The accurate prediction of residential heat load is crucial for effective heating system design, energy management, and cost optimization. In order to further improve the prediction accuracy of the model, this study introduced principal component analysis (PCA), the minimum sum of squares of the combined prediction errors (minSSE), genetic algorithm (GA), and firefly algorithm (FA) into back propagation (BP) and ELMAN neural networks, and established three kinds of combined prediction models. The proposed methodologies are evaluated using real-world data collected from residential buildings over a period of one year. The obtained results of the PCA-BP-ELMAN, FA-ELMAN, and GA-BP models are compared with the neural network before optimization. The experimental results show that the combined prediction models have higher prediction accuracy. The Mean Absolute Percentage Error (MAPE) evaluation indices of the three combined models are distributed between 5.95% and 7.05%. The FA-ELMAN model is the combination model with the highest prediction accuracy, and its MAPE is 5.95%, which is 2.25% lower than the MAPE of an individual neural network. This research contributes to the field by providing a comprehensive and effective framework for residential heat load prediction, which can be valuable for building energy management and optimization.

The ultra-fast laser heating process of nano-films is characterized by an ultra-short duration and ultra-small space size, in which the classical Fourier law based on the hypothesis of local equilibrium is no longer applicable. Based on the Cattaneo–Vernotte (CV) model and the dual-phase-lag (DPL) model, the two-dimensional analytical solutions of heat conduction in nano-films under ultra-fast laser are obtained using the integral transformation method. The results show that there is a thermal wave phenomenon inside the film, which becomes increasingly evident as the elapse of the lag time of the temperature gradient. Moreover, the wave amplitude in the vertical direction is much larger than that in the horizontal direction of the nano-film. By comparing the numerical result of the two models, it is found that the temperature distribution inside the nano-film based on the DPL model is gentler than that of the CV model. Additionally, the temperature distribution in the two-dimensional solution is lower than that in the one-dimensional solution under the same Knudsen number. In the comparison results of the CV model, the maximum peak difference in the thermal wave reaches 75.08 K when the Knudsen number is 1.0. This demonstrates that the horizontal energy carried by the laser source significantly impacts the temperature distribution within the film.

Grain is an important material for human survival. However, the expanding world population is contributing to a growing grain shortage. In order to reduce the loss of grain due to pests and mold during storage, mechanical ventilation as the main method of ventilation has crucial research significance. This article proposed and analyzed the ventilation effect and the migration characteristics of ozone in the grain pile under the modified ventilation network (MVN) and compared it with the original ventilation network (OVN). The study found that the temperature, moisture, and ozone concentration in the grain pile of the two ventilation networks are not evenly distributed in the vertical direction, showing a layered pattern. That is, with an increase in grain stack height, the temperature and moisture content of the grain stack are higher, and the ozone concentration is lower. Moreover, in the pre-ventilation period, the average temperature decline rate of the MVN was 1.25 °C/d, which was better than that of the OVN (0.84 °C/d), and the maximum temperature difference between the MVN and the OVN was 0.89 °C. The vertical ducts added to the MVN improved the ventilation effect, maintaining high ozone concentrations within the grain pile. Notably, on the sixth day of fumigation, the average ozone concentration of the MVN exceeded that of the OVN. The MVN can solve the shortcomings of the OVN, where air intake and fumigation gas rise slowly in the vertical direction. These findings hold substantial significance for optimizing ventilation network structures, devising effective fumigation strategies, and enhancing the insecticidal effects of grain storage.

The Discrete Element Method (DEM) has been widely employed to investigate the behavior of particle systems at a macroscopic scale. However, effectively simulating the gradual filling of bulk cereal grains within silos using the DEM remains a formidable challenge due to time constraints. Thus, there is a critical need to develop a simplified model capable of substantially reducing the computational time required for simulating cereal grain accumulation. This study introduces a Layered Filling Method (LFM) designed to expedite the computational process for cereal grain piles within silos. By utilizing particle kinetic energy as a specific criterion, this model identifies particles as stable situations when their kinetic energy drops below a designated threshold. Throughout the filling process, lower particles that were judged to satisfy the condition of stability are isolated, forming sub-heaps that are exempt from persistent detection. The whole particle heap is subsequently segregated into multiple sub-piles and a main pile till the process’s culmination, and these divisions are merged back together. In order to validate the model’s feasibility and accuracy, a comparative analysis was performed on the characteristics of the porosity and airflow patterns of grain piles generated using the LFM and the progressive filling method (PFM), respectively. The research results indicate that there is a marginally higher porosity value in the grain pile simulated by the LFM in comparison to the PFM. However, the average relative error remains below 5.00%. Both the LFM and PFM exhibit a similar spiral upward trend in the simulated airflow paths. Notably, the LFM demonstrates a substantial reduction in the time required to construct grain piles.

The gray model simplifies calculations by ignoring phonon polarization, but sacrifices a certain level of computational accuracy. In effect, the frequency and wavevector of phonons form complex polarization patterns, which means their propagation modes and vibrational directions have different influences. Therefore, based on the phonon dispersion relations in silicon, the lattice Boltzmann method is used to analyze the phonon transport characteristics in nano-silicon films under ultrafast laser excitation. The results show that the total energy density distribution obtained by superimposing acoustic and optical branches exhibits multiple wave-like behaviors. Among them, the acoustic branch has excellent transfer capability, dominating the rate at which the total energy density reaches a steady state distribution, while the optical branch has stronger heat capacity characteristics, with a greater impact on the peak value of the total energy density. When the heat transfer approaches a steady state, the longitudinal optical branch surprisingly contributes up to 52.73%. This indicates that the often-neglected optical phonons should also receive sufficient attention. Additionally, compared to the results of the gray model, it is found that the dispersion model is preferred when more attention is paid to the propagation characteristics during phonon transport.

The post-combustion carbon capture process with monoethanolamine/piperazine (MEA/PZ) blends encounters notable modeling and optimization challenges. These arise from strong thermodynamic–kinetic nonlinear coupling, as well as limited availability of high-quality experimental data. To address this, we propose a machine learning and multi-objective optimization strategy driven by physicochemical features. By extracting explicit physical features and embedding physicochemical constraints into data-driven models, this study evaluated the predictive performance of three distinct algorithms based on wet-wall column experimental data. These algorithms included natural gradient boosting (NGBoost), sure independence screening and sparsifying operator (SISSO), and gaussian process regression (GPR). Subsequently, an optimization problem aimed at minimizing $P_{{\mathrm{CO}}_2}^{*}$ and maximizing $k_{\mathrm{g}}^{\prime }$ was formulated. The multi-objective beluga whale optimization (MOBWO) algorithm was then employed for global optimization and benchmarked against the traditional non-dominated sorting genetic algorithm II (NSGA-II). Results indicate that the Gaussian process regression (GPR) model performed best when it was enhanced by physicochemical features and optimized via Bayesian hyperparameter tuning. It achieved R² values of 0.989 and 0.953 for $P_{{\mathrm{CO}}_2}^{*}$ and $k_{\mathrm{g}}^{\prime }$, with average absolute relative deviations (AARDs) kept below 15.7% and 12.2% respectively. Feature importance analysis validated the underlying physical laws. Specifically, temperature dictates thermodynamic equilibrium, while CO₂ loading limits mass transfer kinetics. In the optimization phase, MOBWO outperformed NSGA-II by generating a more uniformly distributed Pareto front. Decision-making analysis further identified three typical operating regimes encompassing kinetics-dominant, thermodynamics-dominant, and comprehensive equilibrium conditions. This framework provides a robust paradigm for small-sample modeling and optimization in complex chemical processes.