Control engineers typically use either PID controllers or model-based control design tools to design control systems. PID often is insufficient to control complex systems, and model-based control methods depend on good process models that are difficult to develop and maintain. In addition, designing a model-based control system requires high-level expertise and is time-consuming.
To help automation engineers solve various tough control problems, CyboSoft released a major revision to its MFA Control Toolbox for MATLAB Software. Model-Free Adaptive (MFA) control, as its name suggests, is an adaptive and robust control method that does not require process models. Based on the core MFA technology, CyboSoft developed a set of general-purpose MFA controllers, each of which solves a tough control problem.
This product can help bridge the gap between the academic and industrial world because its the only off-line control simulation software for which a number of field-proven, real-time control products of the same family are available, says CyboSoft CEO Dr. George Cheng. They include many MFA control products offered by CyboSoft and various MFA-embedded or integrated products and systems offered by vendors in the process control, machine control and building control markets.
The MFA Control Toolbox for MATLAB software includes a set of embedded MFA controller S-functions that work as control functional blocks in the Simulink environment. Anyone familiar with Simulink can configure and use the MFA control blocks to perform off-line simulations. It is a useful tool for research, development, design, teaching and lab testing.
MATLAB and Simulink are the most popular simulation tools, says CyboSoft director of engineering Steve Mulkey. Using the MFA Control Toolbox, extensive off-line simulations can be run to prove the concept. Once the simulation is successful, the same MFA controllers can be selected and launched when using the real-time MFA control products in the real application. This, adds Mulkey, allows tremendous savings in research and development time and equipment investment, as well as minimized development risks and time to market.
One of the largest steel companies in the world is using the MFA Control Toolbox for MATLAB software in many applications, says Cheng. Engineers first run MFA control simulations against the MATLAB process models they have already built. They then apply the same control strategy and the MFA controllers in the real plant. The simulation results prove the concept and can alleviate any concerns by the management and plant users.
In this new version, says Mulkey, a flexible demo system is implemented to let the user compare the performance of MFA versus PID, where different types of MFA controllers are used to meet various control challenges.
The MFA Control Toolbox for MATLAB software runs in Windows operating system. The Standard Edition includes six of the most popular MFA controllers for processes characterized as single-input/single-output, nonlinear, anti-delay, robust, feed-forward or pH.
The Professional Edition adds six additional special MFA controllers for the following processes: multiple-input/multiple-output, flex-phase, anti-inverse, time-varying, anti-delay pH and XRT (exothermal reaction run-away).