Motivation and aim
The energy transition leads to increasing planning needs in the electricity industry like:
1. Electricity grid development planning
2. Bidding zone configuration assessment
3. German hard coal phase out analysis
Today therefore used electricity market simulation software requires ongoing and fixed IT expenditures. Further, only dedicated data research and calibration simulations create a usable input data set with close to reality results.
Maon aims to develop an electricity market simulation and to provide it as a browser based Web Service. This approach enables simulations without the installation of software or hardware. Moreover, scenarios can be generated with already calibrated input data sets for Europe-wide and yearly simulations.
The new approach involves no command lines nor direct application of optimization tools. Instead, the user is supported by parameterization tools and data quality checks. Additionally, results as unit commitment and market prices are prepared for evaluations and applications as welfare analysis or power-flow simulations.
Access to simulations and data is available via the browser-based Graphical User Interface and via the Web Service for embedding machine-to-machine processes. Figure 1 depicts this software architecture and the used technology stack.
Figure 1: Software architecture and technology stack
The Graphical User Interface provides features for up- and download, parameterization, data visualization and the simulation. Such features are based on the underlying Web Service. Due to this software architecture, simulations can be carried out on demand (software-as-a-service).
Database and processing
The electricity market simulation handles a Europe-wide input and output data set with degrees of freedom in international exchange, demand, renewables, thermal and hydro power plants for one time coupled year with a hourly granularity. Figure 2 displays the processed data.
Figure 2: In- and output of the Simulation Core respectively Database
If a positive initial plausibility, feasibility and complexity check occurs (lasts a few minutes) the procedure guarantees a break-free simulation run. In case of failure, the user gets a problem description and a solution proposal. Hence, manual data processing can be reduced.
The Simulation Core uses files formatted in comma-separated values (csv). It is based on a three-staged optimization problem solving procedure displayed in Figure 3. Input data is checked upfront and raw results are processed afterwards for further usage.
Figure 3: Three-staged optimization problem solving procedure
First, on-off dispatch decisions are neglected and non-linear equations linearized. The resulting problem solves a barrier method. Second, determined net positions per bidding zone are fixed. Based on such, sub-problems per bidding zone are further divided into weekly sub-sub-problems. Such are solved successively via a branch-and-cut method. Afterwards, derived on-off dispatch decisions are fixed. Third, based on derived operation points the problem is linearized and solved. The procedure handles numerical problems internally via optimizer restarts or re-calibrations. During simulation runs identical software and hardware is used. Thus, the Simulation Core works automatic and provides reproducible results.
The new approach enables web-based, automatic and reproducible electricity market simulations without the need for permanent provision of software and hardware for users.