4. Modelling Approach
ELL's market simulation model, EMarketNZ as described in section 4.1 is used to model the two scenarios from 2003/2004 through 2007/2008, each scenario with 72 years of historical inflow data relating to the years from 1931/1932 to 2002/2003.
All forecasting requires assumptions to be made, and the first step in the modelling process is to define and develop these assumptions. The base line assumptions are detailed in section 5 and were discussed between the MED and ELL.
4.1 EMarketNZ-Market Simulation Model
EMarket simulates the operation of the New Zealand spot market for electricity, including the generators and their price-quantity offers. Thermal generators are configured to offer bands of energy into the market. For instance, a generator might be configured to offer a base load band of 100MW at 0c/kWh, a firming band of 250MW at 3c/kWh, and a peaking band of 50MW at 5c/kWh.
Hydro generators have negligible real short run marginal costs but are energy constrained i.e. the "fuel" resource available to them is limited by their storage and expected inflows. Their inflows can also be highly volatile. The four major hydro generators, Waitaki, Manapouri, Clutha and Waikato, are modelled in some detail including their component reservoirs and the flows between these reservoirs.
Each hydro generator has a two stage offer, the first being its must run band offered at its must run price, with the second offer based on the "water value." The water value and optimum release for each major system is calculated using stochastic optimisation15 using 72 years of historical weekly inflow data, with reference to the thermal offers in the simulation run, but constrained by a dry year security of supply criteria entered by the user. This is discussed in more detail in section 5.11.
Small hydro generators are also optimised but on the assumption that they are price takers.
At each time zone, all offers are submitted to the dispatch module. For the purposes of long term modelling, time zones of Week Day, Week Night, Other Day and Other Night are used. The module obtains the demand from the demand input data and then dispatches generation to minimise the "cost" of dispatch as set by the generators' offers, and as seen by the purchasers in aggregate, taking account of conservation of energy, grid flows, grid losses and constraints on generator output and line capacities.16
Nodal prices at each modelled node are then calculated as the "shadow price" of each nodal energy constraint. Therefore, they include the price effect of all constraints mentioned above, including any effects due to marginal losses and line constraints.
Back to Top