Session: 08-01: Deployment and Analysis of CSP Subsystems

Paper Number: 168351

168351 - Estimating Value of Information for Heliostat-Washing Operations at Concentrating Solar Power Plants With Monte Carlo Simulation 

Abstract: 

Concentrating solar power central-receiver plants use thousands of sun-tracking mirrors, i.e., heliostats, to reflect sunlight to a central receiver, which collects and uses the heat to generate electricity. Over time, soiling reduces the reflectivity of the heliostats and, therefore, the efficiency of the system. Current industry practice sends vehicles to wash heliostats using schedules that are either predetermined or updated in response to reflectance data collected on site at regular intervals, which incurs cost due to the labor required to obtain measurements. This work presents a novel framework for evaluating heliostat-washing strategies at concentrating solar power (CSP) plants under uncertain soiling conditions, with a focus on determining how frequently soiling data should be collected and how measurement inaccuracies influence operational decisions. Leveraging a discrete event simulation (DES), our approach captures the stochastic nature of precipitation events, dust storms, and daily soiling accumulation, as well as measurement error. The integrated model combines SolarPILOT, used to assess heliostat field performance, with the System Advisor Model to obtain system-level energy output.

Unlike prior studies that rely on predetermined or purely threshold-based cleaning schedules without considering measurement uncertainty, this work explicitly models the discrepancy between true and measured reflectance. In doing so, we reveal the impact of both measurement frequency and measurement accuracy on the performance of the plant when using schedule-driven and measurement-driven policies. This allows us to estimate the value of information as a function of instrument accuracy and the frequency of measurements taken, allowing the operator to optimize the frequency of recalibrating the instrument and taking measurements in the field.

The approach is demonstrated through a collection of contrived case studies of operating plants with different wash technologies available. The results indicate that estimating the value of information through this method can be a useful method to determine data collection frequency and assess potential investment in measurement instrumentation.

Presenting Author: Justin Kilb Colorado School of Mines

Presenting Author Biography: Justin Kilb is a Ph.D. candidate in Operations Research with Engineering at the Colorado School of Mines. His research explores generative methods and optimization, with applications in energy systems and defense technologies.