DBI-MAT
An energy system modelling and optimization tool, written in python, to model local energy systems (microgrids).
Back
Target application
The tool is used to determine optimized scenarios to integrate renewable energy sources, hydrogen applications or other processes in an economical and ecological feasible way. In previous project the focus was on the technical and economical feasibility of the utilization of hydrogen or oxygen produced via electrolysis.
Methodology
Figure 1 describes the structure of the DBI-MAT tool.
- At first the user must define all technical components
- Then, further component-specific parameters must be set.
- When all components are initialized, the ports of each component can be linked to balance groups and its priorities and the technical system (microgrid) is complete.
- Before solving the system, temporal data (e.g. weather data or energy supply/demand),can be set, too.
- The microgrid gets solved technical first and can be solved economical afterwards.
- The results can then either be used to optimize the microgrid further, exported or used for visualizations.
Figure 1: Schematic DBI-MAT chart
Solving a balance group
A balance group counts as solved, if an equilibrium is archived. E.g. surplus power can be fed into the grid and power
demand, not covered by a power source, can be drawn from the grid. The feed-in or feed-out profile of the passive grid
component is not fixed, instead it is calculated from the time series of the power source and the consumer, as
illustrated in Figure 2.
Figure 2: Solving a Balance Group
But, a solution is not necessarily achieved, if the passive component is subject to limits (e.g. maximal or minimal amount of energy, which can be fed-in or drawn from a grid). In this case a more advanced logic has to be implemented in order to ignore the limits, alter the time series of active components or, if multiple passive components are available, solve other components again.
Technical Solution
The balance groups are solved by the tool one after another. The order depends on the configuration of the technical system and the intermediate dependencies of the balance groups and its components. Thereby all material- and energy flows within the system are calculated. After calculating each balance group for each time step, an electric energy flow or a hydrogen mass flow results for each hour of the year for the in- and outputs of each component. The energy- or massflows for each component at each time step are output as the solution of the technical calculation.
Economical Solution
After the technical solution is calculated, the economical solution can be calculated via annuity method. The technical solution has to be calculated first, because its results, the temporal data, might be necessary to calculate demand-related costs, e.g. demand of deionized water for electrolysis, which depends on the electricity consumption.
Temporal resolution
The tool works with different temporal resolutions. The temporal resolution is defined by the used time series (load profiles) which are set as an input of the system. Typical temporal resolutions which have been used in previous projects are 15 minutes or 1 hour.
Results and Post-Processing
The Results can be returned in typical python data structures, which can either be exported to other file-formats, in order to share the results with customers, or directly used for further analysis or visualization, with existing python packages (e.g. plotly or matplotlib)
Energy System Model
The following Energy System consists of several components which interact with each other. The components are connected through 4 balance groups, two electric and two hydrogen balance groups, as shown in Figure 3. It is assumed, that the wind farm exists already and has an installed power of 50 MW, while the size of the PV plant is not defined yet. The electrolyser system consists of a 20 MW alkaline electrolyser for the base load and a 10 MW PEM electrolyser. The pipeline is a repurposed natural gas pipeline and the hydrogen storage is an oversized salt cavern plus surface facility.
Figure 3: Technical System of an Energy Park
First electric balance group
The wind farm, PV plant, power grid and the electrolysis system are connected to the first electric balance group. The PV system and wind farm components are active. The profiles of the RE plants are set by available wind and solar radiation data, and the electrolysis primarily takes the RE power. In this process, electricity is left over when the rated power of the electrolysis is exceeded or the minimum load of the electrolysis is not reached. The surplus electricity is then fed into the power grid and remunerated at a fixed price. In order to increase the full load hours of the electrolysis, an operating point can be selected up to which the electricity supply of the RE plants is supplemented by grid electricity. For this electricity purchase a working price is set.
Second electric balance group
The second electrical balance group considers the 'balance of plant' (BoP) of the electrolysis and the power grid. BoP is defined as an active component and has a defined power consumption that is independent of the electrolysis load. Only when switching between operating and standby mode, the BoP stalls demand change. The power is drawn from the grid at a fixed energy price.
Hydrogen balance group
The first hydrogen balancing group includes the electrolysis, the transport pipeline and the underground storage. The second balancing group includes the transport pipeline, which is connected to the consumer. Priorities for the passive components can be set in the tool. In the system definition, it was determined, that the transport pipeline has priority over the hydrogen underground storage. The hydrogen produced is primarily delivered directly to the customer via the transport pipeline. With the help of the underground storage, the tool tries to provide the customer with a constant amount of hydrogen throughout the year. The annual average of the hydrogen supplied is determined. If the annual average is exceeded or undercut, the hydrogen is stored in or withdrawn from the underground storage. At the same time the tool checks, whether the amount of hydrogen produced by the electrolyser exceeds the capacity of the salt cavern or the geological limitation of 10 bar pressure in- or decrease within a 24-hour window. The underground storage and the pipeline are 'out of scope' in this research.
Scenarios
Four different scenarios are analysed, as per description below.
- Scenario 1
- Scenario 2
- Scenario 3
- Scenario 4
Figure 3-1: Alternations of the Technical System for Scenario 1
In Scenario 1, it is assumed, that the electrolysis plant is supplied with electricity from a 50 MW wind farm only. Solely the energy demand of BoP is covered by the purchase of green electricity from the power grid to simplify the scenario.
Figure 3-2: Alternations of the Technical System for Scenario 2
In this scenario, it is considered, that the wind farm is built as planned with a nominal capacity of 50 MW. The electrolysis system is connected to the power grid and is always operated at full load due to the grid supply. Thus, the electrolysis plant reaches 8760 full load hours per year. The use of wind power always has priority over the grid purchase.
Figure 3-3: Alternations of the Technical System for Scenario 3
Scenario 3 considers the case, where the electrolysis plant, with the exception for BoP, does not draw any electricity from the grid. A PV plant is considered as an additional source of renewable electricity.The scaling of the wind farm and the PV plant is optimized in a first step.
Figure 3-4: Alternations of the Technical System for Scenario 4
Scenario 4 considers the case, where the electrolysis plant is supplied with renewable electricity from the wind farm
and the PV plant. The electrolysis plant is connected to the power grid and always consumes electricity until it reaches
the maximum capacity. Accordingly, the electrolysis plant reaches 8760 full load hours per year, as in scenario 2. The
use of electricity form the wind farm and the PV plant have always priority over the purchase of grid electricity. For
this scenario the scaling of the wind park and the pv plant is optimized with respect to the total hydrogen costs.
Results
The sensitivities are expressed as the relative change over the percentage deviation of th parameters from their respective base values.
Sensitivity Analysis
- Scenario 1
- Scenario 2
- Scenario 3
- Scenario 4
Sensitivity Analysis of Scenario 1 - 50 MW Windpark, no grid, no PV plant
The change in the prime cost, with a relative change in the parameters, with respect to the baseline, was largest for the wind farm CAPEX and for the imputed interest rate. Both parameters cause more than 2% change in the cost price for a 5% change in the base value. The smallest influence on the production costs was found for the electricity purchase costs. Since in this scenario no electricity is purchased from the grid for the electrolysis process itself, but only for the BoP, the total electricity consumption is low and the electricity price therefore only has a minor impact. In this scenario a yearly hydrogen production of 2133 t was calculated.
Sensitivity Analysis Scenario 2
In the base-case hydrogen costs of 7.28 €/kg were calculated. The costs for the best and worst case are 5.23 €/kg and 8.23 €/kg, respectively. The cost of the base case is about 72 % of the base case, while the cost of the worst case, relative to the base case, is about 115 %. In the best-case scenario, most of the costs were caused by teh gas storage and the wind farm. On the other hand, in the base- and worst-case electricity procurement was determined to be the largest cost center.This is due to the high influence of the electricity price on the cost price, since in this scenario a large amount of energy is purchased from the power grid. Due to the connection to the power grid the electrolyser runs for 8760 hours a year. Therefore the annual hydrogen production in this scenario is about 4636 t.
Sensitivity Analysis of Scenario 3
In the sensitivity analysis of this scenario, the interest rate was found to have a large impact on the hydrogen production costs. For the imputed interest rate, a change in the hydrogen production cost of 2.69 % to the base cost, with a change in the interest rate of 5 % of the base value, was determined. In absolute terms, the hydrogen production cots change by 0.589 €/kg when the interest rate changes by 1 %. A high influence on the hydrogen production costs was furthermore determined for the CAPEX of the Wind park and the PV plant. As in Scenario 2 the electrolyser runs for 8760 hours a year, therefore the amount of hydrogen produced is the same, 4636 t.
Optimization
To find out, which size of the wind park and pv plant leads to a minimal LCOH a hyperparameter optimization, based on a grid search approach is used.
In this scenario the wind park was optimized in a step size of 6MW installed capacity and the pv plant with a step size of 6.5 MW installed capacity.
The optimal results for the wind park are 54 MW nominal power and for the PV plant 105 MW. The optimum is also very flat, as seen in Figure 4-3-2, therefore a deviation from the optimum of a few Megawatts leads to good results.
Optimization Results of the nominal power of wind and pv compared to LCOH
The ratio of PV to wind is 2:1, and the ratio of electrolysis to wind is 1.8:1, while the ratio of electrolysis to PV is 3.5:1.
Sensitivity Analysis
Sensitivity Analysis of Scenario 4
The sensitivity analysis has shown that the electricity price has the largest influence on hydrogen production costs in scenario 4.
Optimization
To find out, which size of the wind park and pv plant leads to a minimal LCOH a hyperparameter optimization, based on a grid search approach is used.
In this scenario the wind park was optimized in a step size of 2.5 MW installed capacity and the PV plant with a step size of 6 MW installed capacity.
The optimal results for the wind park are 30 MW nominal power and for the PV plant 64 MW. The optimum is also very flat, as seen in Figure 4-3-2, therefore a deviation from the optimum of a few Megawatts leads to good results.
Optimization Results of the nominal power of wind and pv compared to LCOH
As in Scenario 3, the ratio of PV to wind is 2:1, and the ratio of electrolysis to wind is 1.8:1, but the ratio of electrolysis to PV is 2.1:1.
Conclusion
In summary, each scenario leads to different ranges of production costs for green hydrogen, as shown in Figure 5. By comparing secenario 1 to the others the largest impact on the production costs is the availability of renewables. An additional PV plant or green electricity from grid has the largest impact, due to an increase in full load hours of the electrolyser. A decisive factor that is not considered is the customers' willingness to pay. In this research, the gas storage is also an underground salt cavern. In the considered scenarios, the cavern storage is responsible for a large share of the prime costs, since the electrolysis is the only customer. The reason for the decision to use a salt cavern is the research character and therefore not considered in detail. It is assumed that the cost price will decrease the more customers use the storage in the future.
Figure 5: Comparing the LCOH of Best-, Worst- and Base-Case of each Scenario by share
Authors
Dr. Martin Pumpa | Ing. Frank Fischer (M.Sc) | Martin Heckner (M.Sc.)