Pilot Plant Analysis, Experiments, and Control for the Hybridization of Transient Solar Heat with Conventional Utilities

Rowe, Scott Christian

Graduate Thesis Or Dissertation

Pilot Plant Analysis, Experiments, and Control for the Hybridization of Transient Solar Heat with Conventional Utilities Public Deposited

Analytics

Citations

Citeable URL: https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/q524jp00q

Abstract

The direct capture of solar heat is now commercial for electrical generation at 550 ℃ (1000 ℉), which has provoked interest in solar driven approaches to commodity and fuels production at higher temperatures. However, conventional commodity and fuels facilities often operate continuously regardless of weather and nighttime conditions. Conversely, direct sunlight is immediately lost upon shading by clouds and sunset. Beyond inconvenience, this intermittency has the potential to destroy high temperature equipment through thermal fatigue and thermal shock. To overcome interruptions in solar availability we propose the inclusion of direct sunlight in commodities and fuels production as a supplement to conventional electrical heating. Within this regime conventional utilities are ideally sourced from sustainable stored or orthogonal energy sources. Control is needed to substitute solar, which can be lost within seconds during transient weather, with electrical heat. To explore control strategies for the alternation of solar and electrical heat a new facility was constructed at the University of Colorado, Boulder. Specifically, a 45 kW 18 lamp high-flux solar simulator was erected that approximates the sunlight found in actual concentrated solar plants. Calorimetry was analyzed for the measurement of extreme radiance in this testbed. Results from calorimeter design were applied to radiation measurement from the lamps, which were capable of delivering 9.076±0.190 kW of power to a ∅10 cm target with a peak flux of 12.50 MW/m² (12,500 “suns”). During this characterization a previously unknown observer effect was seen that differentiates radiative heat from lamps and the energy delivered by sunlight in actual concentrated solar facilities. This characterization allowed confident experimentation within the lamp testbed for control studies on a 15 kW solar-electric tube furnace for commodities and fuels production. Furnace electric heat was manipulated by four different linear control strategies for the rejection weather transients reproduced by the high-flux solar simulator lamps. These included feedback, feedforward feedback, model predictive control, and model predictive control with a weather forecast. It was found that model predictive control with a forecast best maintained furnace conditions. Prior researchers have suggested that forecasts would be useful in solar control, which was shown across simulation and experiment.

Creator