Prediction control method with peak current constraint for Boost converter

Abstract

The invention discloses a prediction control method with peak current constraint for a Boost converter, and under the control method, the Boost converter is high in interference immunity and high in dynamic response speed. The method mainly comprises the following steps: constructing a mathematical model of the Boost converter; designing an outer ring non-smooth integral controller; designing a discontinuous super-spiral observer to estimate the uncertainty of the input voltage; designing an inner ring model prediction controller; constructing a peak current model of the Boost converter, and designing a current constraint according to the peak current model; and designing a switching function through inner loop model predictive control and peak current constraint to obtain a Boost converter composite control method. According to the invention, the current constraint capability of the Boost converter in the starting stage is obviously enhanced, the Boost converter can quickly respond when being disturbed, and the control precision of the output voltage following the reference voltage is improved.

Description

Technical Field

[0001] This invention belongs to the field of DC-DC Boost converter control for electric vehicles, and specifically relates to a predictive control method with peak current constraint for Boost converters. Background Technology

[0002] Boost converters, due to their boost capability, are widely used in various industrial scenarios, especially in the electric vehicle field. In practical applications, Boost converters often face challenges such as load variations and power supply voltage fluctuations, and typically require sufficiently fast dynamic response. Traditional methods can only achieve this by increasing the control gain, but this approach, in turn, can lead to current overload. Furthermore, Boost converters generate extremely large inductor currents during startup, which may damage the hardware system. Therefore, it is necessary to study the current constraint problem of Boost converters.

[0003] Most existing model predictive control algorithms employ an outer-loop PI control and an inner-loop model predictive control scheme, with current constraints primarily achieved by limiting the upper limit of the output duty cycle. However, the outer-loop PI controller can only achieve asymptotic convergence. Furthermore, as a typical linear control strategy, the PI controller cannot provide ideal control performance for nonlinear systems. Therefore, it is necessary to design a controller that can achieve faster control convergence speed and is suitable for nonlinear systems such as Boost converters. While limiting the upper limit of the output duty cycle to achieve current constraint is simple, setting the maximum duty cycle limit too large will result in poor current constraint performance, while setting it too small will affect the boost converter's boost ratio and dynamic response speed. Therefore, there is an urgent need to design a control method that can accurately constrain current while ensuring good dynamic response speed and robustness. Summary of the Invention

[0004] To address the shortcomings of existing Boost converter control schemes, namely the inability to improve dynamic response speed and robustness while ensuring current constraints, this invention proposes a predictive control method with peak current constraints for Boost converters.

[0005] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0006] A predictive control method with peak current constraint for Boost converters:

[0007] For the Boost converter operating in continuous inductor current mode, the state-space averaging method is used to establish its mathematical model. Considering the input voltage variation, the actual mathematical model of the Boost converter is constructed. The forward Euler method is used to discretize the mathematical model of the Boost converter.

[0008] Using the non-smooth integral controller as the outer loop of predictive control, the reference current output by the non-smooth integral controller is obtained.

[0009] The input voltage uncertainty in the actual mathematical model of the Boost converter is estimated using a discontinuous superspiral observer, which is then used to calculate the compensated current constraint control quantity.

[0010] Delay compensation technology is used to compensate for the inductor current measurement value;

[0011] The predictive control inner loop targets the discrete mathematical model of the Boost converter, uses the model predictive control algorithm to differentiate the objective function, obtains the optimal control quantity u, and performs delay compensation to obtain the compensated optimal control quantity u'; the objective function is constructed based on the reference current and inductor current values ​​output by the non-smooth integral controller.

[0012] The peak current model is used instead of the average current model to predict the peak inductor current at the next moment, and the current constraint control quantity is calculated based on the set maximum current limit and the inductor current measured at the current moment. And perform delay compensation to obtain the compensated current constraint control quantity. ;

[0013] Based on whether the peak inductor current exceeds the set value in the next cycle, select the output as u' or It acts on the Boost converter.

[0014] Furthermore, the non-smooth integral controller is:

[0015]

[0016]

[0017] in, For the error of the Boost converter output voltage, For reference voltage, This refers to the output voltage. The reference current output by the non-smooth integral controller. For intermediate quantities, constants , , , , These are the adjustable positive parameters of the non-smooth integral controller.

[0018] Furthermore, the discontinuous superspiral observer is:

[0019]

[0020] in, The system control input duty cycle signal is given, and L is the inductance parameter. Indicates the nominal value of the input voltage. This represents an estimate of the uncertainty related to the input voltage. This represents the filtered perturbation estimate. This indicates the inductor current estimation error of the Boost converter. For symbolic functions, This represents the estimated inductor current, and the observer parameters satisfy... , , , and This is the adjustable observer parameter gain.

[0021] Furthermore, the uncertainty related to the input voltage at time k+1 is: ,in, The switching cycle time. express The first-order difference.

[0022] Furthermore, the objective function is: Where b is the weighting factor for the current error term in the objective function. The reference current output by the non-smooth integral controller. The value of the inductor current at time k+1.

[0023] Furthermore, the optimal control quantity after compensation ,in, This is the input voltage.

[0024] Furthermore, the compensated current constraint control quantity ,in, It is a manually set maximum current limit. The slope of the inductor current rise during one switching cycle.

[0025] Furthermore, the u' or This is determined using the following switching function: ,in, This represents the peak value of the inductor current after compensation within one switching cycle.

[0026] The beneficial effects achieved by this invention are as follows:

[0027] (1) The present invention uses a non-smooth integral controller, which significantly improves the robustness and dynamic response speed of the Boost converter compared with the traditional PI controller, reduces the influence of the nonlinear characteristics of the Boost converter on the control effect of the controller, and combines the non-smooth integral controller with model predictive control, thereby improving the robustness and dynamic response speed of the Boost converter.

[0028] (2) The peak current constraint used in this invention solves the prediction error caused by the average current model used in traditional predictive control schemes and improves the prediction accuracy. At the same time, by back-deriving the control quantity u* through the peak current model, the precise constraint on the inductor current is achieved, which significantly improves the control accuracy of the Boost converter in the startup phase.

[0029] (3) The discontinuous superspiral observer used in this invention estimates the input voltage uncertainty of the Boost converter, which reduces the prediction error of the peak current model and improves the prediction accuracy.

[0030] (4) The switching function used in this invention enables free switching between two control quantities, while ensuring the robustness of model predictive control and the accuracy of peak current constraint, thereby improving the robustness of the Boost converter and ensuring its excellent dynamic performance and current constraint capability. Attached Figure Description

[0031] Figure 1 Here is a diagram of the Boost converter topology;

[0032] Figure 2 This is a schematic diagram of the control structure of the present invention;

[0033] Figure 3 This is a schematic diagram of the outer loop non-smooth integral controller structure of the present invention;

[0034] Figure 4 This is a schematic diagram of the inner loop controller structure of the present invention;

[0035] Figure 5 This is a schematic diagram of the discontinuous superspiral observer structure of the present invention;

[0036] Figure 6 This is a flowchart illustrating the implementation of the method of the present invention;

[0037] Figure 7 The waveform of the inductor current during the startup phase of the Boost converter;

[0038] Figure 8 The output voltage waveform of the Boost converter when the input voltage changes abruptly.

[0039] Figure 9This is a waveform of the output voltage of the Boost converter when the load resistance changes abruptly. Detailed Implementation

[0040] The following specific examples illustrate the implementation of the present invention, and those skilled in the art can easily implement it based on the content disclosed in this specification. To make the objectives, technical solutions, and effects of the present invention clearer, the technical solutions in the examples of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. It should be noted that the described embodiments are only some embodiments of the present invention, and not all embodiments. They are only used to explain the present invention and are not intended to limit the present invention.

[0041] The Boost converter topology of this invention includes a DC voltage source. The components include a controllable switching device S, a diode D, an inductor L, a capacitor C, and a load R, where the DC voltage source is... The capacitor C is connected in series with the inductor L and then in parallel with the controllable switching device S. The capacitor C is connected in parallel with the load R and then in series with the diode D, and finally in parallel with the controllable switching device S. The topology is shown in Figure 1. The control block diagram is shown in Figure 2. The sensor collects the output voltage of the Boost converter. and inductor current This provides feedback signals for subsequent control. With reference voltage The difference signal is input to the non-smooth integral controller, and the output is the reference value of the inductor current. ; Input the discontinuous superhelical observer to obtain the predicted observations. ; After time delay compensation, the inductor current after time delay compensation is obtained. ; , and The system consists of two modules: peak current constraint control and model predictive control. The outputs of these two control modules are selected by a switching module. Figure 4 ), to obtain the final control signal ; The signal is fed into the PWM generator, and the generated PWM drive signal acts on the switching devices of the Boost converter to achieve closed-loop control.

[0042] In this embodiment, the specific parameter is: input voltage =12V, expected output voltage =24V, load R=75 Capacitor C=470 Inductance L=330 The system frequency is f=100kHz.

[0043] like Figure 6 As shown, the implementation process of predictive control with peak current constraint for a Boost converter is as follows:

[0044] Step 1: For the Boost converter operating in continuous inductor current mode, establish its mathematical model using the state-space averaging method. Then, considering the input voltage variation, give the actual mathematical model of the Boost converter.

[0045] The mathematical model for the Boost converter is as follows:

[0046]

[0047] in, , These represent the output voltage and inductor current, L, C, R represents the inductance parameters, capacitance parameters, input voltage, and load resistance, respectively. The system control input duty cycle signal, This indicates that the switch is on. This indicates that the switch is off.

[0048] Next, considering the input voltage variation, the actual mathematical model of the Boost converter is given:

[0049]

[0050] in, The expression for the uncertainty related to the input voltage is: , This indicates the nominal value of the input voltage.

[0051] Step 2: Discretize the mathematical model of the Boost converter using the forward Euler method;

[0052] A discrete mathematical model of the Boost converter is established using the forward Euler method:

[0053]

[0054] in, , These are the output voltage and inductor current values ​​measured by the sensor at time k, respectively. , These are the predicted values ​​of the output voltage and inductor current at time k+1, respectively. The switching cycle time (set to) ), Let be the system control input duty cycle at time k.

[0055] Step 3: Use a "non-smooth integral controller" as the predictive control outer loop to obtain the reference current;

[0056] like Figure 3 As shown, the "non-smooth integral controller" used in the predictive control outer loop is:

[0057]

[0058]

[0059] in, For the error of the Boost converter output voltage, The reference current output by the non-smooth integral controller. For intermediate quantities, constants , , , , These are the adjustable positive parameters of the non-smooth integral controller. In this embodiment, the specific parameters are set as follows: , , , , .

[0060] Step 4: Estimate the input voltage uncertainty using a discontinuous superspiral observer;

[0061] Discontinuous superhelical observer for estimating input voltage uncertainty ( Figure 5 The construction of ) is as follows:

[0062]

[0063] in, This represents an estimate of the uncertainty related to the input voltage. This represents the filtered perturbation estimate. This indicates the inductor current estimation error of the Boost converter. This represents the estimated inductor current, and the observer parameters satisfy... , , , and For the adjustable observer parameter gain, the specific parameter in this embodiment is set as follows: , , , , .

[0064] At time k+1 The value can be calculated using the following formula:

[0065]

[0066] in, express The first-order difference.

[0067] Step 5: Use delay compensation technology to adjust the inductor current value. Provide compensation;

[0068] Delay compensation technology replaces the current measurement with the calculated current value at the next moment to avoid the computational delay introduced by the digital signal processor in a real system. After compensation, the predicted inductor current expression is:

[0069]

[0070] Step 6: Predictive Control Inner Loop. For the discrete mathematical model of the Boost converter, the model predictive control algorithm is used to differentiate the objective function and obtain the optimal control quantity u.

[0071] The objective function in inner-loop model predictive control can be expressed as:

[0072]

[0073] Where b is the weighting factor of the current error term in the objective function, and its value is a real number greater than 0. In this embodiment, b is set to 10.

[0074] For the objective function By taking the partial derivative with respect to the duty cycle u and setting it equal to 0, the optimal duty cycle can be calculated. for:

[0075]

[0076] After considering delay compensation, its expression is:

[0077]

[0078] Step 7: Use the peak current model to replace the mathematical model of the Boost converter, predict the peak inductor current at the next moment, and calculate the control quantity based on the set maximum current limit and the current measured at the current moment. ;

[0079] The peak current model is:

[0080]

[0081]

[0082]

[0083] in, and The slope of the inductor current rise during one switching cycle. The slope of the inductor current decrease during one switching cycle. and The rise time of the inductor current within one switching cycle. The time it takes for the inductor current to decrease within one switching cycle;

[0084] The peak inductor current during one switching cycle can be expressed as:

[0085]

[0086] in, This represents the estimate at time k by the nonlinear superhelical observer. The peak current of the inductor at time k+1, after considering time delay compensation, is expressed as:

[0087]

[0088] Current constraint control quantity It can be calculated using the following formula:

[0089]

[0090] in, It is a manually set maximum current limit value, which is an empirical value (its range is 3~5A). In this embodiment, it is set to 4A.

[0091] Compared to the traditional average current model, the peak current model can accurately calculate the peak value of the inductor current within a switching cycle, rather than the average value, thus improving prediction accuracy. Considering delay compensation, the expression for the current constraint control quantity is:

[0092]

[0093] Step 8: Based on whether the peak inductor current in the next cycle exceeds the set value, select the output of predictive control as the optimal duty cycle u. Or current constraint control quantity The final Boost converter control method based on non-smooth integral and peak current constraint is obtained.

[0094] The switching is implemented using the following switching function:

[0095]

[0096] in, It is a symbolic function.

[0097] When predicting the peak inductor current in the next cycle Greater than the maximum current set by the user At that time, output control quantity When predicting the peak inductor current of the next cycle Less than the maximum current set by the user At that time, output control quantity By switching between the two control variables, it is possible to achieve precise control of the inductor current while retaining the strong robustness of model predictive control.

[0098] Based on this embodiment, experiments will be conducted to further demonstrate the effectiveness of the present invention.

[0099] The experimental results are presented below under three conditions. First, the current waveform of the Boost converter during the startup phase (i.e., from the start of system response to the steady state) is tested to verify the precise current constraint effect of the present invention. Second, the voltage waveform of the Boost converter under sudden increases and decreases in input voltage is tested to verify the robustness of the method of the present invention. Third, the voltage waveform of the Boost converter under sudden increases and decreases in load resistance is tested to verify the robustness of the method of the present invention.

[0100] Specifically, in the current constraint test experiment, the Boost converter starts up and begins boosting at 0s. In the input voltage surge experiment, The input voltage suddenly changes from 12V to 16V. The input voltage suddenly changes from 16V to 12V. In the load resistance sudden change experiment, Load resistance 75 Mutation to 150 , Load resistance 150 Mutation to 75 .

[0101] Case 1: Current response waveform during the startup phase of the Boost converter

[0102] like Figure 7 As shown, the maximum inductor current of the Boost converter during the startup phase is 3.98A, which is precisely constrained to the set maximum current value of 4A.

[0103] Case 2: Voltage response waveform of the Boost converter when the input voltage changes abruptly

[0104] like Figure 8 As shown, The input voltage suddenly changes from 12V to 16V. When the input voltage suddenly changes from 16V to 12V, the overshoot / undershoot of the output voltage is very small, demonstrating the good robustness of this invention.

[0105] Case 3: Voltage response waveform of the Boost converter when the load resistance changes abruptly

[0106] like Figure 9 As shown, Load resistance 75 Mutation to 150 , Load resistance 150 Mutation to 75 When the load resistance changes abruptly, the overshoot / undershoot of the output voltage is very small, demonstrating the good robustness of this invention.

[0107] Although the present invention has been described according to various specific embodiments, those skilled in the art will recognize that the invention can be practiced with modifications within the spirit and scope of the claims. Therefore, any obvious improvements, substitutions, or modifications that can be made by those skilled in the art without departing from the essence of the invention are within the scope of protection of the present invention.

Claims

1. A predictive control method with peak current constraint for a Boost converter, characterized in that: For a Boost converter operating in continuous inductor current mode, a mathematical model is established using the state-space averaging method. Considering the input voltage variation, a practical mathematical model of the Boost converter is constructed. The mathematical model of the Boost converter is discretized using the forward Euler method; Using the non-smooth integral controller as the outer loop of predictive control, the reference current output by the non-smooth integral controller is obtained. The input voltage uncertainty in the actual mathematical model of the Boost converter is estimated using a discontinuous superspiral observer, which is then used to calculate the compensated current constraint control quantity. Delay compensation technology is used to compensate for the inductor current measurement value; The predictive control inner loop targets the discrete mathematical model of the Boost converter, uses the model predictive control algorithm to differentiate the objective function, obtains the optimal control quantity u, and performs delay compensation to obtain the compensated optimal control quantity u'; the objective function is constructed based on the reference current and inductor current values ​​output by the non-smooth integral controller. The peak current model is used instead of the average current model to predict the peak inductor current at the next moment, and the current constraint control quantity is calculated based on the set maximum current limit and the inductor current measured at the current moment. And perform delay compensation to obtain the compensated current constraint control quantity. ; Based on whether the peak inductor current exceeds the set value in the next cycle, select the output as u' or It acts on the Boost converter.

2. The predictive control method with peak current constraint according to claim 1, characterized in that, The non-smooth integral controller is: in, For the error of the Boost converter output voltage, For reference voltage, This refers to the output voltage. The reference current output by the non-smooth integral controller. For intermediate quantities, constants , , , , These are the adjustable positive parameters of the non-smooth integral controller.

3. The predictive control method with peak current constraint according to claim 2, characterized in that, The discontinuous superspiral observer is: in, The system control input duty cycle signal is given, and L is the inductance parameter. Indicates the nominal value of the input voltage. This represents an estimate of the uncertainty related to the input voltage. This represents the filtered perturbation estimate. This indicates the inductor current estimation error of the Boost converter. For symbolic functions, This represents the estimated inductor current, and the observer parameters satisfy... , , , and This is the adjustable observer parameter gain.

4. The predictive control method with peak current constraint according to claim 3, characterized in that, The uncertainty related to the input voltage at time k+1 is: ,in, The switching cycle time. express The first-order difference.

5. The predictive control method with peak current constraint according to claim 4, characterized in that, The objective function is: Where b is the weighting factor for the current error term in the objective function. The reference current output by the non-smooth integral controller. The value of the inductor current at time k+1.

6. The predictive control method with peak current constraint according to claim 5, characterized in that, Optimal control quantity after compensation ,in, This is the input voltage.

7. The predictive control method with peak current constraint according to claim 5, characterized in that, Compensated current constraint control quantity ,in, It is a manually set maximum current limit. The slope of the inductor current rise during one switching cycle.

8. The predictive control method with peak current constraint according to claim 5, characterized in that, The u' or This is determined using the following switching function: ,in, This represents the peak value of the inductor current after compensation within one switching cycle.

Images