README.md 2.17 KB
Newer Older
Pablo Aznar's avatar
Pablo Aznar committed
1
2
# SecureGrid

Pablo Aznar's avatar
Pablo Aznar committed
3
4
Deep Learning based Attack Detection System for Smart Grids

Carlos A. Iglesias's avatar
Carlos A. Iglesias committed
5
6
7
8
9
10
## Required modules
* Keras
* There is a recognized bug in numpy 1.16.3. Thus, the numpy version should be 1.16.1.
pip uninstall numpy
pip install --upgrade numpy==1.16.1

Pablo Aznar's avatar
Pablo Aznar committed
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
## Usage

In order to detect attacks, the power consumption values of the houses are analyzed. For that reason, first, the needed DataFrames to feed the neural network (autoencoder) have to be created.

For this purpose, the notebook dataframe_creation is used. This notebook generates .pkl files that contain the DataFrames with the necessary data. In addition, these DataFrames contain the following features:

| Feature  | Description |
| ------------- | ------------- |
| Day  | Current day of the first window value  |
| Hour  | Current hour of the first window value  |
| Minute  | Current minute of the first window value  |
| Pn  | Power consumption window values  |
| Mean  | Mean of the window values  |
| Mean_i - Mean_i-1  | Difference between the mean of the window values and the mean of the previous window values  |
| s  | Standard deviation of the window values |
| Pn - P1  | Difference between the last and first value of the window  |
| Q1  | First quartile of the window values  |
| Q2  | Median of the window values  |
| Q3  | Third quartile of the window values  |
| IQR  | Interquartile range of the window values  |

Once the DataFrames are created, they are used to feed the autoencoder. Therefore, the conv1d_autoencoder.py file has to be configured.

The normal_data_path variable has to contain the path to the .pkl file that contains data without attacks, that is to say, a normal behaviour of the houses. In addition, the attack_data_path variable has to contain the path to the .pkl file that contains the data that is wanted to be analyzed in order to detect attacks.

Furthermore, in order to train the autoencoder, the DO_TRAINING variable has to be set to True.

Finally, the following command executes the system:

	$ python conv1d_autoencoder.py

## Results

Once the system is executed, it generates the predicted_labels.csv file that contains the labels that classify every entry of the DataFrame into attack (1) or normal behaviour (0).