Comments (2)
The IEEE14 grid used is not fully configured correctly due to missing information, which leads to weird line loading values.
I'm not a developer of grid2op, but I can try to explain the loss:
Previously, only a high line loading (in %) was penalized so that the overall loadings were minimized. In the new loss, the system loss (energy lost as heat in the various components of the grid) is to be minimized, which is simlar to before, but not quite. This is the term losses_cost.
redisp_cost is a cost term derived by the summing the power produced by the generators that can be redispatched (controlled). This is to be minimized as well.
Both terms are adjusted by marginal_cost, which is the maximum cost term of the active generators that can be redispatched. This penalizes the use of the most expensive generator in general.
The goal is to not use the redispatchable generators (iirc they should model fossil fuel power plants) and minimize the overall system losses and the new loss is an approximation of that.
from grid2op.
Thank you @jhmenke your explanation is crystal clear and totally correct :-)
The answer to the original question has been carried out in mail and a dedicated meetings.
For the community, in case someone wonder the same things:
Most of the problems came from the the reuse of l2rpn2019 data for this new environment. Let me recap here and closing this issue:
First, the time information is missing, which might be a minor problem.
The date and time information are present, as a default value is used. Data start by default on January 1st, 2019.
Second, the definition of the thermal limit is confusing for us. I think the thermal limit is defined as apparent power in the previous version. Now it is defined as follows with significantly different magnitude.
We have noticed this issue. The underlying "problem" is the voltages used in the default case14 of pandapower. A new environment will be available in the next release (0.5.7) that should be more realistic on this point of view.
Even if I change this to the previous value, it will be changed again during the calculation.
Don't try to use any data from the l2rpn 2019 environment into this one. This is a bad idea and will lead to wrong behaviour.
Similarly, the state.rho values also become very strange. I think it might be a bug in the program.
After investigation, there is not bug at all in this regard. Bug was probably due (lack of information to reproduce this bug) to the use of l2rpn 2019 environment.
The setpoint of the latest version will lead to the diverge of power flow even when I take no action, which needs to be modified.
This environment is completely different from the l2rpn2019. This is not a bug but a desired feature.
Jan-Hendrik perfectly explain the change of reward (see above post).
Thanks a lot for the constructive remarks
Benjamin
from grid2op.
Related Issues (20)
- Prevent the environment from disconnecting a chosen set of lines even in case of overflow HOT 3
- Wrong importation when importing a file named "utils.py" after grid2op HOT 2
- In 1.9.9
- Write the grid2op Markov Decision Process in the documentation
- Work on the reward module
- Bug when no gen redispatching but some storage units on the grid
- Do not start with ref topo == everything to busbar 1 HOT 1
- Codalab scoring throws an error when filtering chronics HOT 16
- Wrong total time step when providing "max_step" in ScoreL2RPN2020 HOT 1
- 'h_forecast' bug when creating env without forecast for l2rpn_idf_2023 HOT 4
- Expand Observation + Action to check for game-ending topology changes HOT 4
- Simulation on current state reapply actual_dispatch when a redispatch action has been applied before HOT 7
- Being able to specify position of elements (powerlines, loads, generators, storages directly in the layout)
- Minor Change in Runner to allow for parallel computing with tensorflow models and Pool HOT 2
- Switch problems for maintenance functions in the environment("l2rpn_wcci_2020") HOT 2
- Tutorial 1: Grid2Op integration with existing frameworks HOT 2
- [Feature] Set init state HOT 1
- In reset, also add possibility to change the parameters, forecast parameters, max iter and skip time steps HOT 2
- Improve the handling of "name_chronics_to_backend"
- Optimal Clustering of Substations for Topology Optimization Using the Louvain Algorithm HOT 7
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from grid2op.