ABB Ability™ Expert Optimizer
Several challenges may exist when controlling a grinding process. There is usually a strong coupling between the variables, large time delays, uncontrollable disturbances, nonlinear process behavior and a lack of instrumentation. The need for advanced control techniques are therefore clear. ABB’s Expert Optimizer is geared to tackle these problems with an extensive set of tools in its arsenal.
ABB Ability™ Expert Optimizer constantly looks to minimize process variability and drive the process to an optimum, defined by the user. The tools available in the EO toolbox include: Model Predictive Control, Fuzzy Logic, Rule Blocks & Logic as well as the ability to develop Soft Sensors and Inferentials.
Figure 4 Expert Optimizer UI
For the problem at hand, a Model Predictive Controller (MPC) solution was developed. With the availability of the digital twin platform the MPC could be rigorously tested, adapted, and tuned before being subjected to a benchmarking exercise. A comprehensive number of simulations were run in the simulation environment to test the performance of the control strategy. The MPC solution in EO was then compared to the current control strategy onsite. The main results from this exercise are listed below.
MPC vs Existing control strategy
- Improved throughput – Ranged from 1% to 3%
- Improved maximization of torque – Ranged from 1% to 7%
- Improved disturbance rejection – Ranged from 0.5% to 3.5%
- Improved response time and circuit stability – Reduced STD > 5%
General advantages of MPC
- Multivariable control
- Constraint aware control
- Optimization of key process variables
- Maximization of torque
- Maximization of feed
- Optimize grind – by minimizing speed
- Uses the RPM in the entire load-range
- Better at handling multiple conditions
Further work could see the inclusion of our future solution ABB Ability™ Cascade Monitoring (for more information please contact us) into the Expert Optimizer solution strategy. ABB Ability™ Cascade Monitoring is a device that attaches to the shell of the mill to transmit milling features such as toe angle and strike amplitude wirelessly to a base station for consumption in the control system and Expert Optimizer. Benefits here could be an increase in grinding efficiency by reduced variation around milling features and possibly reduced liner wear due to minimizing liner strikes above certain angles specified by operators.