DC/DC converter control using suggested artificial intelligent controllers

ABSTRACT


INTRODUCTION
In view of the way that the energy of the solar is sustainable and condition well-disposed vitality source, the use of solar energy is quickly spreading [1]- [3]. One of the important parts in solar system is the DC/DC converters which are the interface between the source and the load. The usage of artificial intelligence in tackling and control different problems has been quickly looked into and examined in the course of recent decades [4]- [6]. The control process of DC/DC converter is considered as pivotal function in the conversion of power [7]. Where these converters have many important applications in our life such as aircrafts, computer power supplies, switching mode regulators, and solar chargers [7]. Martinez et al. [8] proposed output regulation for the DC/DC buck converter using fuzzy logic. They used type-1 and type-2 fuzzy logic. They presented simulation and experimentation for the voltage controller. In the second paper [9], the authors suggested fuzzy logic with buck converter. They showed the role of this controller and its applications in the control algorithm. They evaluated the buck under different loads and power quality indices like total harmonic distortion and power factor. Maji et al. [10], the authors suggested control method for buck converter utilizing fuzzy logic and robust controller. They used power metal oxide semiconductor field effect transistor (MOSFET) as a switch for the buck. On the other hands, the fuzzy and robust controller was based on pulse width modulation (PWM) to control the switch. The authors designed the fuzzy logic to enhance the converter by MATLAB/Simulink. However, when used the robust controller, the buck output  [11], have been controlled the non-isolated buck converter with proportional-integral-derivative (PID) controller. Where this controller measures the difference between process output and the set point. MATLAB simulation has been used to evaluate the performance of the buck. Patil and Jagtap [7], construct a system to control the buck converter. Where the authors controlled the output of the buck and keep constant without changing the load and the input. They utilize the adaptive fuzzy controller which simulated in MATLAB. They supposed that their control system will increase the efficiency and power of converter. Poodeh et al. [12], the authors introduce different fuzzy logic control law algorithms to control the DC/DC converter. They suggested these controllers to eliminate the overshoot and decrease the settling time. Fuzzy PI, fuzzy PD, and fuzzy PID are proposed for DC/DC converter. The work was done using MATLAB software. Corcau and Dinca [13], give a DC/DC buck converter controlled from 24 V to 12 V with 500 W power. The whole system is constructed in SIMPOWERSYS in MATLAB. They used fuzzy with PI configuration. In that case, the fuzzy controller is built for two controllers in parallel, the first one for proportional component and the second for integrative component. As mention they suggested this structure because of the less inference rules are needed as well as the tuning is easier. Priyadar Shini and Rai [14] designed PID controller for buck boost and Cuk DC/DC converters under the continuous conduction mode. They make the output constant with variations in the input. Safarinejadian and Jafartabar [15], they proposed control system for the output of the buck converter. They designed hybrid fuzzy logic controllers to perform this job. Three methods are used, fuzzy PI, two-level fuzzy hybrid, and fuzzy PD+I. They found that the PD+I have minimum steady state error. While the twolevel is suitable for less settling time. Ugale et al. [16], the authors suggested control for buck-boost DC/DC converter including a power semiconductor device as a switch. They develop a fuzzy logic produced the PWM pulse as a duty cycle of the MOSFET. This required to reach to the desired voltage. Duranay and Guldemir [17] proposed fuzzy control for DC/DC buck converter. Input voltage and output load are changed and the responses of fuzzy are tested in this work.
In this paper various controllers have been presented to control the duty cycle of the DC/DC buck converter for the closed loop system. The rest of the paper is organized as: section 2 presented the overview on buck converter; section 3 proposed the controllers; while the simulation results are in section 4; finally, the conclusion in section 5.

OVERVIEW ON BUCK CONVERTER
The buck is a step down chopper converter. As in Figure 1, the input voltage is higher than that of the output. The switch is used for regulation and keep the desired level for the output. The switching frequency is the rate of change from ON to the OFF for this switch. where: is the duty cycle which is the ratio of the output to the input.
where: ) is the output current The switch is controlled by a signal which will help to control the output voltage [18]. Buck converter can work in two modes of operation, switch ON mode and switch off mode. At the closed position of the switch as in Figure 2, the input voltage will be divided between inductor and the load. Therefore, the energy in the inductor will increase. While at the open position of the switch in Figure 3, the inductor will discharge the energy to the load. Here, the switch is transistor that will be ON or OFF according to the gate signal. ON switch, produces the input voltage which is the sum of the inductor voltage and load voltage. In OFF switch, the inductor voltage is equal to the resistor voltage. Therefore, the inductor is the source at OFF switch. Also, in this converter the capacitor is used to decrease the output voltage ripple as well as the filtering of currents harmonic [19]. In addition to that, the freewheeling diode is utilized to transfer the energy from inductor to the load.

Proportional, integral, and derivative controller
In most projects of process control, PID controller is used due to it is a simple in operation and its rendering is convinced [20]. In order to arrange the performance of this controller, its parameters can be adapted which are proportional, integral, derivative [21]. These parameters are optimized using different methods then, The PID is tuned according to [22]; large positive input to the PID means the should be big, then small, as well as , thus rapid the output of the system; very small input, produces small , large and , in that case the overshoot will be reduced by output and the response will be fast. That is why it should use another intelligent method. The MATLAB simulation of the PID controller is shown in Figure 5.

The proposed fuzzy logic controller
Fuzzy logic controller (FLC), is a controller used in different areas and applications such as the industrial, dish washes, and Televisions. FLC utilized the linguistic form rather than using the mathematical form for knowledge [23]. It is simple and low cost so it widely used especially for nonlinear systems [24]. Fuzzy controller consists of the following parts  The fuzzifier. In the fuzzifier stage, the data of the input will convert to the fuzzy values. Then the second stage is transfer of the fuzzy values to the (IF) and (THEN) rules. In this stage there is also a membership function in the system like Gaussian, Trapezoidal, and triangle. While the decision making will decide the output values. Figure 6 shows the fuzzy logic structure. While the fuzzy editor in MATLAB is shown in Figure 7.

Fuzzy supervisory-PID controller
The fuzzy supervisory-PID (FSC-PID) [25] sort endeavors to give nonlinear activity for the controller utilizing fuzzy reasoning and the parameters of the PID are adjusted dependent on a fuzzy inference system. This controller consists of two parts, fuzzy and PID. In this system, the controller is the PID control, while the role of fuzzy is to control the PID through its parameters. MATLAB simulation for this type is shown in Figure 8.

SIMULATION RESULTS
All the results have been done using laptop Dell with Windows 7 Ultimate operating system ,where the CPU is Intel Core i5 2.5 GHz, RAM is 4 GB. The work is done using MATLAB R2008a. Thus, the buck converter has been simulated using MATLAB as shown in Figure 9.
The values of the component of the DC/DC converter are 0 = 40 Ω, Inductor = 40 , and Capacitor = 4 µ . First, control the buck converter using PID controller as shown in Figure 10. Where = 100, = 10, = 0.01 which are calculated by try and error method for the best values. Second, fuzzy controls the buck converter as shown in Figure 11. In this case triangle membership function is used. Finally, fuzzy supervisory-PID is presented to control the DC/DC converter as shown in Figure 12.  Figure 13 shows with different time instants. Hence, PID controller adjusts the duty cycle of converter. Here, the ripple is high with steady state error which is illustrated in Figure 14. In the fuzzy controller part, as given in Figure 15, the ripple is less than before but with big settling time. In Figure 16 the fuzzy supervisory-PID controls the converter with excellent results and very fast response which outperforms the others in terms of the error. In Figure 17 all controllers have been compared and illustrated in one figure. The graph of is presented in Figure 18. As shown in Figure 19, the controller is still not able to follow the reference purely and need improve by other controllers. In Figure 20, the fuzzy controller track the set voltage, however the settling time is larger compares to the rest controllers. Fuzzy supervisory-PID has good performance if it is compared with the two pervious controllers as in Figure 21. With different steps of , the controllers are compared in the same plot as shown in Figure 22. Finally, table 1 presents the steady state values.

CONCLUSION
In this work, various control techniques have been introduced to control the switching of the buck converter. Even if the controllers are competitive, but FSC-PID controller outperforms them in all its performance. The results are done using MATLAB software. The artificial intelligent control is used as well as PID control. Where the PID controller is used alone, then fuzzy logic controller, after that fuzzy-PID controller is implemented. The set voltage varied from (5, 10, and 15) to test the ability of the controllers to do the job.