Model-free adaptive chaos control for the boost converter

Abstract

In order to solve the dynamic analysis and chaotic motion control problems of the current-mode-controlled Boost converter system, a state equation model in the continuous current mode is established. The evolutionary path of the periodic motion of the system through the doubling bifurcation to the chaotic motion is analyzed by the bifurcation diagram, Poincaré sections and phase portraits. Then, according to the fundamental principle of the resonant parametric perturbation (RPP) method, a controller is designed based on the model-free adaptive control (MFAC) method to output the perturbation amplitude, and the disturbance quantity is calculated according to the perturbation amplitude. By adding a small amplitude disturbance to the controllable parameter of the controlled system, the chaotic motion of the system is guided to expected periodic orbit. The advantage of this method is that no precise mathematical model (PMM) of the system is required. At the same time, the design of the controller is simple and easy to implement.

Keywords

Boost converter chaos control model-free adaptive control resonant parametric perturbation

Introduction

Since the meteorologist Lorenz proposed chaos theory, researchers have discovered chaos in many engineering fields.¹ As the core part of switching power supply, DC-DC converter is widely used in instruments, electric vehicles, green energy and other systems.^2–4 In practical operation, DC-DC converter will produce complex and abundant nonlinear physical phenomena (such as bifurcation, chaos, etc.),^5,6 which will not only greatly affect the stability of the converter, but also cause the converter system failure.⁷

In order to suppress the chaotic motion of DC-DC converter system, researchers have made active exploration. Ito et al.⁸ adopted the strategy of perturbation pulse-frequency modulation to realize chaotic motion control of Buck converter. Jia⁹ researched the bifurcation and chaotic motion generated by the voltage-mode-controlled Buck converter with the change of input voltage. She proposed a self-feedback chaos control method, which successfully stabilized the chaotic motion of Buck converter to period-1 orbit. In view of bifurcations and chaos in Buck converter system, Luo et al.¹⁰ adopted the periodic pulse control method to adjust the output pulse width of the comparison amplifier and controlled the system to the desired periodic motion. Zhang et al.¹¹ adopted sinusoidal voltage compensation strategy based on the monodromy matrix theory to reduce the duty cycle of the system. The chaos control of V² controlled Buck converter system was realized. El Aroudi et al.¹² extracted control signals from the errors between the inductive current and the constant parameters of the converter system, proposed a new self-compensation technique, and realized the chaotic motion control of the DC-DC converter system. Morcillo et al.¹³ analyzed the evolutionary path of the DC-DC converter system leading to chaos through period-doubling bifurcation with the change of circuit parameters. An adaptive ramp control strategy with self-tuning offset and amplitude was proposed to realize the chaotic motion control of the system within a wide range of circuit parameters. Fu et al.¹⁴ used Integral Absolute Error (IAE) as the control performance indicator, and introduced Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) respectively to find the optimal value of PID controller parameters, and the optimized PID controller was applied to realize chaos control of the voltage-mode-controlled Buck converter system. Wu et al.¹⁵ explored the bifurcation behavior and chaotic motion of the Buck-Boost converter system. Based on the exponential delayed feedback control method, the feedback adjustment function in the parameter feedback control law was subtracted from the reference current. The feedback intensity was adjusted by setting the adjustment coefficient in the feedback adjustment function. The improved control strategy was used to control the system to stable periodic orbit. Deivasundari et al.¹⁶ uncovered the chaotic phenomena of a master-slave controlled parallel Buck converter system. The system was controlled to regular periodic orbit successfully by adding an optimal phase-shifted sinusoidal disturbance to the reference voltage. Bao et al.¹⁷ controlled the chaotic motion of the system to a stable periodic motion by introducing a suitable ramp compensation current into the current-mode-controlled Boost circuit. Luo et al.¹⁸ studied the voltage-mode-controlled Buck converter system and used OGY method to suppress the chaotic motion of the system. Lu et al.¹⁹ adopted the time-delay feedback chaos control method to realize the chaos control of the current-mode-controlled Boost converter system. Yang et al.²⁰ studied the dynamics and chaotic motion of the switched-inductor Buck-Boost converter system with the memristive load, and realized chaos control by adopting resonant parametric perturbation (RPP) method. The principle of RPP chaos control method is to change the chaotic motion state of the system into regular motion state by applying a specific frequency of micro-perturbation to the circuit parameters.²¹

Over the years, researchers have made some achievements in the study of chaos control of DC-DC converter systems through continuous efforts. However, most of the existing studies rely on the precise mathematical model (PMM) of the controlled system to design the controller, which will not be applicable or cannot obtain satisfactory control effect when the system is difficult to accurately model. For this reason, researchers have explored the control methods that do not depend on the precise model of the controlled system. For voltage-mode-controlled Buck converter system, Huang²² used Radial Basis Function Neural Network (RBFNN) to learn dynamic characteristics of chaotic system of Buck converter, and adopted the trained RBFNN model to realize chaos control. Ling²³ took current-mode-controlled Buck-Boost converter system as the research object, and introduced GA to optimize the parameters of Back Propagation Neural Network (BPNN) controller, and used the optimized BPNN controller to output the perturbation amplitude to realize chaos control. Although researchers have actively explored the chaos control of DC-DC converter systems without the need for PMM of the system, the current research results are still few and need to be further explored and deepened.

Model-free adaptive control (MFAC) method can effectively solve control problems that are difficult to be modeled accurately because of its advantages of not relying on the mathematical model information of the controlled system and requiring only the input/output (I/O) data of the controlled system.²⁴ Therefore, it is widely used in various fields such as urban road traffic network control,²⁵ network communication delay compensation,²⁶ robot motion control,^27,28 electric vehicle and motor control,^29,30 etc. However, the study of introducing MFAC method into chaos control of DC-DC converter system has not been reported publicly.

Therefore, aiming at the chaos control problem of DC-DC converter system, this paper proposes a model-free adaptive chaos control (MFACC) method which does not depend on the PMM of the system. Based on the idea of combining RPP method and MFAC method, a model-free adaptive chaos controller is designed only using the I/O observation data of the system, so as to output the perturbation amplitude to be determined in the disturbance term. By substituting the perturbation amplitude into the calculation formula of the disturbance term, the disturbance quantity can be calculated. The chaotic motion generated in the system can be guided to a stable periodic orbit by perturbing the controllable circuit parameter. By using MATLAB simulation software and taking the Boost converter system as an example, the results show that the chaos control method proposed in this paper is feasible and effective.

Chaotic motion analysis of the system

In this paper, the chaotic motion control problem of the current-mode-controlled Boost converter is studied. The circuit diagram of the system is shown in Figure 1. In Figure 1, the main circuit of the Boost converter is composed of the input voltage V_in, an inductor L, a MOSFET switch S, a diode D, a capacitor C, a resistor load R and other electronic components. To facilitate the analysis of the working principle of the circuit, assume that all the components in the circuit system shown in Figure 1 are ideal.

Figure 1.

Schematic diagram of the current-mode-controlled Boost converter.

According to the circuit schematic diagram of the system shown in Figure 1, and combined with the theoretical knowledge of related circuit, the working principle is analyzed.³¹ Using the laws of KCL and KVL, the inductive current i_L and the capacitive voltage v_c are determined as the state variables of the system, that is $x = {[i_{L} v_{c}]}^{T}$ , the state equations of the system are

\begin{matrix} \overset{\cdot}{x} = [\begin{matrix} {\overset{\cdot}{i}}_{L} \\ {\overset{\cdot}{v}}_{c} \end{matrix}] = [\begin{matrix} 0 & - \frac{1}{L} \\ \frac{1}{C} & - \frac{1}{RC} \end{matrix}] [\begin{matrix} i_{L} \\ v_{c} \end{matrix}] + [\begin{matrix} 0 & \frac{1}{L} \\ - \frac{1}{C} & 0 \end{matrix}] [\begin{matrix} i_{L} \\ v_{c} \end{matrix}] H + [\begin{matrix} \frac{1}{L} \\ 0 \end{matrix}] V_{in} \end{matrix}

(1)

where H is the signal values of switch S on and off, 1 corresponds to on and 0 corresponds to off.

When some circuit parameters of the current-mode-controlled Boost converter system change, the system may exhibit rich and complex dynamic behaviors. According to the stroboscopic mapping sampling method, the state variables (v_c and i_L) of the system are sampled respectively at the time t = nT (where T is the switching period, n = 1,2,3,…). And take Poincaré cross sections at the time t = nT. The Boost circuit parameters are selected as shown in Table 1. I_ref is used as the bifurcation parameter to conduct numerical simulation of the system dynamics. The bifurcation diagram of the system inductive current i_L changing with I_ref is shown in Figure 2.

Table 1.

The parameter setting of each element in the circuit.

Elements	Parameters	Values
Voltage	V_in	10 V
Inductor	L	1 mH
Capacitor	C	12.5 μF
Resistor	R	20 Ω
Switching period	T	100 μs
Reference current	I_ref	0.5 A–5.5 A

Figure 2.

The bifurcation diagram of i_L with I_ref as the bifurcation parameter.

In Figure 2, when the reference current I_ref changes within a specific range, the system shown in Figure 1 has stable period-1 motion, but with the increase of I_ref, bifurcation will occur and the system will eventually evolve into chaotic motion. There are some periodic windows after the system has gone into chaos, and this is one of the typical characteristics of chaotic motion of nonlinear system.

Figure 3 shows the Poincaré section and phase portrait of period-1 motion of the system. It can be seen from Figure 3 that when I_ref = 0.7 A, a point is shown on the Poincaré section. The phase portrait is a closed curve. The Poincaré section and phase portrait of period-2 motion of the system are shown in Figure 4(a) and (b) respectively. In Figure 4, when I_ref = 1.8 A, there are two points on the Poincaré section, and the phase portrait is partially dented compared with the orbit of period-1. Figure 5 is the Poincaré section and phase portrait of the chaotic motion of the system. Furthermore, from Figure 5, when I_ref = 2.75 A, the Poincaré section presents a point set with fractal structure. The phase portrait has infinitely nested self-similar structures.

Figure 3.

Period-1 motion generated in the system (I_ref = 0.7 A): (a) the Poincaré section and (b) the phase portrait.

Figure 4.

Period-2 motion generated in the system (I_ref = 1.8 A): (a) the Poincaré section and (b) the phase portrait.

Figure 5.

The chaotic motion generated in the system (I_ref = 2.75 A): (a) the Poincaré section and (b) the phase portrait.

MFAC of the chaotic motion

Control strategy

According to the analysis in Section 2, when I_ref changes within a certain range, the system evolves into chaotic motion through period-doubling bifurcation (when I_ref = 2.75 A, the system is chaotic). It can be seen that if the value of I_ref is changed reasonably, the survival conditions of the chaotic motion can be eliminated, so as to achieve the purpose of suppressing the chaotic motion.

Based on this, according to the principle of RPP method, a MFAC strategy is proposed for the chaotic motion of the Boost converter. Firstly, a circuit parameter Q which has a great influence on the system and is easy to change is selected, so that the survival conditions of the chaotic motion can be changed by perturbation of Q, and the system can be guided to expected periodic orbit. Secondly, based on the idea of MFAC, the equivalent dynamic linearization of the Boost converter system is carried out. That is to say, the pseudo partial derivative (PPD) is estimated online by the I/O data of the system to build the dynamic linearized data model of the system. After obtaining the model, the controller is further designed. By constantly extracting the effective information in the I/O data of the system, the controller can accurately learn the nonlinear characteristics of the controlled system, so as to determine the reasonable perturbation amplitude applied to Q.

The structure diagram of the control system designed according to the above control strategy is shown in Figure 6.

Figure 6.

The structure diagram of the control system.

In Figure 6, Q is the circuit parameter that has a great influence on the system and is easy to change; δ_asin(2πf_dt) is the disturbance term (i.e. △Q), δ_a is the perturbation amplitude (i.e. the output of the model-free adaptive controller), f_d is the disturbance frequency (f_d is equal to the switching frequency). v_o is the output voltage of the system; d is the difference between the output voltages of the system after two adjacent iterations, that is, d(k) = |v_o(k)−v_o(k−1)|; d^* is the expected value of the difference between the output voltages of the system after two adjacent iterations; e(k) is the error between the expected difference and the actual difference between the output voltages of the system after two adjacent iterations, that is, e(k) = d^*−d(k).

Controller design

Given the I/O data (discrete time I/O data) of the controlled Boost converter, we can described the system as follows

\begin{matrix} d (k) = f (d (k - 1), . . ., d (k - n_{d} - 1), δ_{a} (k - 1), . . ., \\ δ_{a} (k - n_{δ_{a}} - 1)) \end{matrix}

(2)

where d(k) is the output of the Boost converter at time k, d(k) = |v_o(k)−v_o(k−1)|; δ_a(k) is the input of the Boost converter at time k (i.e. the output of the controller at time k); n_d and n_δa represent the unknown order of the system respectively; f (·) is an unknown nonlinear function.

The Boost converter system shown in equation (2) satisfies three assumptions: the I/O data of the system is measurable and controllable; the partial derivative of f (·) with respect to the control input signal δ_a(k) of the system exists and is continuous; The system is generalized Lipschitz, that is to say, for any value of k and Δδ_a(k) = δ_a(k)−δ_a(k−1) ≠ 0, there is $| Δ d (k + 1) | \leq p | Δ δ_{a} (k) |$ , p is a constant, and $Δ δ_{a} (k) = δ_{a} (k) - δ_{a} (k - 1)$ , $Δ d (k) = d (k) - d (k - 1)$ . Therefore, the nonparametric time-varying dynamic linearization model of the Boost converter system shown in equation (2) can be established as follows

d (k + 1) = d (k) + ϕ (k) Δ δ_{a} (k)

(3)

where d(k) represents the distance between two adjacent points on the Poincaré section, which can reflect the trend of the system approaching stable periodic motion (i.e. the difference between the output voltages v_o of the system after two adjacent iterations, d(k) = |v_o(k)−v_o(k−1)|), where v_o(k) is the value of the output voltage v_o of the system at time k; $ϕ (k)$ represents the PPD and is the characteristic parameter of the nonparametric time-varying dynamic linearization model shown in equation (3) (i.e. the pseudo gradient of d(k + 1) with respect to δ_a(k)). It is related to system I/O data up to the sampling time k. It is $ϕ (k)$ that can achieve self-adaptation.

Based on equation (3), we consider the control input criterion function as follows

J (δ_{a} (k)) = {| d^{*} (k + 1) - d (k + 1) |}^{2} + λ {| δ_{a} (k) - δ_{a} (k - 1) |}^{2}

(4)

where d^* (k + 1) is the expected output (i.e. the expected value of the distance between two adjacent points on the Poincaré section, i.e. the expected difference between the output voltages v_o of the system after two adjacent iterations); λ is the weight factor.

By substituting equation (3) into equation (4), the chaotic motion control algorithm of Boost converter can be obtained as follows according to $\frac{\partial J (δ_{a} (k))}{\partial δ_{a} (k)} = 0$

δ_{a} (k) = δ_{a} (k - 1) + \frac{ρ ϕ (k)}{λ + {| ϕ (k) |}^{2}} [d^{*} (k + 1) - d (k)]

(5)

where ρ is the step sequence.

Since $ϕ (k)$ is an unknown parameter, it needs to be estimated using the I/O data of Boost converter. After obtaining the estimated value $\hat{ϕ} (k)$ , d(k + 1) can be calculated by substituting $\hat{ϕ} (k)$ into equation (3). Therefore, considering the estimation criterion function $J (ϕ (k))$ as defined in equation (6) and the optimal condition shown in equation (7), based on equations (2) and (6), the estimation algorithm of the PPD is obtained as shown in equations (8) and (9).

\begin{matrix} J (ϕ (k)) = {| d (k) - d (k - 1) - ϕ (k) Δ δ_{a} (k - 1) |}^{2} \\ + μ {| ϕ (k) - \hat{ϕ} (k - 1) |}^{2} \end{matrix}

(6)

\frac{\partial J (ϕ (k))}{\partial ϕ (k)} = 0

(7)

\begin{matrix} \hat{ϕ} (k) = \hat{ϕ} (k - 1) + \frac{η Δ δ_{a} (k - 1)}{μ + Δ δ_{a} {(k - 1)}^{2}} \\ [Δ d (k) - \hat{ϕ} (k - 1) Δ δ_{a} (k - 1)] \end{matrix}

(8)

\hat{ϕ} (k) = {\hat{ϕ}}_{1}, if sign (\hat{ϕ} (k)) \neq sign ({\hat{ϕ}}_{1}) or | \hat{ϕ} (k) | \leq ε

(9)

In equation (8), η is the step sequence; μ is the weight coefficient, μ > 0. In equation (9), ${\hat{ϕ}}_{1}$ is the initial value of the PPD; ε is a sufficiently small positive number and is a definite value; sign (·) is the sign function.

In summary, the model-free adaptive chaos controller of the current-mode-controlled Boost converter system can be written as

δ_{a} (k) = δ_{a} (k - 1) + \frac{ρ \hat{ϕ} (k)}{λ + {| \hat{ϕ} (k) |}^{2}} [d^{*} (k + 1) - d (k)]

(10)

As can be seen from equation (10), the design of the model-free adaptive chaos controller does not depend on the model information and structural information of the system, but is only related to the I/O data of the controlled system. In the control process, only one one-dimensional parameter value (i.e. the estimated value of the PPD) waits to be adjusted online.

In summary, the proposed MFACC method of the Boost converter system only uses the I/O data of the controlled system to design the controller. The perturbation amplitude δ_a(k) output by the model-free adaptive controller is used to calculate the disturbance quantity (i.e. δ_asin (2πf_dt)). Then the disturbance quantity is applied to the controllable circuit parameter Q to realize the chaotic motion control of Boost converter. The proposed method is not only simple to calculate but also easy to implement.

Simulation research and analysis

Control effects of the chaotic motion

By means of simulation, the proposed MFACC method is used to control the chaotic motion generated in the Boost converter.

It is found that the chaotic motion can be caused by the changes of the input voltage, the inductance, the capacitance, the resistor, the reference current and other parameters in the circuit system. Therefore, a reasonable controller is designed to output a tiny perturbation quantity imposed on one of the above circuit parameters, so that the survival conditions of the chaotic motion change, so that chaos control can be realized. Due to space limitation, this paper considers that the reference current in circuit parameters has a great influence on the system and is easy to change, so only the reference current is selected as the excitation parameter, and I_ref is replaced by I_ref + δ_a sin(2πf_d t). Therefore, this paper only shows the effect of the chaos control when the perturbation quantity output by the controller is imposed on the reference current of the system.

In order to compare the effects before and after controlling chaotic motion, the controller starts to exert control on the chaotic motion at the 200th iteration of the system. The input of the system (i.e. the output of the controller) δ_a = 0 before the control is applied.

Figure 7 richly shows the control effect after adopting MFACC method. During the simulation, according to a large number of tests, the six initial parameters of the controller are set as follows: ${\hat{ϕ}}_{1}$ = 0.7, λ = 0.00036, ρ = 1.95, μ = 0.91, η = 1, ε = 10⁻⁵.

Figure 7.

The effect of chaotic motion control of the system (I_ref = 1.3369 A). (a) The period-1 orbit of the system after control, (b) the phase portrait of v_c−i_L, (c) the waveform of i_L, and (d) the waveform of v_c.

In Figure 7, the chaotic motion can be stabilized quickly (at k = 250) to the orbit of period-1. At this time, the phase portrait is a closed curve. The waveforms of the inductive current and the capacitive voltage both indicate that the system is moving at period-1.

Figure 8 shows the simulation results of chaos control when the corresponding parameters of the controller are set as: ${\hat{ϕ}}_{1}$ = 0.02, λ = 0.00036, ρ = 1.95, μ = 0.91, η = 1, ε = 10⁻⁵.

Figure 8.

The effect of chaotic motion control of the system (I_ref = 1.9954 A). (a) The period-2 orbit of the system after control, (b) the phase portrait of v_c−i_L, (c) the waveform of i_L, and (d) the waveform of v_c.

In Figure 8, the chaotic motion can be stabilized quickly (at k = 229) to the orbit of period-2. At this time, the phase portrait has a partial depression compared to the orbit of period-1. The waveforms of the inductive current and the capacitive voltage both indicate that the system is in the period-2 motion.

Through simulation studies, it can be found that the expected targets can also be set as other orbits of period-n when applying the method in this paper to the chaotic motion of the system shown in Figure 1 under the condition that the controller parameters are selected appropriately. However, due to the consideration of space, only the control renderings of orbits of period-1 and period-2 are given here. The control effects of the remaining periodic orbits are not discussed.

It can be seen from the above results that the proposed MFACC method is effective and the control effect is ideal for the chaotic motion control of the current-mode-controlled Boost converter system.

The influence analysis of controller parameters on chaos control effect

The simulation study found that the values of the controller parameters can have an impact on the chaos control effect. In the following, the relationship between different values of controller parameters and the effect of chaos control is analyzed and compared for the case where the expected control target is period-1 motion. The case where the control target is the period-2 motion is similar to the case where the control target is moving in period-1. For the sake of space, it will not be described in this section.

Analysis of influence of single parameter value of controller on chaos control effect

The chaos control of Boost converter is simulated by changing the value of one of the parameters ${\hat{ϕ}}_{1}$ , λ, ρ, μ, η in turn, and the rest of the parameter settings are the same as in Section 4.1, in order to analyze the influence of different values of the single parameter of the controller on the chaos control effect, and part of the simulation results are shown in Table 2 (each parameter of the controller takes 4 simulation experiment results, 5 parameters total 20 simulation experiment outcomes). Due to space considerations, the rest of the simulation outcomes are not repeated.

Table 2.

The influence of single parameter value of controller on chaos control effect.

No.	${\hat{ϕ}}_{1}$	λ	ρ	μ	η	Iteration
1	0.5	0.00036	1.95	0.91	1	266
2	0.6	0.00036	1.95	0.91	1	258
3	1	0.00036	1.95	0.91	1	271
4	2	0.00036	1.95	0.91	1	285
5	0.7	0.00031	1.95	0.91	1	274
6	0.7	0.00032	1.95	0.91	1	261
7	0.7	0.00035	1.95	0.91	1	264
8	0.7	0.00038	1.95	0.91	1	263
9	0.7	0.00036	1.85	0.91	1	273
10	0.7	0.00036	1.90	0.91	1	256
11	0.7	0.00036	1.96	0.91	1	263
12	0.7	0.00036	2	0.91	1	261
13	0.7	0.00036	1.95	0.77	1	273
14	0.7	0.00036	1.95	0.79	1	263
15	0.7	0.00036	1.95	0.83	1	270
16	0.7	0.00036	1.95	0.96	1	272
17	0.7	0.00036	1.95	0.91	0.97	272
18	0.7	0.00036	1.95	0.91	0.99	261
19	0.7	0.00036	1.95	0.91	1.02	279
20	0.7	0.00036	1.95	0.91	1.05	281

It can be seen from Table 2 that although chaotic motion of Boost converter can be guided to the desired period-1 orbit under each set of parameters, the control effect is different from that in Section 4.1: for example, when ${\hat{ϕ}}_{1}$ = 2 (i.e. the controller parameters are set as No. 4 in Table 2), the chaotic motion can be controlled into stable period-1 motion when the number of iterations reaches the 285th generation; When λ = 0.00031 (i.e. the controller parameters are set as No. 5 in Table 2), the system can be stabilized on the period-1 orbit when iterating to the 274th generation; When ρ = 1.85 (i.e. the controller parameters are set as No. 9 in Table 2), the chaos can be controlled as stable period-1 motion when the iteration number is 273; When μ = 0.77 (No. 13 in Table 2), the chaotic motion can be controlled when the 273rd iteration is reached; When η = 1.05 (No. 20 in Table 2), the chaotic motion can be controlled when the number of iterations is 281. Compared with the control effect achieved when the number of iterations is 250 in Section 4.1, the control speeds under the above parameter settings are all significantly slower.

To better evaluate the influence of single parameter value of the controller on the chaos control effect, this section selects part of the simulation experiment results from Table 2 to draw the effect diagram of chaotic motion control of the system. In order to make the figure more clear and intuitive, the selected outcomes are the simulation experiment results with the minimum iteration times of each parameter of the controller (i.e. the fastest control speeds). Five parameters are five simulation experiment results in total. The simulation result in Section 4.1 is added for comparative analysis, and the chaos control effect diagram as shown in Figure 9 is drawn.

Figure 9.

Chaos control effect diagram of single parameter value of the controller: (a) global control effect (range of system iteration: 0–1000) and (b) local control effect (range of system iteration: 240–275).

Figure 9(a) is a global rendering of the control applied to chaotic motion when the system iterates 200 times (the system iterates 1000 times in total). In order to compare and analyze the influence of controller parameters on the speed of chaos control more clearly, the region range of system iteration times from 240 to 275 is marked with an elliptical circle on Figure 9(a), and the local chaos control effect diagram as shown in Figure 9(b) is obtained after enlarging the region. It can be clearly observed from Figure 9 that: when ${\hat{ϕ}}_{1}$ = 0.6 (i.e. the controller parameters are shown as No. 2 in Table 2), the chaotic motion of the system is controlled as stable period-1 motion when the number of iterations reaches the 258th generation, shown in green in the figure; When λ = 0.00032 (i.e. the controller parameters are set as No. 6 in Table 2), the system is stably controlled on the period-1 orbit when iterating to the 261st generation, which is drawn with cyan color in the figure; When ρ = 1.90 (i.e. the controller parameters are set as No. 10 in Table 2), the chaotic motion is controlled as stable period-1 motion when the number of iterations is 256, and the chaos control effect is drawn in magenta; When μ = 0.79 (i.e. No. 14 in Table 2), the system is controlled to move in period-1 at the 263rd iteration and is shown in yellow; When η = 0.99 (i.e. No. 18 in Table 2), the chaotic motion is stabilized to the motion of period-1, denoted by black, when the number of iterations is 261. Compared with the expected control effect achieved when the number of iterations is 250 in Section 4.1 (shown in blue in Figure 9), the control speeds under the above parameter settings are still slower.

Evidently, the value of each parameter of the controller will have an impact on the chaos control effect. To obtain a relatively satisfactory control effect, the relatively optimal controller parameter must be selected.

Analysis of the influence of different parameter combinations of the controller on the chaos control effect

Simultaneously changing the values of any three parameters/four parameters/all parameters of the controller (5 simulation experiment results for each case, 15 simulation outcomes in total), the chaotic control effect obtained is analyzed and compared, and the results are shown in Table 3. In Table 3, No.1 to No.5 are the simulation results when three parameters are varied simultaneously; No.6 to No.10 are the simulation results when four parameters are changed at the same time; No.11 to No.15 are the results when all parameters are changed in the meantime. Table 3 shows that different combinations of controller parameters will have corresponding effects on the chaotic control effect of the system. According to the simulation outcomes in Section 4.1, it can be found that only the relatively optimal combination of controller parameters can obtain the relatively optimal control effect.

Table 3.

The influence of different combinations of parameters of the controller on the effect of chaotic control.

No.	${\hat{ϕ}}_{1}$	λ	ρ	μ	η	Iteration
1	0.9	0.00091	1.93	0.91	1	279
2	0.5	0.00063	1.95	0.32	1	280
3	0.7	0.00057	1.59	0.98	1	281
4	0.7	0.00025	1.95	0.97	0.97	272
5	0.7	0.00036	1.91	0.64	0.75	276
6	0.6	0.00042	1.82	0.63	1	270
7	0.7	0.00028	1.82	0.86	1.03	285
8	0.8	0.00093	1.93	0.22	1	274
9	1	0.00032	1.95	0.34	0.94	274
10	1	0.00076	1.97	0.77	1	281
11	0.17	0.00084	1.16	0.21	0.61	270
12	0.23	0.00056	1.25	0.71	0.41	268
13	0.58	0.00095	1.81	0.55	0.80	274
14	0.9	0.00029	1.88	0.37	0.95	273
15	1	0.00075	1.72	0.78	0.58	272

To better evaluate the influence of different parameter combinations of the controller on the control effect and let readers read more clearly, this section selects three groups of parameters corresponding to the fastest chaos control speeds No.4, No.6, and No.12 from No.1 to No.5, No.6 to No.10, No.11 to No.15 in Table 3 respectively. In addition, the parameters in Section 4.1 are added for comparison, and the effect diagram of the chaotic motion control of the system as shown in Figure 10 is drawn. In Figure 10, green, cyan and magenta are respectively used to show the chaos control effect under the corresponding parameters of No.4, No.6, and No.12. Draw the control effect under the parameters in Section 4.1 in blue. The comparative analysis process is similar to that in Section 4.2.1. Due to limited space, it will not be repeated in this section.

Figure 10.

Chaos control effect diagram of different parameter combinations of the controller: (a) global control effect (range of system iteration: 0–1000) and (b) local control effect (range of system iteration: 240–285).

To sum up, not only the change of a certain parameter in the controller will affect the chaos control effect, but also the simultaneous change of multiple parameter values of the controller will affect the chaos control effect. Only by selecting the relatively optimal combination of controller parameters can the relatively optimal control effect be obtained. It can be seen that the chaotic motion of Boost converter can be effectively controlled by using the proposed method as long as the controller parameters are properly selected.

Conclusions

In this paper, a MFACC method of the Boost converter system is proposed by taking advantage of the advantages that the MFAC method does not depend on the PMM of the system and only uses the I/O data of the system to design the controller. The method achieves the purpose of guiding the chaotic motion of the system to periodic orbit. The simulation results show that the controller designed in this paper can meet the control requirements well. The controller quickly makes the system search automatically and finally stabilizes it in the expected periodic orbit. In addition, reasonable selection of the controller parameters is the guarantee of good control quality. The research results of this paper provide a new idea and method to solve the control problems of chaotic motion of Boost converter in engineering practice.

In the future, in order to further improve the design efficiency of the controller, the next step is to carry out the corresponding research on chaos control of DC-DC converter system by combining intelligent optimization algorithm and MFAC, expecting to obtain the corresponding research results in the construction of control performance index function and optimization algorithm configuration.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is supported by National Natural Science Foundation of China (Grant No. 51665027), Youth Science and Technology Fund program of Gansu Province (Grant No. 21JR7RA328), Introduction of talents research launch project of Shanghai Institute of Technology (Grant No. YJ2021-18 and YJ2021-19). Collaborative Innovation Fund Project of Shanghai Institute of Technology (XTCX2023-20).

ORCID iD

Ning-zhou Li

References

Lin

, et al. Robust chaos suppression of uncertain unified chaotic systems based on chattering-free sliding mode control. Meas Control 2022; 55(5-6): 321–329.

Liu

Wang

Zhu

, et al. Quadratic buck-boost converter based on switched inductance. Trans China Electrotech Soc 2022; 37: 190–197.

Wang

, et al. Characteristics analysis of buck converter with switched inductor in CCM and DCM. Trans China Electrotech Soc 2022; 37(16): 4155–4168.

Bao

Zhang

Bao

, et al. Dynamical effects of memristive load on peak current mode buck-boost switching converter. Chaos Solitons Fractals 2019; 122: 69–79.

Tian

Wang

Tao

CX.

Research on chaos and its control in current-mode switched-inductor converter. Electr Drive Locomot 2017; 4: 36–40.

Wang

Tao

Tian

Research on nonlinear behavior control in switched-inductor buck-boost converter. Electr Drive 2017; 47(9): 24–28.

Sun

Yang

Liu

HC.

Bifurcation in discontinuous current-programmed buck-boost converter with switched-inductor structure. J Harbin Inst Technol 2015; 47(2): 55–61.

Ito

Asahara

Kousaka

, et al. Clock pulse modulation for ripple reduction in buck-converter circuits. Chaos Solitons Fractals 2018; 111: 138–145.

Jia

Applying a self-feedback method to control chaos in the buck converter. Turk J Phys 2017; 41: 560–575.

10.

Luo

Wang

Chen

, et al. Research on bifurcation behaviour and chaos control in dc-dc buck converter. Acta Phys Sin 2003; 52(1): 12–17.

11.

Zhang

Yang

Long

, et al. Mechanism of instability behaviors and stabilization on V2 controlled buck converter. Acta Phys Sin 2013; 62: 21.

12.

El Aroudi

Mandal

Giaouris

, et al. Self-compensation of dc–dc converters under peak current mode control. Electron Lett 2017; 53(5): 345–347.

13.

Morcillo

Burbano

Angulo

Adaptive ramp technique for controlling chaos and subharmonic oscillations in DC–DC power converters. IEEE Trans Power Electron 2016; 31(7): 5330–5343.

14.

Tian

, et al. Analyses and control of chaotic behavior in dc-dc converters. Math Probl Eng 2018; 2018: 1–13.

15.

Zhang

XP.

Chaos control of buck-boost converter in current mode. Proc CSU-EPSA 2018; 30(8): 51–57.

16.

Deivasundari

Uma

Vincent

, et al. Non-linear intermittent instabilities and their control in an interleaved dc/dc converter. IET Power Electron 2014; 7(5): 1235–1245.

17.

Bao

Liu

Bifurcation of Boost converter with two boundaries and its stability control by ramp compensation. Acta Phys Sin 2009; 58(5): 2949–2956.

18.

Luo

Xiao

Tang

SC.

Study on chaos model and control of voltage-mode buck converter. Comput Simul 2014; 31(9): 151–155.

19.

Zhou

Luo

, et al. Time-delayed feedback control of chaos in Boost converter and its optimization. Acta Phys Sin 2007; 56(11): 6275–6281.

20.

Yang

Wang

Dynamic analysis of the switched-inductor buck-boost converter based on the memristor. Electronics 2021; 10: 452.

21.

Zhou

Tse

Qiu

, et al. An improved resonant parametric perturbation for chaos control with applications to control of dc/dc converters. Chin Phys 2005; 14(1): 61–66.

22.

Huang

GX.

Study of controlling chaos in high-intensity accelerator and DC/DC switching power converter by neural network. MA.E. Thesis, Guangxi Normal University, China, 2004.

23.

Ling

TF.

Based on current controlled dc-dc converter chaos control research. MA.E. Thesis, Chongqing University of Technology, China, 2021.

24.

Hou

Dong

Jin

ST.

Model-free adaptive control with coordinates compensation for automatic car parking systems. Zidonghua Xuebao 2015; 41(4): 823–831.

25.

Jin

, et al. Model-free adaptive predictive control for an urban road traffic network via perimeter control. IEEE Access 2019; 7: 172489–172495.

26.

Liu

GP.

Networked multi pipeline pressure synchronous coordinated predictive control based on the model free adaptive method. Int J Eng Syst Modelling Simul 2019; 11(2): 60–68.

27.

Yang

Xiang

Hao

, et al. Research on PSA-MFAC for a novel bionic elbow joint system actuated by pneumatic artificial muscles. J Mech Sci Technol 2017; 31(7): 3519–3529.

28.

Yang

Gao

Chen

, et al. Hammerstein adaptive impedance controller for bionic wrist joint actuated by pneumatic muscles. IEEE Access 2019; 7: 47–56.

29.

Liu

Guo

, et al. A modified model-free-adaptive-control-based real-time energy management strategy for plug-in hybrid electric vehicle. Energy Sci Eng 2022; 10(10): 4007–4024.

30.

Wang

Zhang

Research on model-free adaptive control of electro-hydraulic servo system of continuous rotary motor. IEEE Access 2022; 10: 31165–31174.

31.

Zhou

Chen

Xie

, et al. Application of resonant parametric perturbation to the chaos control in boost converter and its optimization. Acta Phys Sin 2004; 53(11): 3676–3683.