This document summarizes an electrical grid stability dataset from UCI. It describes 11 predictive attributes like reaction time and power consumed/produced for electricity producers and participants. The goals are stability, a value measuring system instability, and a stability label (stable/unstable). Correlations between attributes and stability are provided. The document also references a related paper on decentralized smart grid control that aims to stabilize the power system through distributed generation and price-based incentives for consumers.

1. Electrical
Grid
Stability
Simulated
電網穩定性模擬
10766012 Jason 陳遠任

2. Data Set Information
• From UCI
• http://archive.ics.uci.edu/ml/datasets/Electrical+Grid+Stability+
Simulated+Data+#
• The analysis is performed for different sets of input values
using the methodology similar to that described in
• [Schà ¤fer, Benjamin, et al. 'Taming instabilities in power grid
networks by decentralized control.' The European Physical
Journal Special Topics 225.3 (2016): 569-582.].

3. Attribute Information
• 11 predictive attributes, 1 non-predictive(p1) :
• 1. tau[x]: reaction time of
Tau1 - the value for electricity producer.
• 2. p[x]: nominal power consumed(negative)/produced(positive)
p1 = abs(p2 + p3 + p4)
• 3. g[x]: coefficient (gamma) proportional to price elasticity
g1 - the value for electricity producer.

4. No.1
No.2
No.3No.4
electricity producer
participants
Electricity Producer

5. Attribute Information
• 2 goal fields :
• 4. stab: the maximal real part of the characteristic equation root
(if positive - the system is linearly unstable)(real)
• 5. stabf: the stability label of the system (categorical:
stable/unstable)

6. Correlation of
Variable stab
tau1 0.275761
tau2 0.290975
tau3 0.280700
tau4 0.278576
p1 0.010278
p2 0.006255
p3 -0.003321
p4 -0.020786
g1 0.282774
g2 0.293601
g3 0.308235
g4 0.279214
stab 1.000000

7. References from paper
• Decentral Smart Grid Control (DSGC)
• star-like topology星形拓撲結構：
connect large plants with regional consumers
• distributed generation分佈式發電
become unindirectional
• An alternative approach without massive communication
between consumers and producers directly utilizes the grid
frequency to adjust production and consumption.
• The frequency increases in times of power excess while it
decreases in times of underproduction.

8. References from paper
• Instead of paying a constant price for electric power, consumers
are presented with a linear price-frequency relation
• The aims of DSGC is to stabilize the power system by
encouraging consumers to lower their consumption in times of
high load and low production and increase consumption in times
of low load but high production.

9. Original version

11. Regression using
linear regression

12. Classification using
Tree

13. Classification using
Logistic regression

14. Modified version

15. Sampling at First

16. Regression using
Linear regression
Training Data Testing Data

17. Regression using
Linear regression

18. Classification using
Tree
Training Data Testing Data

19. Classification using
Tree

20. Classification using
Logistic regression
Training Data Testing Data
Encoding

21. Classification using
Logistic regression

22. Summary
• Linear regression
• Tree
• Logistic regression