 |
Data Mining for Energy Saving |
Background:
Aluminum factory uses 5,000 electrolytic cells in aluminum production. An electric current
of 100,000A is applied to each of the cells. The anode of the cells is made of an
electro-conducting alloy (Fe-P-Si-Mn-S).
Problem:
High electric power consumption caused by the high resistance of micro-cracks
formed in the cell alloy during production. The voltage drop across each cell is 0.4-0.5
volts, and the total power consumption by the 5,000 cells in the aluminum production line
is about 0.5x100,000x5,000 = 250,000,000 watts = 250,000 KW
Goal:
Reduce power consumption by reducing the micro-cracks in the alloy.
Method:
Use history data from operations to build a good model that can be used to control and
reduce the formation of micro-cracks in the cell alloy.
Finding:
Factors affecting micro-cracks |
How to control/optimize |
| Mn | Increased |
Si |
Reduced |
C |
No change |
S |
No significant influence |
P |
No change |
Result:
Micro-cracks were significantly reduced, and the conductance of the alloy was much higher.
Save electricity by 3,000,000 KWH per year