Minimum Variance Benchmarking
Developed in Canada in the early 1990s, this method compares the performance of a controller against that of a minimum variance controller. It is simple to use and requires the least amount of plant data and information.
When to Use MV
- When only a rough guideline is needed and simplicity is a premium.
- When variations of control signal and actuator movements are not important.
- When very tight control of the process output is demanded without concerns for power or energy used.
When Not to Use MV
- When there are tight limits on actuator movements and control signal variations.
- When the controller is in practice restricted to being low order and of a particular form.
- When the cost involved is not simply an output variance or tracking error.