SmartCSP – Application of AI methods to improve operations and maintenance in CSP power plants.

Duration: 10/2021 - 03/2024
Contracting Authority/ Sponsors: Bundesministerium für Wirtschaft und Klimaschutz (BMWK)
Project Partners: Mondas GmbH, RWTH Aachen
Project Focus:  
Im Projekt betrachtete CSP Kraftwerkskomponenten
© Fraunhofer ISE
Im Projekt betrachtete CSP Kraftwerkskomponenten. Vorne das Parabolrinnenfeld – die Wärmequelle und in der Mitte das Trockenkühlsystem – die Wärmesenke.
Predictive control/improvement of O&M strategies using insights from machine learning
© Fraunhofer ISE
Predictive control/improvement of O&M strategies using insights from machine learning.

The goal of the SmartCSP project is to make available and usable the large amounts of data generated so far in CSP power plants, not only in the control at the component level, but also at the power plant level for the operating personnel. The data use takes place via online monitoring for one or more CSP power plants currently in operation for the solar field -- the "hot end" or heat source -- and the turbine cooling system -- the "cold end" or heat sink. The data collected on the solar field and cooling system will be used for direct performance monitoring, but it will also be evaluated for further operational optimization using machine learning algorithms.

The SmartCSP project combines applied research on AI methods for operation optimization with a practical approach of cost-effective system integration in new and already in operation plants. On the one hand, this is intended to optimize the operation and maintenance (O&M) strategy of the plant components at the power plant level, and on the other hand, it should be possible to draw conclusions in order to further develop the components themselves. Approaches and techniques that have already been developed in other industries through the digitalization of systems in relation to the development of Industry 4.0 are to be used consistently. For example, additional data required to achieve the project goals will not be collected using complex and expensive measurement technology, but rather by using a large number of inexpensive sensors and networking them within the online monitoring system.

The concrete work objectives of the project are:

Online monitoring system of the CSP power plant

- Establishment/use of a cloud-based or SQL server-based platform for the evaluation of the existing sensor-generated time series data acquired during the operation of CSP power plants, in order to be able to use them in global power plant control and operation optimization, instead of only for local component control as has been the case to date.

- Combination, development and testing of the use of currently available low-cost sensor technology in the context of solar thermal power plants for the cost-optimized generation of additionally required information.

Optimization of solar field operation

- Better understanding of solar field status based on existing but unused operational data and enhanced sensor technology.

- AI-based generation of recommendations and clear instructions to improve collector operation.

Optimization of cooling system design and operation.

- Set-up of additional measurements for air velocity, temperature, and pressure for online data collection on the cooling system of a CSP power plant and their analysis.

- Creation and validation of a detailed CFD model of the cooling system considering the environment and the real fan geometry, 

Holistic system simulation of the CSP power plant - Digital Twin

- Set-up of a power plant system simulation which is fast enough for annual simulations - including validation by real measurement data

- Investigation of potential measures to increase effectiveness / improve plant operation

More Information on this Topic:

Research Topic

Concentrating Solar Collectors

Research Topic

Solar Thermal Power Plants and Industrial Processes