Cost Reduction of Cement Manufacture by Technique

An elevator conveyor feeds the output from the ball type mill to a high efficiency separator that extracts a product cut on the basis of particle size; oversized material is recycled to the mill. Prior to the installation of automated control, the performance of the circuit was maintained by manually manipulating the feed rate, separator speed and separator draft, since these three parameters have a strong impact on process behavior. Routinely-measured plant variables and regular off-line Blaine analysis provided data on which to base remedial action.

The analysis is based on determining the pressure drop across a packed bed and can only be executed off-line. At Vulcan Materials, sampling and associated measurement is carried out every two hours to check product quality. The data can be used for process control but clearly do not provide real-time insight into the process.

Recognizing the limitations of a manual approach, the operations team decided to invest in an automated control solution to optimize operation, focusing on energy consumption and product quality. Vulcan Materials selected a control package from Pavilion Technologies specifically designed for optimizing cement finishing circuits.

At the heart of the Pavilion Cement Grinding Application solution is a multivariate process model that predicts performance from a range of inputs. Automatic manipulation of controlled variables on the basis of these predictions keeps the plant within a defined operating window targeted to optimization goals. The objective is to maintain product quality (Blaine and 325 meshes) within specifications and to reduce variability and improve operational stability. Maximise fresh clinker feed rate subject to equipment constraints.

The specific details of the Vulcan Materials circuit were captured by developing a control scheme uniquely tailored to the plant, which was configured via process testing. The configuration procedure defines the mathematical relationship between different parameters for the individual unit, producing an accurate model for dynamic and continuous optimisation. When the plant is operational the model runs in real time, driving the system towards an optimal operating point within the defined constraints.

A feedback loop measures process response and compares actual response to desired values. A steady state optimiser calculates the correct changes to manipulated variables in order to drive controlled variables to target. An integrated dynamic controller then projects these into the future, developing a series of control actions for each manipulated variable and predicting the effect on the controlled variables in order to meet multiple targets simultaneously. The purpose is to minimise a certain objective function such that the predicted process outputs are as close as possible to the desired reference trajectories. Furthermore, results from an in-house lab study showed that one-day strength correlated much more closely with certain particle size parameters produced by Hongxing and professional  than it did with Blaine. Moving away from Blaine to on-line particle size measurement seemed the next logical step.