Adrian Gonzalez Gonzalez

Developing an accurate concentrated solar power (CSP) performance model requires significant effort and time. The power block (PB) is the most complex system, and its modeling is clearly the most complicated and time-demanding part. Nonetheless, PB layouts are quite similar throughout CSP plants, meaning that there are enough historical process data available from commercial plants to use machine learning techniques. These algorithms allowed the development of a very accurate black-box PB model in a very short amount of time. This PB model could be easily integrated as a block into the PM. The machine learning technique selected was SVR (support vector regression). The PB model was trained using a complete year of data from a commercial CSP plant situated in southern Spain. With a very limited set of inputs, the PB model results were very accurate, according to their validation against a new complete year of data. The model not only fit well on an aggregate basis, but also in the transients between operation modes. To validate applicability, the same model methodology is used with a data from a very different CSP Plant, located in the MENA region and with more than double nominal electric power, obtaining an excellent fitting in the validation.

The energy production of concentrated solar power (CSP) plants not only depends on their design, but also of the weather conditions and the way they are operated. A performance model (PM) of a CSP plant is an essential tool to determine production costs, to optimize design and also to supervise the operation of the plant. The challenge is developing a PM that is both easy enough to be useful during the earlier stages of the project, and also useful for supervision of plant operation. This requires one to be able to describe the step between the different modes of operation and to fit the response to transient meteorological phenomena, not so relevant in terms of aggregate values, but crucial for the supervision. The quasi-dynamic performance model (QD-PM) can predict the net energy exported to the grid, as well as all the key operational variables. The QD-PM was implemented using Matlab-Simulink of Mathwoks (MA, USA) with a modular structure. Each module is developed using specific software and a state machine is used to simulate the sequence between the operation modes. The validation of the PM is made using one complete year of commercial operation of a 50 MWe CSP plant situated in Spain. The comparison between the actual data and the results of the model shows an excellent fit, being especially noteworthy as follows the transients between the different CSP operation modes. Then, QD-PM provides an accuracy better than the usual PM, and, almost, as good as that of a fully dynamic model but with a shorter simulation time. But, the main advantage of the QD-PM is that it can be use not only in the feasibility and design stages, but it can be used to supervise the operation of the plant.