Maximising dispatchability of combined solar power plants

Abstract

A challenge of solar energy production is the intermittency of solar availability. Photovoltaic (PV) arrays can be supplemented with battery storage to maintain steady output even during solar transient events. Concentrating solar power plants (CSP) solve this problem with thermal energy storage (TES), allowing production to be inexpensively shifted away from periods of solar availability. Hybrid PV-CSP plants combine the benefits of both methods, with the PV plant being used to produce electricity when the solar resource is available, and the CSP plant with large TES being used to provide the remaining power to track loads. However, during cloud transients, the power output of PV arrays dramatically changes, much faster than the dynamics of a CSP plant. This paper looks at the use of short term (up to 10 minutes) predictions of solar availability to anticipate these events and enable the plant to better track its required load.In order to do this, models of both the CSP and PV components of the hybrid plant are developed with suitable time dynamics. A model predictive control method is used to control the power output of the CSP plant and simulate the performance of the combined plant for 1-1.5 hour periods of variable solar availability. Three different methods of solar availability prediction are used: no predictions, perfect knowledge, and predictions of direct normal irradiance (DNI) from Pidgeon (2014).For the periods simulated, the use of perfect predictions reduced the RMSE of plant power output with respect to a reference power output by 35-50%. By comparison, the causal predictions from Pidgeon (2014) only gave a 5-10% improvement. Both of these methods also reduced the size of battery storage required to supply extra power to meet the required load, which can contribute to lowering the costs of such a hybrid plant. This shows that better predictions of solar availability forecasts can improve the load tracking ability of a PV-CSP plant. This work also gives some insight into improvements that can be made to solar availability predictions to improve the performance for this kind of application.

UQ eSpace

“…As such, solar energy is expected to play an increasingly important role in the Australian energy sector, and there are clearly opportunities to transition from fossil fuels towards incorporating large scale solar PV and solar thermal systems in remote mining sites and communities. Furthermore, hybridising variable renewable energy such as PV, with solar thermal and/or diesel generators has been recommended to increase reliability and penetration (Choros, 2016). Hybrid systems integrating renewables with existing diesel generator systems are gaining traction internationally as they provide a suitable route to decrease reliance on fossil fuels and attain lower energy costs (Adaramola et al, 2017).…”