Computer Systems and Software
Balancing the scale
A manufacturer of steel turned to statistical analysis software to reduce the scale on rods to acceptable levels. Eston Martz reports.
O
neSteel manufactures and distributes more than 40 000 products to some 30 000 customers in the construction, manufacturing, housing, mining and agricultural industries. The Australia-based company’s products
range from thick beams to thin steel wires. OneSteel discovered its Laverton facility was producing rods with excessive variations in mill scale—the iron oxides formed during processing. While some scale is inevitable, the scale must be removed to prevent wear on expensive equipment, avoid defects in the final product, and limit yield loss. To reduce the scale on rods to acceptable levels, a OneSteel team turned to Minitab Statistical Software to guide their efforts.
The challenge
During processing, OneSteel’s raw metals are shipped between facilities. At the Laverton Rod Mill, steel is heated to temperatures over 1000°C and rolled into rods that receive initial cooling in water boxes. Next, the rods are formed into a pattern of rings, placed on a conveyor to be cooled further, and finally coiled together. High-temperature processing at Laverton caused iron
oxides, or mill scale, to form on the rod’s surface. Scale levels vary depending on the type of rolling mill used, the temperature of the steel, and how long it’s heated. Scale is abrasive and reduces the life of the drawing die which, in turn, can reduce quality. To remove the scale, the Geelong Wire Mill uses reverse bend descaling, a mechanical process of passing the rods over a series of sheaves. This bending removes the brittle scale without harming the flexible steel. However, scale levels on the Laverton rods often exceeded
the desired level for mechanical descaling. The result was decreased yield and costs including:
l Clean up and disposal. l Operation and maintenance of the scale extraction system. l Lower run rates. l Shorter die life. l Increased lubricant usage. l Lower product quality.
In one year, the Geelong Wire Mill spent over $800 000 for housekeeping, lubricant, and dies, so reducing scale on the Laverton rods could potentially create enormous savings for OneSteel. Team members from the rod and wire mills included a Six Sigma Black Belt, technical managers, process engineers, quality professionals, a metallurgist, and equipment operators. For their comprehensive study using the DMAIC method, the team defined defective rods as rods with descalability lower than 95 per cent. Next, they sampled the rods arriving at Geelong from Laverton. The team measured scale weight as a proportion of the
total product (steel rod coils) in kg/tonne. They found that 62 per cent of samples exceeded 7kgs of scale per tonne, with variations up to 3kg per tonne. This results in a significant yield loss since variation of just 1kg per tonne equals $75 000 in yield. Geelong loss estimates were around $220 000 per year. Throughout the project, team members turned to Minitab
to analyse data and to communicate their findings. For example, Minitab’s 2-sample t-test proved the difference in scale weights between grades of steel rods, while a box plot clearly displayed the information. In addition, they used Minitab’s powerful regression analysis
to reveal the relationship between laying head temperature and scale weight, and a scatterplot displayed the relationship. The team also applied Minitab’s Probability Density
Function to actual sampling data and compared it to a desired
26
www.engineerlive.com Fig. 1. Minitab’s powerful graphics made it easy for everyone in the project to see the difference in die wear between the trial run (red) and the
control run (black).
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36