Article Preview
TopIntroduction
Conventional controllers are the most widely used controllers in industry to control the dynamics of any process (Shinskey, 1998; Basilio, & Matos, 2002). But certain short comings of conventional controllers restrict its use in critical applications. Conventional PID controllers fail to deliver good results in the system where certain system parameters are not known or change over time, in such cases adaptive control techniques may be used as an alternative (Benjelloun, Mechlih, & Boukas, 1993; Tsai, Huang, Chen, & Hwang, 2004). One particular way of handling this problem is the use of model reference adaptive controller (MRAC). Here, a reference process model whose dynamics in response to a reference input should be followed by the actual process (Astrom, & Bjorn, 2001). MRAC is a direct tuning strategy with adjusting mechanism that enables it to achieve desired result even in system parameter variations (Cirrincione, & Pucci, 2005; Kersting, & Martin, 2017).
In MRAC, the process output always tracks the response of a reference model and accordingly it adjusts the controller parameters to provide desired performance. However, often it is observed that MRAC fails to deliver good results for the higher order, complex and nonlinear systems as it is unable to change adaptation gain automatically according to the process requirement (Mushiri, Mahachil, & Mbohwa, 2017). In such cases, fuzzy or other soft-computing techniques may be used as an alternative (Naskar, & Pal, 2017). In the process industry, generally a skilled human operator always tries to manipulate the process input, usually by adjusting the controller gain based on the current process states to get the process ‘optimally’ controlled. The proper tuning of the output control action / adaptation gain is very important, as it is equivalent to the controller gain. This output gain has been given the highest priority because of its strong influence on the system performance and stability of the system (Basilio, & Matos, 2002).
The fuzzy logic system is a soft computing tool that can be used to powerfully get things done. It deals with uncertainty and provides an inference structure that enables appropriate human reasoning capabilities. The human brain interprets imprecise and incomplete sensory information provided by perceptive organs. The fuzzy set theory provides a systematic calculus to deal with such information linguistically, and it performs numerical computation by using linguistic labels stipulated by membership functions. Moreover, a selection of fuzzy if- then rules forms the key component of a fuzzy inference system that can effectively model human expertise in a specific application (Naskar, & Pal, 2017). Fuzzy logic reflects how people think and it attempts to model our sense of words, our decision making and our common sense. As a result, it is leading to new and more human intelligent systems. Fuzzy logic unlike Boolean or crisp logic, deals with problems that have vagueness or imprecision. It deals with uncertainty in engineering by attaching degrees of uncertainty to the answer to a logical question. The fuzzy knowledge-based system can track the system parameter variations easily and it able to deal with the problem of uneven load, inertia and set point variations (Pal, & Mudi, 2012; Koo, 2001; Abid, Chtourou, & Toumi, 2009).
The paper suggested a modified version of MRAC using fuzzy logic which is capable to adjust the system parameter variations by changing the adaptive gain as required. The design fuzzy based MRAC is expected to provide improved performance as it includes the process dynamics. The developed controller is further modified using a set-point adjustment scheme and subsequently its performance is investigated in different processes with system parameter variations.
In this section, the proposed work is presented through a flowchart representation as follows: