Abstract
The optimal use of reservoir hydropower is a complex problem. To describe the marginal value of a reservoir connected to a hydropower plant the term water value is used. In an electricity system with both hydropower and thermal capacity, water value is equal to the absolute change of total production costs when there is one unit more water available. Water value is a function of time and reservoir filling level. The less water in the reservoirs, the more valuable it is. Annual water inflow is unequally spread over the year and therefore affects the available amount of water and the water values.
Increasing the share of wind power in electricity generation causes integration costs in the system. These costs are due to e.g. increased need for regulation. Reservoir hydropower is efficient in balancing the system with wind, but the use of reservoirs has to be reoptimised. Wilmar Joint Market Model (JMM) is a stochastic electricity market model created for the study of the effects of large scale wind integration.
In this study a new method is presented to calculate water values for the JMM using VTT’s Markkinahintamalli (MH model). MH model is a tool for predicting electricity production costs in the Nordic countries. Wilmar Long Term Model, which is the original water value component of the JMM, and a method following historical reservoir levels were used as a comparison. The advantage of the MH model is the fast optimisation of water reservoirs using stochastic dynamic programming. It is also possible to calibrate the interplay with the JMM, unlike with the other two methods. Water values are calculated from the total cost function of the MH model. Water value is the derivative of the total cost function with respect to reservoir level.
JMM was used to simulate electricity production in the Nordic countries during the year 2001. Three methods to calculate water values were compared. The most important results are marginal production costs, water reservoir levels and electricity production by generation type. These were compared with the historical data. The results with the new method are promising. The simulated reservoir levels throughout the year are very close to the historical values. Production shares are not accurate with any of the methods, but the production time series appears realistic and stable with the new MH model method.
Additional work will be needed to calibrate and enhance the JMM and the MH model. The most important areas of development are the modelling of different kinds of reservoirs within both the JMM and the MH model, the defining of realistic plant outages as well as the share of runofriver power and the estimation of the variable costs of hydropower production in the MH model. The longterm variation of wind power generation should also be included in the MH model.
Increasing the share of wind power in electricity generation causes integration costs in the system. These costs are due to e.g. increased need for regulation. Reservoir hydropower is efficient in balancing the system with wind, but the use of reservoirs has to be reoptimised. Wilmar Joint Market Model (JMM) is a stochastic electricity market model created for the study of the effects of large scale wind integration.
In this study a new method is presented to calculate water values for the JMM using VTT’s Markkinahintamalli (MH model). MH model is a tool for predicting electricity production costs in the Nordic countries. Wilmar Long Term Model, which is the original water value component of the JMM, and a method following historical reservoir levels were used as a comparison. The advantage of the MH model is the fast optimisation of water reservoirs using stochastic dynamic programming. It is also possible to calibrate the interplay with the JMM, unlike with the other two methods. Water values are calculated from the total cost function of the MH model. Water value is the derivative of the total cost function with respect to reservoir level.
JMM was used to simulate electricity production in the Nordic countries during the year 2001. Three methods to calculate water values were compared. The most important results are marginal production costs, water reservoir levels and electricity production by generation type. These were compared with the historical data. The results with the new method are promising. The simulated reservoir levels throughout the year are very close to the historical values. Production shares are not accurate with any of the methods, but the production time series appears realistic and stable with the new MH model method.
Additional work will be needed to calibrate and enhance the JMM and the MH model. The most important areas of development are the modelling of different kinds of reservoirs within both the JMM and the MH model, the defining of realistic plant outages as well as the share of runofriver power and the estimation of the variable costs of hydropower production in the MH model. The longterm variation of wind power generation should also be included in the MH model.
Translated title of the contribution  Improving water values in a stochastic electricity market model 

Original language  Finnish 
Qualification  Master Degree 
Awarding Institution 

Supervisors/Advisors 

Award date  5 May 2010 
Publisher  
Publication status  Published  2011 
MoE publication type  G2 Master's thesis, polytechnic Master's thesis 
Keywords
 water value
 wind power
 hydropower
 electricity market model
 optimisation
Fingerprint Dive into the research topics of 'Improving water values in a stochastic electricity market model'. Together they form a unique fingerprint.
Cite this
Rinne, E. (2011). Vesiarvolaskennan kehittäminen sähkömarkkinamallissa. Tampere University of Technology. http://urn.fi/URN:NBN:fi:tty201103291059