Numerical Investigation of Pipelines Modeling in Small-Scale Concentrated Solar Combined Heat and Power Plants

• Main Authors: Roberto Tascioni, Luca Cioccolanti, Luca Del Zotto, Emanuele Habib
• Format: Article
• Language: English
• Published: MDPI AG 2020-01-01
• Series: Energies
• Subjects: numerical analysis; renewable energy; dynamic simulation; concentrated solar power plant; organic rankine cycle system; thermal losses
• Online Access: https://www.mdpi.com/1996-1073/13/2/429

In this paper four different detailed models of pipelines are proposed and compared to assess the thermal losses in small-scale concentrated solar combined heat and power plants. Indeed, previous numerical analyses carried out by some of the authors have revealed the high impact of pipelines on the performance of these plants because of their thermal inertia. Hence, in this work the proposed models are firstly compared to each other for varying temperature increase and mass flow rate. Such comparison shows that the one-dimensional (1D) longitudinal model is in good agreement with the results of the more detailed two-dimensional (2D) model at any temperature gradient for heat transfer fluid velocities higher than 0.1 m/s whilst the lumped model agrees only at velocities higher than 1 m/s. Then, the 1D longitudinal model is implemented in a quasi-steady-state Simulink model of an innovative microscale concentrated solar combined heat and power plant and its performances evaluated. Compared to the results obtained using the Simscape library model of the tube, the performances of the plant show appreciable discrepancies during the winter season. Indeed, whenever the longitudinal thermal gradient of the fluid inside the pipeline is high (as at part-load conditions in winter season), the lumped model becomes inaccurate with more than 20% of deviation of the thermal losses and 30% of the organic Rankine cycle (ORC) electric energy output with respect to the 1D longitudinal model. Therefore, the analysis proves that an hybrid model able to switch from a 1D longitudinal model to a zero-dimensional (0D) model with delay based on the fluid flow rate is recommended to obtain results accurate enough whilst limiting the computational efforts.