Optimization of flotation plant performance using micro-price analysis

"Froth flotation circuits are widely used throughout the mineral processing industry for the concentration of fine particles. It had been estimated that nearly 2 Gt (2.2 billion st) of ores are treated globally each year using this versatile separation process (Fuerstenau, 1999). Due to the high feed tonnage, a large economic incentive exists for flotation plant operators to optimize the performance of this critical processing circuit. Prior to the 1950s, the optimization of flotation circuits primarily involved the manual adjustment of operating variables based on visual observation and sporadic sampling (Lynch, Harbort and Nelson, 2010). During the next several decades, however, the advent of improved on-line instrumentation coupled with the rapid development of high-speed microprocessors allowed the industry to develop more advanced control systems that could quickly adjust operational parameters in response to disturbances in ore characteristics and plant flows. These early online controls typically involved feedback loops designed to stabilize cell performance through the automatic manipulation of process variables such as pulp levels, gas rates and chemical dosages. The set-points for these controllers were often set by plant operators based on one or more operator objectives, such as maximizing throughput, recovery or grade. More recently, these regulatory controllers have been interfaced with high-level data display and management systems that provide operators with the ability to better optimize flotation plant performance based on mathematical models and rule-based algorithms (Greet, 2010).While tremendous strides have been made during the past several decades, even the most modern flotation plants are still faced with the long-standing question: “What are we trying to optimize?” From an operator perspective, some common responses include “Maximize tonnage of saleable concentrate” or “Maximize metals recovery at a customer-specified market grade.” These replies may or may not reflect those of management, who often focus on economic goals such as “Maximize profits for each operational unit” or “Maximize rate of return for all corporate assets.” Unfortunately, these high-level objective statements do not always translate well into practice due to the disconnect between plant operating protocols and product value. Attempts have been made by industry to overcome this issue using various types of optimization routines such as linear programming or other optimization engines. While this approach does have some merit, these blind routines require considerable input data and interpretational expertise and provide little or no fundamental insight that can be used to drive improvements in concentrator performance."