Abstract
This paper presents the design and development of a fuzzy peak current controlled (FPCC) single-stage single-phase nonisolated AC/DC high-power factor LED drive. The proposed controller includes a fuzzy logic controller (FLC) in the loop voltage and a peak current controller in the loop current for an integrated nonisolated LED driver to attain a high-power factor (PF). The proposed control avoids complexities related to nonlinearities of the converter. The control action is initially derived from a group of rules written in accordance with experience and intuitive reasoning. The proposed technique is realized using a DSP processor (TI-TMS320F2812), which is capable of executing a high number of instructions in one cycle. A 70 W, 350 mA LED driver operating with an input of 90 V-230 V, 50 Hz was designed and implemented using MATLAB/Simulink. The results of the driver are in accordance with international regulations. The steady-state and transient responses are validated experimentally.
1. Introduction
An LED is a semiconductor unit that emits visible light when a current is passed through it. Lamps have good efficiency, are mercury free, and are long lasting [1, 2]. An LED provides control over light distribution with lenses or small reflectors and affords adaptability in the design of lighting apparatuses [3]. A block diagram of driver circuit is shown in Figure 1. In the case of a sinusoidal supply voltage, the driver must draw a sinusoidal current from the mains. To achieve this, an active filter is used along with a switched-mode power supply in one of two designs, namely, two-stage [4–7] and single-stage [8–12] power factor correction (PFC) topologies.
The two-stage power factor correction approach utilizes an input current shaping converter ahead of the DC/DC converter. Both of the converters are controlled individually to accomplish PFC and fast voltage regulation. This technique is known for its superior performance, such as high-power factor and low input current harmonics, which is achieved at the cost of additional circuitry (including control circuitry). This might not be justified for low-power applications. Subsequently, a one-stage design was proposed by using the concept of integration [13, 14]. In [15], an integrated buck-boost driver with classical controllers is examined for LED applications. It consists of a single control switch operated in buck and boost modes. The main drawback of the topology is a high value of storage capacitance and a voltage ripple. To minimize the capacitance, Soares et al. discussed a slope compensation technique in [16]. However, the ripple content is still increased in the voltage and current responses.
A driver may suffer damage if the current is higher than the rated current. In order to restrict the current, a closed-loop system control is important. Severe working conditions in terms of temperature and current density certainly influence the efficiency of LEDs [17]. The LED is used not only for lighting but also for several other applications such as computer or scoreboard displays. For such applications, the LED needs a driver that can control the voltage for all of the lamps. A fast settling time is also necessary for a scoreboard, due to high lighting motion from one LED to another within 5 to 100 ms. The color changing requires rapid responses in the current value and a stable voltage; therefore, proper control of the current and voltage distributions is required.
In recent years, there has been an increasing interest in control strategies to enhance the dynamic behavior of AC/DC converters. Classical (PI, sliding mode) and intelligent (fuzzy, neurofuzzy) control techniques are used in the literature to improve the performance of drivers. In [18–21], the authors considered converters possessing linear characteristics, but this strategy is not appropriate for some operating conditions. Moreover, there exists uncertainty due to source variations and load changes. Fuzzy logic control (FLC) is advantageous in this regard because it can accommodate parameter variations in the system and can therefore incorporate modeling uncertainty, neglected components, and nonlinearities [22–24]. Knowledge-based FLC attempts to code human knowledge, experience, and acumen to be able to make informed choices about the performance of the system. Knowledge-based FLC is a fuzzy rule base with suitable choices to control the activity of the plant under examination without human intervention [25–29]. In [30], the authors were able to achieve a PF of 0.98 and 13% THD using a digital fuzzy controller.
All passive components constituting the converters can be described by linear characteristics. This assumption is obviously not valid under all operational conditions, particularly since the inductance coefficients vary with the current. Moreover, there exists modeling uncertainty in the presence of input voltage fluctuations, load changes, and values of the component coefficients. Hence, in this work, we propose and develop a FLC in-the-loop voltage and fixed frequency peak current controller in the loop current for an integrated nonisolated LED driver. Unlike conventional controllers, a fuzzy peak current controller (FPCC) is used. To maintain performance under coil magnetic saturation and model uncertainty, FLC is selected because it can deal with the aforementioned uncertainty and variations, as well as neglected elements and nonlinearities. In this work, we consider source variations and load changes. The proposed controller attains a high-power factor (PF) with a low %THD, comparing favourably with conventional methods. Furthermore, the voltage loop driver feeds a regulated voltage to the LED panel. The DSP is a better option to realize the proposed FPCC due to its fast execution. Moreover, the driver is low cost, so that the controller and processor are affordable.
The objectives of this work are as follows: (1)To design and develop an intelligent controller that can act effectively during source variations and load changes(2)To achieve a high-power factor with a driver circuit during source and load variations in order to comply with international standards(3)To attain a low %THD for any source and load variations in order to comply with international standards
2. Analysis of LED Driver
The integrated nonisolated LED driver is depicted in Figure 2, and the corresponding parameters are explained as follows.
2.1. Mains Current and Power
The current through during is the line current (), where is duty cycle and is switching time period. The value of the mains current averaged at line frequency can be obtained using where is the instantaneous peak current, is the switching frequency, and is the peak line voltage. From Equation (1), the line current is sinusoidal in shape, and subsequently, the power factor (PF) is unity by filtering electromagnetic interference (EMI) with a filter. The mean input power () is given by
2.2. Load and Bus Voltages
The load voltage () for the driver can be obtained by equating the supply and load powers. The load power is given as , where is the static resistance of the LED, given by
in which and are the resistance and voltage of the LED, respectively. By assuming 100% efficiency, and . The bus voltage can be calculated from the voltage ratio as follows:
To ensure a high PF, the converter must be operated in discontinuous conduction mode (DCM) and the critical limit for the duty cycle () can be obtained from the voltage transformation via .
2.3. Reactive Components
The value of can be obtained from by assuming there are no losses. The bus capacitance for a known peak-to-peak voltage ripple is
The inductance and capacitance and are obtained from where is the current ripple and is the output voltage ripple.
3. Proposed Fuzzy Controlled LED Driver
The driver is nonlinear and time varying from a control system point of view: (1)The driver circuit possesses nonlinear characteristics owing to the voltage and power of the LEDs. At a nominal working voltage, the characteristics experience a large gradient. This means that a small change in voltage can lead to a significant change of current and therefore to a considerable change in the emitted light. Moreover, when devices are connected to nonlinear loads, there could be several nonlinear relations between the system variables(2)The system is time varying because the parameters of the system change with the temperature and magnetic saturation
A fuzzy approach offers the possibility to model a nonlinear system using knowledge of many non-well-defined relations among the variables of the system and allows for the design of a controller that adapts itself to several working conditions. Thus, fuzzy logic seems a suitable solution both to model and to control drives. The circuit diagram of the driver with a compensation ramp is shown in Figure 3(a). The driver is controlled with a fuzzy voltage loop by a Mamdani type fuzzy inference system, and the outer current loop consists of a peak current control with slope compensation.
(a)
(b)
The error between the reference voltage and converter output voltage and its variation are the inputs to the FLC (see Equations (7) and (8)). The FPCC and its components are shown in Figure 3(b). The output voltage is tracked and compared to the reference signal, and the generated error and change in error are given as inputs to the FLC to obtain a suitable signal to the current controller loop. In the current controller loop, three signals are used in order to generate the pulses to the switches: the inductor signal, the FLC signal, and the compensating ramp signal. Although not considered here, parameter and model uncertainties can be incorporated directly as fuzzy numbers in the fuzzy set theory.
The FLC consists of the following components: a fuzzifier converts crisp data into fuzzy data; a knowledge base contains a data base and a fuzzy rule set; an inference engine deduces the fuzzy control action from the knowledge base and simulates a human decision process; defuzzification yields a crisp value from the fuzzy action.
The FPCC, with the help of the knowledge and experience of an expert, improves the performance of the system. Applying intelligent control techniques for issues identified with uncertainty makes use of IF-THEN rules to define the relation between the inputs and the outputs. Simulation is carried out to check the performance of the design, and if the design is not considered satisfactory, modifications are made to tune the limits of the controller by changing the fuzzy rules and the procedure is repeated for better outcomes.
Several steps are involved in the design of fuzzy controllers as shown in Figure 4: (i) choosing the fuzzy controller inputs and outputs; (ii) choosing the preprocessing that is needed for the controller inputs, and possibly postprocessing that is needed for the outputs; and (iii) design of the four components of the fuzzy controller: the fuzzifier, the rule base, decision maker and defuzzifier.
3.1. Identification of Inputs and Outputs
The inputs of the FC are the error and the variation of error , which are derived from
where is the extant load voltage, is the reference load, is the voltage error, and subscript “” denotes the value taken at the opening of the th switching cycle. The output of the FC is the duty ratio.
3.2. Fuzzifying the Inputs and Outputs
The world of the dissertation of the inputs is separated into 7 fuzzy sets of triangular shape. Outputs are also mapped into several fuzzy regions of numerous singletons.
3.3. Development of the Rule Base and Inference
The proposed system, having rules, is shown in Table 1. It has seven variables, named NB, NM, NS, Z, PB, PM, and PS. Triangular shapes are used as membership functions, and the center of gravity method is used for the defuzzification process.
4. Design Parameters
The driver circuit was built in the laboratory and tested with a universal input (90 V-230 V) range. The driver was designed to provide a power of 70 W, with a rated lamp current of 350 mA. The remaining component values are shown in Table 2.
5. Results and Discussion
5.1. Simulation Results
The proposed fuzzy peak current controlled integrated nonisolated LED driver was developed in the MATLAB/Simulink software package. Figure 5 shows the simulation model, in which a diode rectifier converts the AC source into DC, which is fed to the integrated converter. The output of the integrated converter is fed to the LED panel.
In order to control the driver unit, a proposed control is presented in a feedback loop. Figure 6(a) shows the simulation results of the input voltage of a current waveform rated at a voltage of 230 V and at full load (70 W). In the simulation, the waveform voltage magnitude is scaled to 10 V by multiplying with the factor 0.0307. This result shows that the input current is in-phase with voltage and current waveform following a sinusoidal shape. The respective current harmonic spectrum is also shown in Figure 6(b), which demonstrates that the harmonic content presented in the current wave is below the international specifications (IEC-61000-3-2). The PF and percentage THD at this load are found to be 0.987 and 3.51%, respectively. Figure 7(a) is the output current of the driver, which is maintained at 350 mA. To evaluate the dynamic performance of the controller, a change of load was made at 0.5 sec as shown in Figure 7(b), with a stepwise change in load from 175 mA to 350 mA (at 0.5 sec). An improved dynamic response is observed.
(a)
(b)
(a)
(b)
From Table 3, the performance of the converter is significantly improved with the universal input (90-230 V). The THD improves from 10.82% to 3.51%. The PF of the driver in the entire range is found to be above 0.96.
It is observed from Table 4 that the power factor is higher at full load conditions and the corresponding THD is lower (0.987 and 3.51% at 70 W).
5.2. Experimental Results
An experimental setup was implemented to validate the simulation results. Microcontrollers based on FLCs for power converters were implemented, but due to the high sampling frequency, control DSPs were instead used. The proposed controller was implemented using a TMS320F2812 DSP, which is a 32-bit processor operating at 150 MHz. The digital converter contains the recompense network, fault amp, slant recompense, and PWM generator, which work in a discrete time domain. In this paper, we used a Texas Instruments TMS320F2812 DSP processor. The duty is to initially set to 100% but tripped using the cycle-by-cycle trip feature of the processor. The yield voltage of the converter is connected to a resistive “inspecting divider” arranged such that it is associated with DAC. The voltage is tested and changed over to advanced esteem. A digital reference (REF) is subtracted from the digital value, and the error value is used as an input to the digital controller. This represents the error amplifier and compensation network of the analog equivalent. The yield of the controller is increased by an additional term , which scales the yield of the controller to computerized esteem well-suited for use with the D/A Converter (DAC). The general hardware structure is shown in Figure 8. The output of the converter is sampled by an A/D converter, which gives the digital value to the fuzzy controller. The controller processes the data to produce a PWM signal to the control switch SW1 of the converter.
The line voltage and current waveforms, load current and voltage waveforms, and current harmonic content were observed. The experimental component details are listed in Table 5. Figure 9(a) shows the experimental setup of the proposed FPCC integrated nonisolated LED driver with a controller, while Figure 9(b) is a detailed picture of the controller.
(a)
(b)
The supply current was in-phase with a supply voltage of 230 V (: 300 V/div; : 0.5 A/div). The corresponding measured PF is 0.973, and %THD is 5.67%. The respective harmonic spectra are shown in the left of Figure 10(a), while the load current and load voltage are shown in Figure 10(b). The experimental setup of the driver with an LED lamp (component values as in Table 2) and the lamp under working conditions are shown in Figures 11(a) and 11(b), respectively.
(a)
(b)
(a)
(b)
A comparison of the simulation and experimental results is provided in Table 6, from which it can be concluded that the proposed DSP-based FPCC integrated nonisolated LED driver operating at a low voltage (90 V) gives a low PF and high %THD with respect to the current. The PF increases to 0.987 in the simulations and 0.973 in the experiments at the high voltage (230 V), and the corresponding THD is 3.51% in the simulations and 5.5% in the experiments at a full load of 70 W. From the experimental and simulation results, it is clear that the proposed FPCC integrated isolated LED driver implemented through DSP is suitable for low, medium, and high voltage operation.
Table 7 shows a comparison of the existing topologies with the proposed topology. The power factor is 0.987, close to unity, and the THD is 3.51%, relatively low compared to existing topologies. Furthermore, Table 8 shows a comparison of existing controllers with the proposed controller. It is observed that the proposed controller exhibits better performance.
6. Conclusions
This paper presents a fuzzy integrated nonisolated LED driver designed to achieve a high-power factor (PF). It consists of fuzzy control in a voltage loop and peak current control in a current loop. The voltage loop regulates the output voltage, and the current loop improves the power factor. Fuzzy logic is used in a feedback path and linear programming rule to govern the magnitude of the reference current during this proposed technique in order to regulate the switch’s duty cycle for shaping the input current. The proposed method avoids complexity related to nonlinearity of switching converters and is ready to react quickly to load changes, so controller processing leads to better dynamic performance. The experimental results at steady state showed that the proposed control strategy is capable of producing a power factor value of almost unity under a wide range of supply voltage and load power conditions. The experimental results show that the PF of the driver is 0.973 and the %THD is 5.67% at full load, and therefore, the efficiency of the driver is 86.5%, which meets international regulations (IEC-61000-3-2).
Abbreviations
| PF: | Power factor |
| LED: | Light-emitting diode |
| FLC: | Fuzzy logic controller |
| DSP: | Digital signal processor |
| FPCC: | Fuzzy peak current controlled |
| THD: | Total harmonic distortion. |
Data Availability
Data is available upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This project was partially supported by the National Key Research and Development Program of China (Grant No. 2017YFB0701700). This work was supported by the National University Malaysia Pahang (UMP) (Grant No. PGRS200321), and Mr. A.S. Veerendra is working as a research scholar under UMP’s Doctoral Research Scheme (DRS).