Presentation - University of Wisconsin–Madison

Download Report

Transcript Presentation - University of Wisconsin–Madison

An Initial ANN Approach to LMP
Classification & Prediction
Honghao Zheng
University of Wisconsin – Madison
1
ECE 539 Project
2010 Fall
©Honghao Zheng 2010
Motivation
• Locational Marginal Price (LMP), which is usually
referred to as “shadow price” of the power grid,
gives efficient measurement of power production
and the consumption of energy at the different bus
nodes.
• The prediction of LMP at different zonal price could
benefit the individual biding for the electricity at
different nodes in the power system.
• If we could locate the feature vector, then we could
use ANN method to predict the PMU value at
certain time in certain place.
2
ECE 539 Project
2010 Fall
Previous WORK
[1] Zonal Prices Analysis Supported by a Data Mining
Based Methodology, J. Ferreira, S. Ramos, Z. Vale and J.
P. Soares. IEEE Conference Proceedings, 2010.
[2] Zone Clustering LMP with Location information using
an Improved Fuzzy C-Mean, Se-Hwan Jang, Jin-Ho Kim,
Sang-Hyuk Lee and June-Ho Park, IEEE Conference
Proceedings.
[3] High value wind: A method to explore the relationship
between wind speed and electricity locational marginal
price, Geoffrey McD. Lewis
3
ECE 539 Project
2010 Fall
Methodology
Step No.1
The first step of the project would be manually filtering the
large amount of LMP hourly data into different groups.
The LMP data is downloaded from the website of Midwest
Independent System Operator (Midwest ISO).
Filters: Time = April, Value = LMP, Type = LoadZone
4
ECE 539 Project
2010 Fall
Methodology
Step No.2
Step No.2 mainly concerns with the feature vector
selection.
Major Issue that may influence the value of LMP:
1. Grid structure; 2. Weekday or Weekend (7 days in one
week); 3. Different period in a day (Morning/Noon/Evening)
1. Generate physical position of different load zones;
2. Grant different weights to the seven days;
3. Choose 4 hours to be one period, all have high LMP.
Feature Vector Dimension: 4
5
ECE 539 Project
2010 Fall
Methodology
Step No.2(Cont’d)
Generate Geographic Location:
ALTE
3
DECO
24
IGEE
4
SIPC
1
6
ECE 539 Project
2010 Fall
ALTW
11
DPC
6
MDU
4
TVA
1
AMRN
35
EKPC
1
MGE
1
UPPC
1
CILC
3
FE
15
MP
8
WEC
5
CIN
13
GRE
9
NIPS
4
WPS
2
CONS
21
HE
2
NSP
26
WR
1
CWLD
3
IP
10
OTP
15
CWLP
3
IPL
2
SIGE
4
TOTAL
238
Methodology
Step No.3
The Step No.3 Using MLP Mapping to Test the data
1. Classification Criterion: <35 LOW LMP, 35~50 MID
LMP, >50 HIGH LMP
2. Separate the 28 days in Apr into 4 weeks, labeled W1,
W2, W3, W4.
3. Formulate 3 tests: Training Set (W1, W1&W2,
W1&W2&W3), Testing Set (W2&W3&W4, W3&W4, W4)
Here the testing set functions as the prediction, because in
the future if we know the feature vector, we could use MLP
to predict the LMP value directly.
7
ECE 539 Project
2010 Fall
Simulation Result
Training
T.1
T.2
T.3
Ways of Training
T.1
T.2
T.3
Training Set
W1
W1, W2
W1, W2, W3
Layer = 3, Neurons/Layer = 5
Training Rate
Prediction Rate
71.47%
55.12%
67.98%
54.89%
65.55%
53.20%
Testing Set
W2, W3, W4
W3, W4
W4
Layer = 4, Neurons/Layer = 6
Training Rate
Prediction Rate
83.87%
50.26%
67.467%
53.12%
63.15%
61.50%
Comment:
1. Training Rate does not have necessary relationship
with the Prediction Rate
2. Prediction Rate (Testing Rate) is not that high as
expected.
3. The randomly-generated location may result in the
inconsistency.
8
ECE 539 Project
2010 Fall
Discussion
Disturbance & Ways to Improve
Disturbance:
1. Inconsistency in the location
2. The classification of the LMP may be too rough to
determine the exact position of LMP.
3. Possible feature difference not quite clear.
Ways to Improve:
1. Acquire actual geographic location (longitude,
latitude).
2. Classify the LMP value range smaller.
3. To make the range difference between the features to
be obvious.
9
ECE 539 Project
2010 Fall
Conclusion
1. ANN: quite a useful tool in the power system, yet the
application of prediction for LMP value is rare.
2. The result that has the best performance (63%) is
roughly acceptable, yet not the expected value.
3. Outlook: make the model more realistic; trying to get
the location data from the government; change MLP
algorithm to better suitable for LMP Prediction
10
ECE 539 Project
2010 Fall