TWI721582B - Digital fuzzy controller and control method based on adaptive network based fuzzy inference system for boost converter - Google Patents

Abstract

A digital fuzzy controller and control method based on an adaptive network based fuzzy inference system for the boost converter are disclosed. The digital fuzzy controller includes a boost converter and a digital computer. The boost converter includes an inductor, a switch and an output capacitor. The switch is controlled by a continuous signal to convert input voltage to output voltage. The digital computer includes a processor and a database. The database saves the continuous signal and the sampled signal. The processor access the database to conduct a training process, so as to generate the boost converter’s control model by the adaptive network based fuzzy inference system algorithm. The boost converter’s control model is used to generate the real control signal.

Description

Digital fuzzy control of power converter based on adaptive network fuzzy inference system Device and its control method

The present invention relates to a digital fuzzy controller and a control method of a power converter, in particular to a digital fuzzy controller and a control method of the power converter based on an adaptive network fuzzy inference system.

Direct current-direct current (DC-DC) converter or direct current-alternating current (DC-AC) converter is a very important topic in power conversion, and it is usually implemented by a basic continuous circuit, and its conversion efficiency will be determined by the use of Affected by the control method. When applying fuzzy control to a DC-to-AC power converter, too much time is usually wasted for trial and error learning of input scaled factor, fuzzy rule, and membership function, resulting in Poor performance.

In view of this, how to build a controller or control method for power converters with improved performance is a topic that is emphasized in this field. Therefore, the inventor of the present invention considered and designed a digital fuzzy controller and its control method for a power converter based on an adaptive network fuzzy inference system to improve the lack of existing technology, thereby enhancing the industrial application.

In view of the above-mentioned known technical problems, the purpose of the present invention is to provide a digital fuzzy controller and control method of a power converter based on an adaptive network fuzzy inference system to solve the known problem of continuous control signal generation.

According to one objective of the present invention, a digital fuzzy controller for a power converter based on an adaptive network fuzzy inference system is proposed, which includes a boost converter and a digital computer. The boost converter includes an inductor, a switch and an output capacitor. One end of the inductor is electrically connected to the input power source, the switch is connected to the other end of the inductor, and the other end is connected to the output capacitor. The switch is controlled by a continuous signal to convert the input voltage into The output voltage. The digital computer includes a processor and a database. The database receives continuous signals, and samples the output voltage and reference voltage according to the continuous signals, and stores the sampled signals and corresponding continuous signals in the database. The processor accesses the database to execute the data training program. The sampled signal and continuous signal are used to establish a boost converter control model through an adaptive network fuzzy inference algorithm, and the actual control signal is generated through the boost converter control model. The control signal is sent to the switch.

Preferably, the sampling signal may include the difference value between the output voltage and the reference voltage and the change value between the difference value and the difference value corresponding to the previous continuous signal.

Preferably, the adaptive network fuzzy inference algorithm may include 5 layers of directional node operations, with difference values and change values as input parameters, continuous signals as output results, and a data training program to generate a boost converter control model.

Preferably, each of the directional node operations may include operations with a node parameter set and operations with an empty parameter set.

Preferably, the boost converter may include a DC-DC converter or a DC-AC converter.

According to another object of the present invention, a digital fuzzy control method for a power converter based on an adaptive network fuzzy inference system is proposed, which includes the following steps: setting up a boost converter, which includes an inductor, a switch, and an output Capacitor and switch are controlled by continuous signal to convert input voltage to output voltage; set up a digital computer, which includes a processor and a database, the database receives the continuous signal, and stores the continuous signal in the database; performs sampling procedures, digital The computer samples the output voltage and reference voltage according to the continuous signal, and stores the sampled signal in the database; performs data training procedures, the processor accesses the database, and establishes the sampled signal and the continuous signal through an adaptive network fuzzy inference algorithm The boost converter control model; and the digital computer generates the actual control signal through the boost converter control model, and transmits the actual control signal to the switch to control the boost converter.

Preferably, the sampling signal may include the difference value between the output voltage and the reference voltage and the change value between the difference value and the difference value corresponding to the previous continuous signal.

Preferably, the adaptive network fuzzy inference algorithm may include 5 layers of directional node operations, with difference values and change values as input parameters, continuous signals as output results, and a data training program to generate a boost converter control model.

Preferably, each of the directional node operations may include operations with a node parameter set and operations with an empty parameter set.

Preferably, the boost converter may include a DC-DC converter or a DC-AC converter.

As mentioned above, according to the digital fuzzy controller and control method of the power converter based on the adaptive network fuzzy inference system of the present invention, it can have one or more of the following advantages:

(1) The digital fuzzy controller and its control method of the power converter based on the adaptive network fuzzy inference system can reduce the traditional time spent on input adjustment factors, fuzzy rules and attribution Function to carry out trial-and-error learning problems, improve control efficiency, and further improve the performance of the power converter.

(2) The digital fuzzy controller of the power converter based on the adaptive network fuzzy inference system and its control method can use the calculation method of the adaptive network fuzzy inference system to train the input and output data, and adjust the system model parameters To optimize, the obtained control signal can accurately achieve the required output and reduce the operating error.

(3) The digital fuzzy controller of the power converter based on the adaptive network fuzzy inference system and its control method can be applied to the existing power converter without the need to redesign the complex judgment circuit, reducing the cost of development and production , Thereby reducing production costs.

1: Digital fuzzy controller of power converter

10: Boost converter

20: digital computer

C: output capacitor

L: inductance

R: resistance

T: Toggle switch

Vin: input power

S1~S5: steps

Figure 1 is a schematic diagram of a boost converter according to an embodiment of the invention.

Figure 2 is a schematic diagram of the digital fuzzy controller of the power converter according to the embodiment of the present invention.

Figure 3 is a schematic diagram of an adaptive network fuzzy inference algorithm according to an embodiment of the present invention.

Fig. 4 is a flowchart of a digital fuzzy control method for a power converter according to an embodiment of the present invention.

FIG. 5 is a schematic diagram of the tracking result of the DC-AC converter according to the embodiment of the present invention.

FIG. 6 is a schematic diagram of the tracking result of the DC-DC converter according to the embodiment of the present invention.

In order to facilitate the reviewers to understand the technical features, content and advantages of the present invention and the effects that can be achieved, the present invention is described in detail with the accompanying drawings and in the form of embodiment expressions as follows. The drawings used therein are as follows: The subject matter is only for the purpose of illustration and auxiliary description, and may not be the true proportions and precise configuration after the implementation of the invention. Therefore, it should not be interpreted in terms of the proportions and configuration relationships of the accompanying drawings, and should not limit the scope of rights of the invention in actual implementation. Hexian stated.

[0 1])，以將輸入電壓轉換為輸出電壓。上述電感L與輸出電容C的電壓(x _1c、x _2c )滿足以下的方程式：

Please refer to Figure 1, which is a schematic circuit diagram of a boost converter according to an embodiment of the present invention. As shown in the figure, the boost converter 10 includes an inductor L, a switch T, and an output capacitor C. One end of the inductor L is electrically conductive. Connect the input power Vin, the switch T is connected to the other end of the inductor L, and the other end is connected to the output capacitor C. The boost converter 10 may also include a resistor R, which is connected in series to the contact point of the switch T and the output capacitor C. V _in input power will provide a known constant pressure, so that the current through the inductor L having a voltage x _1c, after the switch output capacitance C T so that an output voltage x _2c. At this time, by providing a continuous signal u _c to the switch T, the switch T can be controlled to switch between on and off ( u _c

[0 1]) to convert the input voltage to the output voltage. The voltages (x _1c , x _{2 c} ) of the above-mentioned inductor L and output capacitor C satisfy the following equation:

y _c = x _{2 c} (3)

Wherein, u _c for the continuous control signal of the switch T, v _in a known constant pressure, R is the resistance R, y _c is the output signal of the boost converter. Boost converter 10 by controlling the continuous signal u _c, allowing a known constant voltage v _in correspondence to a desired output voltage converting x _2c. As for how to build a controller that controls the switch T, it will be described in the following embodiment.

Please refer to FIG. 2, which is a schematic diagram of the digital fuzzy controller of the power converter according to the embodiment of the present invention. As shown in the figure, the digital fuzzy controller 1 of the power converter includes a boost converter 10 and a digital computer 20. The boost converter 10 includes an inductor L, a switch T, and an output capacitor C, which are the same as the boost converter 10 described in the previous embodiment. Therefore, the same content will not be repeated. As mentioned above, the output voltage x _2c of the capacitor C of the boost converter 10 is the output signal. By comparing with the reference voltage x _r , the output voltage x _2c and the reference voltage x _r can be sampled, where the sampled signal can be Including the difference between the output voltage x _2c and the reference voltage x _r (e ( k ) = x _r - x _{2 c} ) and the change between the difference value and the difference value corresponding to the previous continuous signal (△ e ( k) )= e ( k )- e ( k -1)).

Digital computer 20 includes a processor and a database, the database can receive and store controls the switch T, a continuous signal u _c, and the output voltage x _2c and a reference voltage x _r is sampled based on a continuous signal u _c, the sample signal (e ( k ), △ e ( k )) are also stored in the database. The processor can execute the program to access the data in the database to perform the data training procedure, taking the difference e ( k ) and the change value △ e ( k ) of the sampled signal as input parameters, and the continuous signal u _c as the corresponding output result. After data sampling and learning and training, the boost converter control model is established through the adaptive network fuzzy inference algorithm. The digital computer 20 generates the actual control signal u _d through the boost converter control model, and transmits the actual control signal u _d to The switch T of the boost converter 10. The following will further explain the network node analysis architecture of the adaptive network fuzzy inference algorithm.

Please refer to Figure 3, which is a schematic diagram of the adaptive network fuzzy inference algorithm according to the embodiment of the present invention. As shown in the figure, the adaptive network fuzzy inference system 30 is a 5-layer network node structure, which is a directional The network architecture of the node calculation method, through the adjustment of the adjustable parameters of these nodes, can minimize the error of the output. In this embodiment, the adaptive network fuzzy inference system 30 can assume that there are two inputs ( x , y ) and one output z = f , corresponding to the aforementioned difference value, change value, and output continuous signal, and include Two rules: Rule 1: If x is A ₁ and y is B ₁ then f ₁ = p ₁ x + q ₁ y + r ₁

Rule 2: If x is A ₂ and y is B ₂ then f ₂ = p ₂ x + q ₂ y + r ₂

，

與

是一個權重比例，

與

的調整就是由輸出評分的增減來告知輸入量的好與不好參考，透過基於適應性網路模糊推論系統的學習可以找出最好權重比例，本實施例是以此來獲得最近似原先的模型。 In the above rules, x and y are the two input values of the system, A ₁ , A ₂ , B ₁ and B ₂ are syntax symbols, and the final output value is

,

versus

Is a weight ratio,

versus

The adjustment of is to inform the good and bad reference of the input amount by the increase or decrease of the output score. The best weight ratio can be found through the learning based on the adaptive network fuzzy inference system. This embodiment uses this to obtain the closest approximation to the original Model.

，其中x為節點之輸入值，A _i 為語言標籤(linguistic label)，O _i ¹為A _i 之隸屬函數(membership function)。第二層則是接收第一層之輸出，將其相乘後輸出，其滿足輸出規則為

。第三層接收第二層之輸出後，計算其平均值後輸出，其滿足

。第四層與第一層同樣接收參數集輸入，其輸出函數為

。第五層則是將第四層的所有輸出加總，其輸出函數為

。適應性網路模糊推論系統30整合了多種運算法則來訓練參數，藉由資料的訓練程序來建立控制模型，即升壓轉換器控制模型，使得控制誤差能降到最低。 As shown in the figure, if the layer has a node parameter set, it is represented by a box, that is, the output function of the node changes according to the parameter value, such as the first and fourth layer nodes. In contrast, when the layer is an empty parameter set, it is represented by a circle, that is, the output function of the node is fixed, such as other layers. In this embodiment, the node output function of the first layer is

, Where x is the input value of the node, A _i is a language label (linguistic label), O _i ¹ A _i of the membership function (membership function). The second layer receives the output of the first layer, multiplies it and outputs it, and it satisfies the output rule as

. After the third layer receives the output of the second layer, calculates its average value and outputs it, which satisfies

. The fourth layer and the first layer also receive the parameter set input, and its output function is

. The fifth layer is to add up all the outputs of the fourth layer, and its output function is

. The adaptive network fuzzy inference system 30 integrates a variety of algorithms to train parameters, and establishes a control model, that is, a boost converter control model, through data training procedures, so that the control error can be minimized.

Please refer to FIG. 4, which is a flowchart of a digital fuzzy control method for a power converter according to an embodiment of the present invention. As shown in the figure, the digital fuzzy control method of the power converter includes the following steps (S1~S5):

，K為正的常數，

，x _1r及x _2r分別為電感電流的穩定狀態值及電容電壓值，x _2c為可量測的輸出訊號。 Step S1: Set up the boost converter. The boost converter includes an inductor, a switch and an output capacitor. The switch is controlled by a continuous signal to convert the input voltage into an output voltage. For the configuration of the boost converter, please refer to the embodiment described in Figure 1. The original control model of the boost converter 10 can be x _c = f ( x _c ) + g ( x _c ) u _c ( t ), where Control signal at time t

, K is a positive constant,

, X _1r and x _2r are the steady state value of the inductor current and the capacitor voltage value, respectively, and x _2c is the measurable output signal.

Step S2: Set up the digital computer. The digital computer includes a processor and a database. The database receives continuous signals and stores the continuous signals in the database. Please refer to Figure 2 again. The boost converter 10 transmits the control signal u _c ( t ) to the digital computer and stores it in the database.

Step S3: Perform a sampling procedure. The digital computer samples the output voltage and reference voltage according to the continuous signal, and stores the sampled signal in the database. For the control signal at different time t , the output signal voltage is sampled, and the difference value e ( k ) between the output voltage and the reference voltage and the change value △ between the difference value and the difference value corresponding to the previous continuous signal e ( k ) is used as the sampling signal, and the sampling signal is also stored in the database.

Step S4: Carry out the data training procedure. The processor can execute the program to access the database, and through the command execution, the sampled signal and the continuous signal are calculated and trained through the above-mentioned adaptive network fuzzy inference algorithm, and the control signal u _d ( t ) of the boost converter is corrected. FLC ( e ( k ) , △ e ( k )).

Step S5: Establish a boost converter control model. The digital computer generates the actual control signal through the boost converter control model, and transmits the actual control signal to the switch to control the boost converter. The control signal of the modified boost converter is used to replace the continuous signal u _c ( t ) to establish a new boost converter control model x _d = f ( x _d ) + g ( x _d ) u _d ( t ), where The voltages of the inductor L and the output capacitor C are expressed as follows:

y _d = x _{2 d} (6)

By controlling the boost converter continuous signal u _d, that allows a constant voltage v _in known converted into a corresponding desired output voltage x _2d, that it is almost equal to the reference voltage x _r.

sin(20πt)於負載為220Ω的升壓轉換器上，其中直流電源供應為48V，切換開關的頻率為30kHz，電感為12mH，電容為15μF。通過上述的資料訓練程序後，實際控制訊號輸入與已知連續控制訊號輸入的比較如圖所繪示，其控制訊號幾乎完全相同。 Please refer to FIG. 5, which is a schematic diagram of the tracking result of the DC-AC converter according to the embodiment of the present invention. As shown in the figure, it is an embodiment of the tracking of a DC-to-AC converter, tracking the sinusoidal reference signal x _r (t)=100+30

sin(20πt) is applied to a boost converter with a load of 220Ω, where the DC power supply is 48V, the frequency of the switch is 30kHz, the inductance is 12mH, and the capacitance is 15μF. After the above-mentioned data training procedure, the comparison between the actual control signal input and the known continuous control signal input is shown in the figure, and the control signals are almost identical.

Please refer to FIG. 6, which is a schematic diagram of the tracking result of the DC-DC converter according to the embodiment of the present invention. As shown in the figure, it is an embodiment of the tracking of a DC-to-DC converter. The tracking constant reference signal x _r (t)=100 is applied to a boost converter with a load of 220Ω, where the DC power supply is 48V, and the switch is switched The frequency is 30kHz, the inductance is 12mH, and the capacitance is 15μF. After the above-mentioned data training procedure, the comparison between the actual control signal input and the known continuous control signal input is shown in the figure, and the control signals are almost identical.

From the above simulation results, it can be seen that only the available output signal can achieve the same effect as the known continuous signal control, and it can be applied to a DC-DC converter or a DC-AC converter.

The above description is only illustrative, and not restrictive. Any equivalent modifications or alterations that do not depart from the spirit and scope of the present invention should be included in the scope of the appended patent application.

1: Digital fuzzy controller of power converter

10: Boost converter

20: digital computer

Claims (8)

A digital fuzzy controller for a power converter based on an adaptive network fuzzy inference system, which includes: a boost converter including an inductor, a switch and an output capacitor, one end of the inductor is electrically connected to an input A power source, the switch is connected to the other end of the inductor, the other end is connected to the output capacitor, and the switch is controlled by a continuous signal to convert an input voltage to an output voltage; and a digital computer including a processor And a database that receives the continuous signal, samples the output voltage and a reference voltage according to the continuous signal, stores a sampled signal and the corresponding continuous signal in the database, and the processor accesses The database executes a data training procedure, uses the sampled signal and the continuous signal to establish a boost converter control model through an adaptive network fuzzy inference algorithm, and generates an actual control through the boost converter control model Signal to transmit the actual control signal to the switch; wherein the sampling signal includes a difference between the output voltage and the reference voltage and a change between the difference and the difference corresponding to the previous continuous signal . The digital fuzzy controller of a power converter based on an adaptive network fuzzy inference system as described in claim 1, wherein the adaptive network fuzzy inference algorithm includes 5-layer directional node operations, and the difference value and the change The value is used as an input parameter, and the continuous signal is used as an output result, and the boost converter control model is generated through the data training procedure. The digital fuzzy controller of a power converter based on an adaptive network fuzzy inference system as described in claim 2, wherein each of the directional node operations includes a node parameter Operation of number set and operation of empty parameter set. The digital fuzzy controller of a power converter based on an adaptive network fuzzy inference system as described in claim 1, wherein the boost converter includes a DC-DC converter or a DC-AC converter. A digital fuzzy control method for a power converter based on an adaptive network fuzzy inference system, which includes the following steps: setting a boost converter, the boost converter including an inductor, a switch and an output capacitor, the switch The switch is controlled by a continuous signal to convert an input voltage into an output voltage; a digital computer is set up, and the digital computer includes a processor and a database. The database receives the continuous signal and stores the continuous signal in the Database; performing a sampling process, the digital computer samples the output voltage and a reference voltage according to the continuous signal, and stores a sampled signal in the database; performing a data training process, the processor accesses the database, The sampling signal and the continuous signal are used to establish a boost converter control model through an adaptive network fuzzy inference algorithm; and the digital computer generates an actual control signal through the boost converter control model, and controls the actual The signal is sent to the switch to control the boost converter; wherein the sampling signal includes a difference between the output voltage and the reference voltage and a change between the difference and the difference corresponding to the previous continuous signal value. Power conversion based on adaptive network fuzzy inference system as described in claim 5 The digital fuzzy control method of the device, wherein the adaptive network fuzzy inference algorithm includes 5-layer directional node operations, the difference value and the change value are used as an input parameter, the continuous signal is used as an output result, and the data The training program generates the boost converter control model. The digital fuzzy control method of a power converter based on an adaptive network fuzzy inference system as described in claim 6, wherein each directional node operation includes an operation with a node parameter set and an operation with an empty parameter set. The digital fuzzy control method for a power converter based on an adaptive network fuzzy inference system as described in claim 5, wherein the boost converter includes a DC-DC converter or a DC-AC converter.

Images