Search

This paper describes a thermodynamic model library developed with the object-oriented language Modelica, which is both implemented for steady-state and transient heat transfer analyses of Parabolic Trough Receivers (PTRs) installed in Concentrated Solar Power (CSP) plants. For the identification of PTR heat loss mechanisms, this heat transfer model is coupled to a derivative-free hybrid optimization routine developed in Matlab, combining a Particle Swarm Optimization (PSO) algorithm with a Nelder-Mead Simplex (NMS) algorithm with search space boundary constraints.

The operation of solar thermal high temperature receivers requires an accurate knowledge of the temperature distribution on critical parts of the receiver. However, concentrated solar radiation makes it difficult to determine the temperature on irradiated surfaces. Contact thermometry is not appropriate for the use under concentrated solar radiation and also pyrometry fails when external light sources interfere significantly. To avoid distortion of the temperature reading, the measurement has to be performed in a spectral range where the emitted thermal radiation exceeds the reflected solar radiation by a multiple. The measurement in solar blind spectral regions of atmospheric absorption bands offers one possible solution of filtering the solar radiance from the measurement signal. Along with the intensity of the incoming solar radiation and the absorber emittance, the bidirectional reflection properties and the temperature of the object determine the required selectivity of the spectral filter. In atmospheric absorption bands, the influence of the atmospheric absorption on the measurement signal cannot be neglected even for small path length. The paper describes the methods of solar blind pyrometric temperature measurement on solar thermal high temperature receivers and shows the possibilities and limitations of accounting for the atmospheric absorption with models based on radiation transfer calculations. Finally, experimental results recorded with an infrared mirror scanner especially designed for the measurement on a pressurized volumetric receiver are presented and compared to thermocouple readings and results of a thermodynamic model of the receiver.

A clamp-on measurement system for flexible and accurate fluid temperature measurements for turbulent flows with Reynolds numbers higher than 30,000 is presented in this paper. This noninvasive system can be deployed without interference with the fluid flow while delivering the high accuracies necessary for performance and acceptance testing for power plants in terms of measurement accuracy and position. The system is experimentally validated in the fluid flow of a solar thermal parabolic trough collector test bench, equipped with built-in sensors as reference. Its applicability under industrial conditions is demonstrated at the 50 MWel AndaSol-3 parabolic trough solar power plant in Spain. A function based on large experimental data correcting the temperature gradient between the measured clamp-on sensor and actual fluid temperature is developed, achieving an uncertainty below ±0.7 K (2σ) for fluid temperatures up to 400 °C. In addition, the experimental results are used to validate a numerical model. Based on the results of this model, a general dimensionless correction function for a wider range of application scenarios is derived. The clamp-on system, together with the dimensionless correction function, supports numerous combinations of fluids, pipe materials, insulations, geometries, and operation conditions and should be useful in a variety of industrial applications of the power and chemical industry where temporal noninvasive fluid temperature measurement is needed with good accuracy. The comparison of the general dimensionless correction function with measurement data indicates a measurement uncertainty below 1 K (2σ).