The article “Technical and economic effectiveness of large-scale compression heat pumps and electric boilers in energy systems with high penetration levels of wind power and CHP”
submitted for publication to Energy in April 2007 [10], pre-sents the modelling framework COMPOSE for assessing relocation options that combines operational design model energyPRO [18], historical production data from Energinet.dk, and various projections from the Danish Energy Agency’s energy system model RAMSES [23], optionally system data from EnergyPLAN.
COMPOSE, which is an acronym for Compare Options for Sustainable Energy, allows for the evaluation of user-defined energy projects in user-defined systems. The mission is for COMPOSE to combine the strength of energy project opera-tional simulation models with the strength of energy system scenario models in order to arrive at a modelling framework that supports an increasingly realistic and qualified compara-tive assessment of sustainable energy options.
COMPOSE currently allows for the evaluation of a project’s relocation coefficient, economic cost-effectiveness of reloca-tion, economic costs, as well as both local, avoided, and system-wide CO2 emissions and consumption of primary energy resources.
Figure 3 illustrates the overall model flow chart for COMPOSE.
In essence, COMPOSE imports an optimized operational strategy from energyPRO, and combines the resulting hourly
33
energy balance for a given year with projected annual and hourly characteristics of the system in which the project is located.
Figure 4 illustrates the internal structure of COMPOSE. At the core of the model, the user defines a system and a project.
The system consists of five parent components: energy system, economic system, environment system, risk specifica-tions and methodology opspecifica-tions. The project consists of two major child components: process and demand.
One key methodological feature is the way most variables are associated with both an annual projection that describes how the mean annual value will develop over the planning period, as well as an hourly profile that describes how the mean annual value is distributed into hourly values for each year in the planning period. COMPOSE imports annual and hourly profiles from Energinet.dk and RAMSES, and optionally also from EnergyPLAN, while projects, including the optimized hourly production profile for each production unit, are im-ported from energyPRO. It is furthermore possible to localize hourly profiles using monthly climate data from RetSCREEN [24], for example cloning a recorded Danish hourly production profile for solar cell production into a simulated hourly produc-tion profile for Trieste in Italy.
Another key methodological feature is the way the interface between the energy project and the energy system is mod-elled. COMPOSE applies a least-cost dispatch model for central electricity generation that relies on how user-selected candi-date marginal electricity producers are expected to bid and stay in the electricity market (the spot market) according to each producer’s long-term marginal production costs. Rather than relying on short-term marginal costs for identifying operational changes in central electricity generation, COMPOSE suggests a simplified methodology by which the dispatch analysis reflects the long-term consequences of changes.
To support an integrated and consistent evaluation, COMPOSE handles each of the candidate marginal plants in the dispatch model as any other COMPOSE energy project, thereby subject to similar assumptions and algorithms.
34
Figure 3: COMPOSE: Model flowchart.
System Economy
System Energy
System Marginal Dispatch
Environment System
Methodology Risk Variables
Project
Demand Process
Cost Benefit Fuel
Figure 4: COMPOSE: Structure for defining energy project and system.
35
Figure 5 illustrates the structure of the energy system compo-nent by which the marginal dispatch model refers to any number of user-defined projects. Also the energy system holds an hourly projection for electricity demand and intermittent supply, which is of particular importance in arriving at a project’s relocation coefficient.
The analyses in [10] are subject to economic fuel costs and electricity markets projected by the Danish Energy Agency / RAMSES as of February 2007 [25]. While avoided costs in central electricity generation are given by the projected annual and hourly prices in the electricity market - projected annual means according to RAMSES and hourly fluctuations assumed similar to historic markets according to Energinet.dk market data, in [10] the 2006 spot market, optionally according to EnergyPLAN - COMPOSE’s least-cost dispatch model is applied for identifying marginal producers in central electricity produc-tion with respect to avoided primary energy consumpproduc-tion and emissions. Figure 6 illustrates the consequences of this dispatch model for the analyses in [10], where three candidate marginal producers towards 2025 were defined for the West Danish energy system: coal-fired power plants at 48 % fuel-to-electricity efficiency, natural gas fired power plants at 55 % fuel-to-electricity efficiency, and off-shore wind power. The dispatch model settled that for electricity market prices below
€33,3 per MWh, found to be the long-term marginal produc-tion costs of coal-fired power producproduc-tion on given assump-tions, wind power is the marginal producer. Between €33,3 and €44,7 per MWh, coal-fired power plants is the marginal producer, while natural gas fired power plants is the marginal producer spot market prices above €44,7 per MWh, corre-sponding to the long-term marginal production costs of natural-fired power production. Detailed techno-economic assumptions with respect to projected fossil fuel prices, O&M costs, and the electricity spot market are presented in [10] as well as in the database included with COMPOSE.
Finally, COMPOSE allows for the user to specify uncertainty ranges for a number of selected variables including heat demand, electricity demand, intermittent supply, and eco-nomic discount rate. These uncertainties are then applied in extensive Monte Carlo risk assessments, subsequently allow-ing for the computation of statistical means and frequencies for individual results.
36
COMPOSE is included on the CD-ROM that accompanies this thesis, and may also be downloaded from http://energyinteractive.net. COMPOSE is a client-server application with a remote database server and includes features to support interactivity, therefore Internet access is required while using the model.
Hourly Profile Energy
System
Intermittent Production
Annual Profile Electricity
Demand Marginal
Dispatch
Candidate Dispatch Project Candidate Dispatch Project ...
Figure 5: COMPOSE: The energy system component.
0 20 40 60 80 100 120 140 160 180 200
1 8760
Hours
€ per MWh Gas-fired PP is marginal
Coal-fired PP is marginal Wind is marginal Electricity spot market
Figure 6: Marginal dispatch in central electricity generation as a function of 2006 spot market price fluctuations.
37