-
Notifications
You must be signed in to change notification settings - Fork 9
Charging strategies
The core of SpiceEV are the different charging strategies. They decide how to react to events, when to charge the cars and by how much. To see how to set strategy options, refer to this wiki page.
All charging strategies support the EPS option, which defines the difference under which two floating point numbers are considered equal. In other words, the value chosen for EPS determines the precision of the simulation. The smaller it is the more precise the calculations are. The downside to this is an increase running time. For some numerical procedures the algorithm might get stuck completely if EPS is too small. The default is 10^-5.
Every strategy also supports the strategy option allow_negative_soc. This option sets negative SoCs to 0 and continues the simulation instead of aborting. The default value is False, setting it to True will activate the option. The option is passed in strategy_option list.
Your most basic and dumb strategy. Charges one vehicle after the next with full power until the desired state of charge (SoC) is reached. Depending on the grid connector (GC), multiple cars may be charged in one timestep.
May charge above the desired SoC if there is surplus feed-in power or the energy price is cheap.
Stationary batteries supported: ✔️ V2G supported: ✔️
| Strategy option | default | explanation |
|---|---|---|
| CONCURRENCY | 1.0 | Reduce maximum available power at each charging station. A value of 0.5 means only half the power is available. |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
Each car is charged such that it uses its complete standing time to reach the desired SoC. May charge more power (and above the desired SoC) if there is surplus feed-in power or if the energy price falls below a certain PRICE_THRESHOLD.
Implementation notice: uses a binary search to find the minimum viable charging power.
Stationary batteries supported: ✔️ V2G supported: ✔️
| Strategy option | default | explanation |
|---|---|---|
| ITERATIONS | 12 | Minimum depth of binary search to find charging power |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
This algorithm first determines the cheapest group of time intervals sufficient to charge all vehicles according to their needs. Moves all charging events to those time intervals and charges them with full power, similar to the greedy strategy. Only one grid connector supported.
Stationary batteries supported: ✔️ V2G supported: ✔️
| Strategy option | default | explanation |
|---|---|---|
| CONCURRENCY | 1.0 | Reduce maximum available power at each charging station. A value of 0.5 means only half the power is available. |
| HORIZON | 24 | number of hours to look ahead |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
| DISCHARGE_LIMIT | 0 | V2G: maximum depth of discharge [0-1] |
When using this strategy, price information within the next HORIZON hours is evaluated. The goal is to divide standing times into periods of equal prices. A vehicle is now charged such that is uses the entire duration of all periods with the lowest price combined to reach its desired SOC. In case that time is not sufficient the periods of the second cheapest price are used to charge as much of the remaining delta SOC as possible, again in a balanced way with respect to power.
Stationary batteries supported: ✔️ V2G supported: ✔️
| Strategy option | default | explanation |
|---|---|---|
| CONCURRENCY | 1.0 | Reduce maximum available power at each charging station. A value of 0.5 means only half the power is available. |
| HORIZON | 24 | number of hours to look ahead |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
| DISCHARGE_LIMIT | 0 | V2G: maximum depth of discharge [0-1] |
| V2G_POWER_FACTOR | 1 | Fraction of max battery power used for discharge process [0-1] |
Charging strategy that prioritizes times with lower power costs. The idea is to find the minimum viable cost threshold (per car or for the whole fleet). This way, timesteps with less external load and smaller costs are prioritized for loading. In times with low cost, the maximum available power is used, no computation needed.
The computed cost is used to find the corresponding power for each future timestep, which is then distributed to the vehicles. This distribution is done using sub-strategies:
- greedy: vehicles charge as much as possible, one after the other (vehicles below desired SoC charge first)
- needy: power is allocated according to missing power needed to reach the desired SoC
- balanced: power is distributed evenly among vehicles below desired SoC. Surplus is then distributed evenly among all cars
- individual: cost is not computed for whole fleet, but for each vehicle individually
Stationary batteries supported: ✔️ V2G supported: ❎
| Strategy option | default | explanation |
|---|---|---|
| HORIZON | 24 | number of hours to look ahead |
| ITERATIONS | 16 | Minimum depth of binary search to find cost |
| LOAD_STRAT | "greedy" | charging strategy, see above |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
Allocate power according to grid operator schedule. ScheduleForesight looks into the future (until all cars have left, at most 24h) and tries to adjust schedule so that all cars can be charged for the next trip. Implements different sub-strategies:
- greedy: Cars charge as much as possible to reach the set schedule. Vehicles below desired SoC get charged first, then in order of departure.
- needy: Cars charge in proportion to the energy required to reach their desired SoC. All cars charge at the same time to reach the desired schedule. Cars not in need of much power are handled first, as they may fall below their minimum power limit and leave more power to the rest of the fleet.
- balanced: The scheduled power target is distributed evenly among the fleet. If this causes vehicles to fall below their minimum power limit, cars are excluded from charging in order of importance (fully charged, time of departure).
If ScheduleForesight takes too long, you may try to increase the EPS parameter.
Stationary batteries supported: ✔️ V2G supported: (✔️, ScheduleForesight only)
| Strategy option | default | explanation |
|---|---|---|
| LOAD_STRAT | "needy" | charging strategy, see above |
Given a time window of high load, tries to charge outside this window. Different sub-strategies supported:
- greedy: vehicles charge as much as possible, one after the other (vehicles below desired SoC charge first)
- needy: power is allocated according to missing power needed to reach the desired SoC
- balanced: power is distributed evenly among vehicles below desired SoC. Surplus is then distributed evenly among all cars
- individual: cost is not computed for whole fleet, but for each vehicle individually
Stationary batteries supported: ✔️ V2G supported: ❎
| Strategy option | default | explanation |
|---|---|---|
| LOAD_STRAT | "greedy" | charging strategy, see Inverse |
There are time windows during which charging is encouraged and there are those where it is discouraged. These time windows are determined by the grid operator (similar to Schedule strategy). During those windows where charging is encouraged the vehicles are charged with one of the following sub-strategies:
- greedy: charge vehicles that are below their desired SOC level one after the other, the rest is ordered by time of departure (earlier departures charged first)
- needy: charge vehicles with little power missing to desired SoC first, vehicles are charged one after the other
- balanced (DEFAULT): Go through vehicles one by one determining the amount of power for charging such that vehicle uses entire cross section of standing time and charging window
If not all vehicles can be charged during the time windows where charging is encouraged, the rest of the energy is charged in non-charging windows. The remaining energy consumption of the entire fleet is balanced out across all non-charging windows to keep power peaks as low as possible.
| Strategy option | default | explanation |
|---|---|---|
| CONCURRENCY | 1.0 | Reduce maximum available power at each charging station. A value of 0.5 means only half the power is available. |
| HORIZON | 24 | number of hours to look ahead |
| PRICE_THRESHOLD | 0.001 | A price below this is considered cheap. Unit: € / 1 kWh |
| DISCHARGE_LIMIT | 0 | V2G: maximum depth of discharge [0-1] |
| LOAD_STRAT | balanced | Sub-strategies for behaviour within charging windows (see description above for options and explanations) |