The photovoltaic-thermoelectric hybrid power generation system is a promising solar energy technology. However, traditional series-connected photovoltaic-thermoelectric system faces challenges such as mismatched device operating temperature and high thermal resistance. In this study, a bifacial type photovoltaic-thermoelectric hybrid power generation system with a sandwichlike configuration is proposed. An experimental setup is constructed to investigate the effects of different irradiance levels and cooling water flow rates on the performance of the new system under steady-state indoor conditions. Experimental data shows that the output power of photovoltaic module and thermoelectric device in bifacial type system is superior to that of traditional series system. Moreover, increasing irradiance levels can enhance the system’s power generation capacity but may reduce the photovoltaic power generation efficiency. Additionally, increasing the cooling water flow rate can further enhance the system’s output performance.

Carnot battery is an emerging energy storage technology characterized by low cost and independence from geographical constraints. The use of latent heat thermal energy storage in Carnot battery results in high energy density, offering broad application prospects. This paper constructs a Carnot battery system based on a packed bed latent heat storage and establishes its dynamic model. The analysis includes the impact of heat transfer fluid flow rate and organic working fluid mass flow rate on system performance. A functional relationship for the mass flow rates of the heat transfer fluid and organic working fluid has been established at a design power of 1000 kW. Subsequently, using the NSGA-II algorithm for multi-objective optimization with roundtrip efficiency and energy density as optimization objectives. After balancing using the LINMAP method, the roundtrip efficiency and energy density obtained were 62.74% and 12.96 kWh/m³.

The initial frosting phenomenon is a spatially discontinuous phase transformation and nucleation process. The cold surface temperature and characteristics have a direct and significant impact on the frosting phenomenon. It has been shown that the effect of a superhydrophobic surface on frost formation on a cryogenic cold surface is much less obvious than that on a general cold surface. To understand the influences of surface characteristics on frost formation at −190∼ −30°C, the growth and morphological characteristics of frost crystals on hydrophilic and superhydrophobic surfaces at the initial stage of frost formation are experimentally studied. Four frost formation patterns are observed: cold surface condensation frosting, air boundary layer condensation frosting, cold surface sublimation frosting, and air boundary layer sublimation frosting. The four frost formation modes appear independently or co-exist on cold surfaces of different temperatures, and have important effects on the morphology of frost crystals. The cold surface properties (contact angles) have a significant effect on frost formation patterns within a specific range. The cold surface condensation frosting mode is only observed on hydrophilic surfaces at −30°C. The liquid air is only observed on superhydrophobic surfaces at −190°C. The initial frost crystal morphology changes with the cold surface temperature, frost formation mode, and surface properties, and it can be roughly divided into four forms: hexagonal, pine needle, cluster, and flocculant.

Based on the basic two-stage and dual-temperature ejector refrigeration cycle (BTERC), a two-stage and dual-temperature ejector auto-cascade refrigeration cycle (TEARC) driven by two heat sources is proposed by introducing a gas-liquid separator and an evaporative condenser. The thermodynamic performance of the BTERC and TEARC is investigated by adopting a onedimensional constant-pressure mixing model of the ejector and thermodynamic modeling of the cycle. The results show that the specific enthalpy of refrigerant at the low-temperature (LT) evaporator and medium-temperature (MT) evaporator inlet is decreased via the evaporative condenser and separator of the TEARC, respectively, leading to a higher COP and cooling capacity. Under the basic operating condition, the TEARC shows improvements of 9.06%, 24.11% and 7.51% in COP, cooling capacity of the LT evaporator, and cooling capacity of the MT evaporator than those of the BTERC, respectively.

Wind power exhibits significant fluctuations, and accurate wind power prediction can provide a guarantee for operation and maintenance scheduling. Common time series models are only suitable for ultra-short-term forecasting. By modeling the power conversion characteristics of wind turbines using neural networks, it is possible to perform day-ahead predictions based on wind speed forecasts. This data-driven approach can improve the deviations caused by terrain, wake effects, and other factors in theoretical power calculation methods. Neural networks provide a more accurate description of the actual operating characteristics of wind turbines. Furthermore, a Gaussian Mixture Density Network (GMDN) is established to assess power uncertainty. The model automatically learns distribution parameters through training, resulting in a non-prior multivariate distribution that better reflects the inherent structure of the data.