# SecureGrid Deep Learning based Attack Detection System for Smart Grids

This is the software of Deep Learning based Attack Detection System for Smart Grids

It is composed of three modules, which should be executed in order:

* `simulator/`  - Data generation of the power of residential houses ander attacks
* `data-preprocessing/` - Data preprocessing of the results of the simulator
* `attack-detector/` - deep learning module for detecting attacks 

Please, refer to every module for its installation, configuration and execution.

The directory `data` contains the result of the execution of the three modules when the simulation has been set to an `attackPercentageValue` of 0, 10, 20 and 30.
It is organized as follows:
- normal2/ (0 attacks)
    + demo_0.hdf5 (generated by simulator)
    + house_N.pkl (one file for each house generated by data-preprocessing, with N in 0..37)
- anomaly_20_2, anomaly_30_2, files with attacks.