintained to 7.61 h,which proved that phase change enclosure was also applicable in other areas.Machine learning of temperature control effects of phase change walls was carried out by using BP neural network.48 sets of working condition data from the experiments and simulations were used as the training set,and the heating surface and ambient temperature were used as input variables,and the time-durations of indoor thermal comfort was output variable.Finally,the learning deviation between original data and the results obtained from machine learning was within 5%.And through the machine learning model,the temperature control effects of phase change walls under unknown working conditions were predicted,including three sets of environmental temperature data of 18℃and 24℃.The machine learning results were compared with the simulation results,and the deviation could be controlled within 3.5%,indicating that machine learning can accurately predict the temperature control effects of phase change walls. |
