A Power Market Forward Curve with Hydrology Dependence - An Approach based on Artificial Neural Networks
(2014)- Abstract
- This paper develops an hourly forward curve for power markets where the intra-day and intra-week shapes (profiles) depend on the level of the hydrological balance. The shaping model is based on a feed-forward Artificial Neural Network (ANN), which is trained on a historical data set of hourly electricity spot prices from the Nord Pool market and weekly
measurements of the Nordic hydrological balance. The yearly seasonal cycle is estimated with historical electricity forward prices from the Nasdaq OMX Commodities exchange. We calibrate the shaping model to prevailing electricity forward prices and proceed to demonstrate its most important properties. By using comparative static analysis we particulary focus on the hydro dependence... (More) - This paper develops an hourly forward curve for power markets where the intra-day and intra-week shapes (profiles) depend on the level of the hydrological balance. The shaping model is based on a feed-forward Artificial Neural Network (ANN), which is trained on a historical data set of hourly electricity spot prices from the Nord Pool market and weekly
measurements of the Nordic hydrological balance. The yearly seasonal cycle is estimated with historical electricity forward prices from the Nasdaq OMX Commodities exchange. We calibrate the shaping model to prevailing electricity forward prices and proceed to demonstrate its most important properties. By using comparative static analysis we particulary focus on the hydro dependence of the shapes. We conclude the paper with a real world
valuation task. By combining our proposed forward curve with a simple Ornstein-Uhlenbeck process we price a strip of hourly call options on the electricity spot price under different hydrological scenarios. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/4857886
- author
- Green, Rikard LU
- organization
- publishing date
- 2014
- type
- Working paper/Preprint
- publication status
- unpublished
- subject
- keywords
- artificial neural networks, Power markets, seasonality, forward curve
- pages
- 29 pages
- language
- English
- LU publication?
- yes
- id
- d2094c10-f8c3-467f-b7d4-14528c6a9b4d (old id 4857886)
- date added to LUP
- 2016-04-04 13:07:47
- date last changed
- 2018-11-21 21:12:20
@misc{d2094c10-f8c3-467f-b7d4-14528c6a9b4d, abstract = {{This paper develops an hourly forward curve for power markets where the intra-day and intra-week shapes (profiles) depend on the level of the hydrological balance. The shaping model is based on a feed-forward Artificial Neural Network (ANN), which is trained on a historical data set of hourly electricity spot prices from the Nord Pool market and weekly<br/><br> measurements of the Nordic hydrological balance. The yearly seasonal cycle is estimated with historical electricity forward prices from the Nasdaq OMX Commodities exchange. We calibrate the shaping model to prevailing electricity forward prices and proceed to demonstrate its most important properties. By using comparative static analysis we particulary focus on the hydro dependence of the shapes. We conclude the paper with a real world<br/><br> valuation task. By combining our proposed forward curve with a simple Ornstein-Uhlenbeck process we price a strip of hourly call options on the electricity spot price under different hydrological scenarios.}}, author = {{Green, Rikard}}, keywords = {{artificial neural networks; Power markets; seasonality; forward curve}}, language = {{eng}}, note = {{Working Paper}}, title = {{A Power Market Forward Curve with Hydrology Dependence - An Approach based on Artificial Neural Networks}}, url = {{https://lup.lub.lu.se/search/files/6057400/5046239.pdf}}, year = {{2014}}, }