Presentation Details
| Yield forecasting conundrums Keith McIntosh. PV Lighthouse |
Abstract
The utility-scale PV industry contends with large uncertainties in its forecasts and experiments. The energy yield from a PV plant is typically predicted with an uncertainty of �4�8%, and forecast models are not easily evaluated because measurements at test facilities have errors of, say, �2�5%. And yet, for large power plants, even small gains in energy yield�like +0.5%�can translate into significant financial returns.
These large uncertainties, combined with the importance of marginal gains, introduce a variety of conundrums. For example, how can incremental technical advances be demonstrated in the field? How should forecasting models be compared when their differences fall within experimental error? What constitutes sufficient evidence to validate a new model? How should uncertainties inform the acceptance criteria during commissioning?
This talk explores these and other yield forecasting conundrums through practical examples, providing insight into how they can be addressed by experimentation and modelling.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.