Automatic Control of Flotation Columns - Advanced Control and Supervision of Mineral Processing Plants - page 250

reach the top of the vessel as a stable froth, where they are removed, usually consti-
tuting the valuable product (concentrate). The non-floatable particles (hydrophilic),
which do not attach to the bubbles, settle and leave the vessel through the bottom
port as non-valuable product called tailings. This separation scheme is called direct
flotation. In some rare cases, the tailings stream may constitute the valuable product,
whereas the concentrate is the discarded product (inverse flotation). In some flota-
tion devices ( i.e. , columns) the froth is sprayed with clean water, which, if adequate
conditions prevail, moves downwards (bias flow) cleaning the froth from undesired
entrained particles.
Despite the highly sophisticated devices (instrumentation and data acquisition)
often found in mineral processing plants, the huge amount of data usually accumu-
lated in their historians is only partially used. Their utilization for automatic control
purposes is rather limited, especially when compared with other types of process
plants such as chemical or petrochemical industries
For instance, typical regulatory control strategies implemented in industrial
columns are basically limited to: (a) controlling the froth depth by manipulating
the tailing flow rate (to indirectly influence the concentrate recovery); (b) manipu-
lating the gas rate to attain the desired recovery. Since the bias rate is not currently
measured, the cleaning action of the froth zone is, at most, manually controlled by
manipulating the wash water rate. Besides gas rate, none of the gas dispersion prop-
erties, nor the bias rate are used for automatic control purposes.
However, some control strategies based on gas dispersion properties as secondary
variables, such as gas hold-up and gasrate, have been proposed by academic re-
searchers. Recent studies have shown that another gas dispersion property, called
“bubble surface area flux ( S b )”, a combination of gas rate and some average bubble
size, is strongly related to flotation performance. Thus, an extended control approach
considering S b is worth considering. This means controlling bubble size and gas rate
to get a desired S b value. All these matters will be discussed in Sections 6.4 and 6.5.
The metallurgical performance of a flotation column is determined by the valuable-
mineral concentrate grade and recovery, often called primary variables, as they de-
termine the revenue the plant is able to get from the sale of its concentrate. Whereas
the grade can be measured on-line using an X-ray on-stream analyzer (OSA), the
recovery must be estimated from steady-state mass balance, which usually intro-
duces some sort of estimation error, therefore strongly limiting its use for control
purposes. The long sampling times and cost of these OSA devices, usually multi-
plexed, also favour the use of a hierarchical control. In such a scheme, secondary
variables, ideally strongly correlated with the primary ones and measured or at least
precisely estimated, are controlled. In the specific case of the flotation column, the
interface position (froth depth), bias rate and gas hold-up or bubble surface area flux
are convenient secondary variables. Ideally, the controller must adequately deal with
the interactions between the variables and take into account various operational con-
straints. The set points of the secondary variable controllers are optimally calculated
by a supervisory control strategy which optimizes an economic criterion. This cost
function is based on the relationships between the secondary and primary variables,