Hybrid PVT solar systems offer an innovative approach that allows solar energy to be used to simultaneously generate thermal and electrical energy. It is still a challenge to develop an energy-efficient hybrid PVT system. The aim of this work is to develop a mathematical model, investigate the system’s performance based on parameters, include sensitivity analysis in the upper layer mainly photovoltaic part, and provide an efficient and innovative system. Performance analysis of the hybrid system is obtained by establishing a mathematical model and efficiency analysis. The electrical model and thermal model of the hybrid system is also obtained by appropriate and complete mathematical modeling. It establishes a good connection of the system in the context of electrical analysis and power generation. The parameters variation impact and sensitivity analysis of the most important parameters, namely, irradiance, ambient temperature, panel temperature, wind speed, and humidity in the PV panel section, are also obtained using a MATLAB model. The results show the effective increase or decrease in the electrical power and sensitiveness in the output of the system due to this modification. Related MPP values as a result of these parameters variation and their impact on the overall output of the hybrid PVT system are also analyzed.

Hybrid PVT systems simultaneously produce electrical energy using photovoltaic technology and thermal energy using a heat extraction method that collects induced heat from the module. The purpose of this work is to establish a PVT system based on characterization, efficiency study, and performance analysis for both an electrical and a thermal system. A mathematical analysis of the electrical, thermal, and optical model is performed to establish the proposed system. Three types of heat exchanger pipes, including stainless steel, aluminum, and copper, are considered for a heat transfer analysis of the system. The results include temperature profiling, a comparison of the PVT system’s different components, and an overall output and efficiency study for all of the mentioned pipes. Results show that the obtained electrical and thermal efficiency for stainless steel is 0.1653 and 0.237, respectively, for aluminum it is 0.16515 and 0.2401, respectively, and for copper it is 0.16564 and 0.24679, respectively. After comparison, it was found that the overall efficiency for stainless steel is 0.40234, for aluminum is 0.40526, and for copper is 0.41244. Thus, this study will enhance the opportunity to provide an effective hybrid PVT energy management system.

Deep learning models are increasingly used for time-series applications; however, they introduce significant computational challenges, particularly when handling the largest workload datasets. Long Short-Term Memory (LSTM) networks are widely used in time- series applications because they can learn sequential patterns and time-dependent relationships within data. Nevertheless, LSTM training is computationally expensive and time-consuming. This work investigates the performance of LSTM models across heterogeneous computing environments, focusing on the impact of sequential and parallel execution strategies. Experiments were conducted on multiple platforms with different Central Processing Unit (CPU) and Graphics Processing Unit (GPU) hardware using a 13-year (4748 observations) original dataset synthetically expanded to 47,480 observations to evaluate computational workloads rather than additional real-world observations. The results show that task-level parallel execution reduces wall-clock execution time across the tested systems, while predictive performance remains broadly comparable across the tested configurations. Furthermore, the GPU-based implementation achieved lower execution times than the CPU-based implementation in the tested configurations. Overall, this study highlights the importance of HPC and parallel computing for accelerating deep learning workflows in time series applications. The findings confirm that computational performance is highly dependent on hardware capabilities and execution strategies, although accuracy remains stable across heterogeneous platforms.