November 2013

Pelletization partners

Vale and universities seek to perfect ball mill optimization

By Alexandra Lopez-Pacheco

Times are tight for the minerals industry, and optimizing equipment to increase efficiency and save energy is becoming central to doing more with less. So says Luís Marcelo Tavares, associate professor of mining and metallurgy at the Universidade Federal do Rio de Janeiro (LTM/UFRJ) in Brazil, who leads the Laboratório de Tecnologia Mineral research group at the university. “Lower grades, larger plants, and the push towards greater sustainability through minimizing energy and water usage are requiring answers from the whole mining community. These challenges call for skills that go from the depth of theoretical to highly practical knowledge. However, looking around the globe, very few groups reach the minimum size and skill base required to make a measurable impact.”

Two years ago, Brazilian mining company Vale, which has a long relationship with the Laboratório de Tecnologia Mineral, set out to tackle this gap, offering some of its employees the opportunity to pursue master’s degrees. Patrícia Faria, master process engineer at Vale S/A was among them. The company contacted Tavares, who seized the opportunity, bringing in his alma mater, the University of Utah. He agreed to be one of Faria’s supervisors as she embarked on her degree, with her research focused on ball mill optimization. “I could recognize that Patrícia combined the right tools for such an important venture: great personal drive, high intellectual capacity and very good people skills. This later proved to be the key ingredients to successfully manage supervisors in two different continents (which she has so far), as well as to negotiate support for her test work, both online and offline, in the pelletizing plant,” says Tavares.

“Pelletizing is becoming increasingly important,” explains Tavares about the relevance of Faria’s research. “Great attention is often devoted in a pelletizing plant to ensuring product quality by controlling and optimizing the formation of green pellets and their induration. In the pellet feed preparation, attention is mostly focused on guaranteeing that grinding generates a consistent product. However, it has been identified that, very often, such milling circuits are not optimized, thus missing out on a very good opportunity to reduce costs and increase production capacity.”

Faria set out to identify the best means to optimize an existing ball mill in order to save energy, increase production and improve iron ore pellet quality. She spent much of last year in Brazil, reviewing the literature and seeking the best optimization methodology. “In the past, most often industry lab work for optimization was based directly or indirectly on the empirical Bond Work Index,” says Faria, who has been employed with Vale since 2010. “This approach was used extensively, but, according to some researchers, this methodology is useful for new design and not for optimization of operating mills.”

Breaking Bond

Faria’s research led her to conclude the Bond Work Index did not take into account two fundamental elements: mill transport and cyclone size classification. “Bond basically assumes that the breakage kinetics transport in the mill and classification of the processes is characterized by a single parameter, the [Bond] Work Index,” says Faria.

Instead, she found the Population Balance Model is more accurate. “It makes it possible to simulate the industrial mill in a laboratory mill with great accuracy,” notes Faria. “And this is important for optimization because it allows me to do all the experiments to determine the optimal conditions for the industrial mill in a very small, 10-inch batch mill, in a laboratory. If the results are good, I can then scale up and apply it to the industrial mill. So it is an economical way to simulate the best conditions to apply in the industrial mill without the potential loss of energy and production that would be associated with doing the tests in the industrial mill.”

Tavares says the use of the Population Balance Model to improve grinding is seldom done in the context of grinding pellet feed and certainly not at the level of detail that Faria has pursued. “Among the reasons it is not done as often is the lack of skilled engineers, as well as the need to properly calibrate models at different levels before validating at an industrial scale,” he points out. “Only after that exercise can one confidently implant the technology in the plant and place methods to conduct calibration of the model to account for variations due to ore grindability.”

Large-scale results

Faria spent the last year collecting extensive data from an industrial mill at Vale in Brazil. She measured the fresh feed size distribution and, using the Population Balance Model, predicted the mill size distribution and the cyclone overflow size distribution.

She is now using the results of her research and the data she collected at the University of Utah lab to simulate the potential improvements in an industrial-scale mill. Her goal is to meticulously test each parameter, such as solids concentration, mill speed, optimal top ball size and ball load, and a series of other possible variables to determine the exact operational conditions needed to optimize the industrial mill.

In January, Faria will return to Brazil and apply her findings. “Validation of simulations using good data from industrial plants is a key prerequisite in her study,” says Tavares. “As such, collecting good data is often not a trivial task, since very few plants appreciate the effort and care required to conduct a good industrial survey. This will be particularly critical in the follow-up of her work back in Brazil, when she will work on the application of more advanced models of comminution to the plant.”

Ultimately, the research has already been successful in the sense that it is affording Faria the opportunity to profoundly augment her knowledge and understanding. “I’m learning what is behind the theory and methodology,” she says. “There is software that is widely available for optimization. I think it has its limitations, but more importantly if I use the software at my work for the optimization, I don’t know the theory behind it. For me, this is far better and it is knowledge I can build on and share.”

Post a comment


PDF Version