Overview
openCEM is a capacity expansion and dispatch model that simulates the national electricity market (NEM) under a set of technical, cost and policy assumptions. Based on those assumptions, openCEM computes future capacity expansion (i.e. building large-scale generators and storage systems) and dispatch decisions over a number of years into the future that achieve a system-wide lowest annualised cost of operation.
openCEM divides the NEM into 16 planning zones to account for differences in renewable energy resources, fuel costs, electricity demand and connection costs. Each zone contains its own list of generator and storage capacity, and aggregates plants by technology in each respective zone. Wind and solar technologies in a given zone have their own hourly power output traces, building and fuel costs.
A cost minimisation search is performed sequentially for a number of future years (every 5 years starting at 2020 for the pre-run scenarios in opencem.org.au) in which a financial year is simulated using a time-sliced approach to compute capacity decisions and then in full to compute dispatch decisions. New capacity decisions are assumed to be operational during the simulated year. The net of all existing and new capacity computed for one year is carried forward as the starting point to the next. For the first year, initial capacity consists of reported firm capacity by the Australian Energy Market Operator (AEMO) in 2018.
Energy can flow without restriction between all the zones in a region but notional interconnectors of fixed capacity (marked red in the figure) limit the amount of energy transmitted between regions.
By default, openCEM uses AEMO Integrated System Plan (ISP) 2018 data for technology and fuel costs, build limits, existing generation, electricity demand traces and renewable energy resource traces (i.e. wind and solar). For CST, openCEM by default uses "collector" only traces that estimate thermal output performance from a collector field [1]. With collector only traces, CST plants in openCEM can be configured to feature different storage sizes.
openCEM considers a set of generator technologies as displayed below. There are three classes of technology: Generators, storage and hybrids. New technologies will be added as required in future iterations of openCEM. Users may also add and configure other technologies to simulations.
| Technology | Class | Fuel | Dispatchable | Renewable | Flexible | Constraints |
|---|---|---|---|---|---|---|
| Biomass | Generator | Yes | Yes | Yes | Yes | Up to 10.6 TWh of yearly generation NEM wide [2] |
| Combined Cycle Gas Turbine | Generator | Yes | Yes | No | Mid | |
| CCGT with Carbon Capture and Storage (CCS) | Generator | Yes | Yes | No | Mid | (No emissions data *]) |
| Black Coal (existing) | Generator | Yes | Yes | No | Low | Moderate penalty on operating point change |
| Black Coal (new entrant) | Generator | Yes | Yes | No | Low | Moderate penalty on operating point change, New entrant costs and specs |
| Black Coal with CCS | Generator | Yes | Yes | No | Low | Moderate penalty on operating point change, (No emissions data *) |
| Brown Coal | Generator | Yes | Yes | No | Lowest | Steep penalty on operating point change |
| Brown Coal with CCS | Generator | Yes | Yes | No | Lowest | Steep penalty on operating point change, (No emissions data *) |
| Open Cycle Gas Turbine | Generator | Yes | Yes | No | Yes | |
| Solar PV Dual Axis tracking | Generator | No | No | Yes | Yes | |
| Solar PV Fixed Tilt | Generator | No | No | Yes | Yes | |
| Solar PV Single Axis | Generator | No | No | Yes | Yes | Build limits per NEM planning zone |
| Wind (low) | Generator | No | No | Yes | Yes | Build limits per NEM planning zone |
| Concentrating Solar Thermal 3h storage | Hybrid | No | Yes | Yes | Yes | Limited zones where permitted to build |
| Concentrating Solar Thermal 6h storage | Hybrid | No | Yes | Yes | Yes | Limited zones where permitted to build |
| Concentrating Solar Thermal 12h storage | Hybrid | No | Yes | Yes | Yes | Limited zones where permitted to build |
| Pumped Hydro Energy Storage 3h | Storage | No | Yes | N/A | Yes | Limited zones where permitted to build |
| Pumped Hydro Energy Storage 6h | Storage | No | Yes | N/A | Yes | Limited zones where permitted to build |
| Pumped Hydro Energy Storage 12h | Storage | No | Yes | N/A | Yes | Limited zones where permitted to build |
| Battery 1 hour | Storage | No | Yes | N/A | Yes | |
| Battery 2 hour | Storage | No | Yes | N/A | Yes | |
| Battery 3 hour | Storage | No | Yes | N/A | Yes | |
| Reciprocating Engine | Generator | Yes | Yes | No | Yes | |
| Wind (high) | Generator | No | No | Yes | Yes | Build limits per NEM planning zone, separate trace to Wind (low) |
| Hydro | Generator | No | Yes | Yes | Yes | Restricted GWh per year to emulate 10 year average behaviour |
| Gas thermal | Generator | Yes | Yes | No | Low | Moderate Penalty on change of operating point |
* CCS variants not in use because of incomplete/inaccurate NTNDP data at the time of release
Generators simulate a range of technologies whose output is affected by either fuel usage or hourly traces. Each generator technology is configured by specifying the following parameters:
- Build costs (can be specified per year and NEM zone)
- Fixed Operations and Maintenance (FOM) Costs
- Variable Operations and Maintenance (VOM) cost
- Fuel Costs (can be specified per year and per NEM zone)
- Hourly traces (can be specified per NEM zone)
Fuel based generators can dispatch their full nameplate capacity at all times but incur fuel costs when doing so. Trace based generators can dispatch up to the product of their nameplate capacity and hourly trace value.
Storage simulates large scale storage devices. openCEM manages the charges and discharge of storage capacity in each zone at every hour of dispatch calculations to run the system at the least cost. The following parameters configure storage devices:
- Build costs (can be specified per year and NEM zone)
- FOM Costs
- VOM cost
- Hours of charge
- Round trip efficiency
Hybrid technologies combine a trace based generator with a storage device. The storage device is similar to a storage device, but is charged by a "collector", a trace based generator whose size is proportional to nameplate capacity (usually greater). Energy stored in hybrid technologies is dispatched hourly according to the needs of the system to achieve the lowest system cost. It is possible for hybrid technologies to charge and dispatch simultaneously.
- Build costs (can be specified per year and NEM zone)
- FOM Costs
- VOM cost
- Hours of charge
- Collection multiple (a ratio of collected power to nameplate power)
- Hourly traces (can be specified per NEM zone)
openCEM optimises dispatch and capacity expansion decisions by seeking the lowest cost of operating the entire system in a financial year. The optimisation objective is to lower the sum of all annualised costs as described below.
Capacity expansion decisions incur annualised build costs calculated from the cost per MW of the technology, and adjusted by a fixed charge rate. The discount rate is defined by the user and the investment lifetime of technologies is assumed to be 30 years (except for batteries at 15 years and PHES at 50 years).
Annualised capital costs for expansion decisions are carried forward into subsequent years to account for repayment of investments in previous years. They are carried forward as a single lump cost and incremented by the build cost of each simulated year.
Simulations account for variable O&M costs in
All technologies incur a fixed O&M cost in $/MW/y for installed capacity.
openCEM assigns a sufficiently high cost to every MWh of demand that is not satisfied by the system to force the optimisation to reduce the cost of operating the system as much as possible. Typically this cost is set to satisfy or exceed reliability standards.
If a cost for emissions in $/kg is defined in simulations, they will be accounted for in seeking the lowest cost of operating the system and will influence both capacity and dispatch decisions
Shadow costs are costs used only to adjust the behaviour of simulations, and do not correspond to dollar amounts for a scenario. In other words, shadow costs are a provision to avoid numerical aberrations in simulations. These costs are seldom used and are tuned to have minimal influence on results.
- Negligible operational costs for transmission
- A steep cost to prevent exogenous retirement of generation beyond existing capacity
- A steep cost to prevent exogenous building of capacity beyond build limits.
- A steep cost to 'surplus' energy, a model relaxation to prevent infeasibility conditions on the optimisation.
openCEM offers the following policy constraints for scenarios:
- NEM wide renewable energy targets as a minimum ratio of total generation per year.
- NEM wide renewable energy targets as minimum generation in GWh per year.
- Region wide renewable energy targets as a minimum ratio of total generation per region per year.
- NEM wide maximum emissions, specified in MT per year.
- NEM wide emission costs, specified in $/kg
These constraints either penalise or enforce hard limits that influence capacity and dispatch decisions in simulations. Model decisions seek to find the least total cost of running the system under any combination of specified policies.
Electricity transmission between NEM planning regions is modelled using a linear pipieline or "truck route" model for the topology shown below.
Pipeline transmission constraints ensure that at each dispatch hour, the amount of energy transmitted between any two zones is less than the prescribed thermal limits for that link. Transmission capacity expansion decisions can increase the thermal limit on a given link (bidirectionally) at a cost in $/MW/km defined for each link. The decision to upgrade a link is made simultaneously with all the other capacity expansion decisions, considering the trade-off between imports/exports of energy and local generation/storage.
All links have separate forward and reverse thermal limits with initial values as described in [3], except for new links which start with an initial capacity of zero. All links by default assume a 2% transmission loss in each direction, except for inter-regional links which also incorporate AEMO proportioning factors on the applicable direction of the link.
Transmission configuration options can be found in the ZONE_INTERCONS dictionary contained in the const module. For more details click here
The operating reserve constraint defines a margin of minimum available capacity at each hour for each NEM region. Reserve operating capacity is defined as the sum of:
- Not dispatched capacity from flexible generators (quick start)
- Non dispatched but committed capacity from non flexible generators (spinning reserve)
- Non dispatched capacity from storage and hybrid devices, provided stored energy is available (quick start)
The constraint is enforced at each hour of dispatch, and the reserve operating capacity must be greater or equal than a percentage of region demand for that hour. By default operating reserves are set at 7.5pct of demand.
In addition to unserved energy costs, capacity expansion calculations must ensure that unserved energy in each region does not exceed 0.002% of demand.
[1] Australian Concentrating Solar Thermal Industry Roadmap, Appendix 1 Concentrating Solar Thermal Technology Status, August 2018, ITP Thermal Pty, Limited
[2] Australian Bioenergy Roadmap, Clean Energy Council of Australia, 2008.
[3] Modelling Transmission Frameworks Review, Report EPR0019 Roam Consulting, February 2013.
Return to the home page